Linux Format UK

NoSQL: HOW TO GET STARTED WITH RIAK GET STARTED WITH LINUX! Beginner’s guide Try Linux instantly Dual-boot Windows FREE full software! PLUS The top three distros explored!

Get into Linux today!

RASPBERRY

Pi 2!

500%

FASTER

EXCLUSIVE! Hands-on guide and in-depth review of the all-new Pi 2

Made of Wynn

Python 3!

Wynn Netherland on looking after GitHub p44 Blogging

Video streaming

Get into Ghost

Linux Motion

How to install the next-gen blog platform

Build a secure home CCTV system

Start programming today and easily port your code TuxRadar.com

‘Optimising for happiness’ is kind of key to the Ruby community

Welcome Get into Linux today!

What we do We support the open source community by providing a resource of information, and a forum for debate. We help all readers get more from Linux with our tutorials section – we’ve something for everyone! We license all the source code we print in our tutorials section under the GNU GPLv3. We give you the most accurate, unbiased and up-to-date information on all things Linux.

Who we are This issue we’re asking our experts how we can the next gen of Linux users excited and involved with free and open source software? Jonni Bidwell Well, we’ve got Justin Beiber Linux, unfortunately. But we’ve also got an exciting movement of kids learning to code in schools. Working together, or fighting among themselves, is pretty much the spirit of open source. They get to improve on the work of their peers, or lambast their idiotic methods. Just like the LKML

Neil Bothwick Android is crucial to this. People exposed to the iOS ecosystem from an early age take the closed nature of software and services for granted and become more difficult to show the usefulness of free and open software. Show them that they can own and change their devices and they will always want that.

Les Pounder The next generation of FLOSS users are now learning with the best bit of kit: the Raspberry Pi. A low-cost Linux computer that uses open source software to do so much for such a small price. The mix of cheap hardware, free software and a host of varied applications is catnip for parents and kids alike.

New Model Pi Army The global phenomena that is the Raspberry Pi enters its next phase and is taking Linux with it, not so much kicking and screaming, but more skipping merrily through classrooms, workshops and coding clubs around the world. A charitable venture that was originally envisioned to produce a few thousand Pis has become an educational gateway to Linux for literally millions of new users. Crucially not any old users but a new generation of programmers, developers, system admins, kernel contributors, bug hunters and security experts. All open-eyed and open-minded to the open source way. Boys and girls as young as seven are coding with Linux. I can’t emphasise how vital this is, as it has been noted that the age of developers maintaining crucial elements of the Linux world are ageing. We jokingly call them greybeards, but unless a new generation comes along to maintain those distros the open source world is going to find itself in trouble. That’s why it’s amazing news an all-new Raspberry Pi is on its way. We’ve got complete in-depth coverage on the Pi starting on page 26, with a report on the launch event on page 6. We’ll of course have more coverage over the year, so stick with us or even save yourself some money and subscribe on page 34. If you’re new to Linux or looking for a change, join us on page 36 where we’re looking at the new generation of Linux distros and how you can enjoy a fresh install of a new operating system that’ll be faster, more stable and offer you all-new features. It’s a cool aspect of Linux that new developments come along when they are ready, and not when a company thinks it’s time to to sell you something else with a big marketing campaign.

Mayank Sharma To get the next-gen of users we need to pipe FLOSS through next-gen devices. So more open source apps in the Steam app store, more Linux distros for the Raspberry Pi, and more Android in portables. And yeah, a little sprinkle of Snowden-like revelations to get people to distrust proprietary software.

Alexander Tolstoy I think we should think about developing an IDE for kids –that would help greatly. Also, how about writing a C++ code by dragging colourful blocks and balloons, and without any keyboard input? I’m sure, that would create the right amount of joyance about Linux from younger people, even more so than with the excellent Tuxpaint!

Neil Mohr Editor [email protected]

Subscribe today See p34 www.linuxformat.com

March 2015 LXF195 3

Contents

“If more of us valued food above hoarded gold, it would be a merrier world.” J.R.R. Tolkien

Reviews ThinkPad Yoga 11e .......... 15 A must-buy Chromebook for those who want to be able to run it over with a tank.

OpenPi ................................16 A new box with a Raspberry Pi Compute at its heart and open source schematics.

The new

Raspberry Pi 2 has landed! The launch event, p6 In-depth review, p26 Hands-on guide, p28

The OpenPi will suit entrepreneurs with ideas and industrial aspirations.

Firefox 35 ........................... 17 Mayank Sharma tests an even more refined browser and its global comms feature.

Roundup: Scripting languages p20

TigerVNC 1.4.0..................18 Tiger, tiger, burning bright is your VNC got the features or out of the fight?

Don’t call it a comeback. TigerVNC gets a new major release.

Civ: Beyond Earth.............19 Just Civilization 5 in space or something else a little bit different and a little special?

Raspberry Pi 2 .................. 26 A new Raspberry Pi pops out of the silicon oven for Les Pounder to devour.

The Raspberry Pi 2 is six times faster you say? Let’s find out.

4 LXF195 March 2015

Interview We’ve traded an open web in these last few years for more walled gardens Wynn Netherland on how the web has changed p44 www.linuxformat.com

On your FREE DVD Fedora 21, Ubuntu 14.10, PCLinuxOS, ArchBang 2015 and more Killer distros from our cover feature. Treat yourself or a PLUS: HotPicks and free eBook

p96

loved one to an LXF subscription! p34

Don’t miss...

Get into Linux ........................ 36 Join the growing ranks of Linux users. Explore and instantly try three top distros.

Riak NoSQL ................................ 50 NoSQL, Sir? Suit you, Sir? Discover why admins love a bit of NoSQL.

Coding Academy

Tutorials Minecraft/Pi Turtle graphics ................. 70

Python 3 ................................ 84 Jonni Bidwell places a gentle arm around probably the least loved sequels in programming history and humbly asks that you embrace the parentheses.

Discover exciting graphical things you can do with a turtle and Minecraft on your Raspberry Pi. You will never look at trees the same way again…

Julia ........................................ 88 Mihalis Tsoukalos explains how to start programming in Julia. This functional language may be great for high-performance numerical and scientific computing, but it’s also a capable general-purpose programming language.

Regulars at a glance News............................. 6 Subscriptions ...........34 Back issues ...............68 Faster Raspberry, Pi! Pi! Also Ubuntu

Our digital editions and print/digital

Don’t look back in anger, hit the past

Core’s got its beady eye on IoT and

deals are now available on Android!

with a sturdy back issue. Why not buy LXF193 – it’s very orange.

Canonical teases details of its phone.

Sysadmin...................58 Mailserver................... 11

Dr Chris dons his poetry cap and

In this month’s news bag, some

then batch processes his holiday

Next month ...............98 An anonymity themed issue so

question the importance of Stallman

snaps. He also concludes his look at

private that even we don’t know

while others offer other ways to erase OpenLDAP, which is nice.

what’s in it!

a hard disk (other than shooting it). Alexander Tolstoy is a man of many

Les Pounder celebrates three (-ish)

picks that are just too hot to handle

years of the Raspberry Pi.

Dropbox Uploader, Gloobus-preview,

Is there life after Bash? We corral

MPS-Youtube, KWave, Linuxbrew,

other scripting languages that meet

SuperTuxKart, Aualé, Hollywood

the needs of a busy sysadmin.

Technical Melodrama and Boomaga.

Neil Bothwick says, step away from your mouse, but not so far that you can’t reach the keyboard to manipulate images using the power of your command line.

Motion Detect and record............. 76

Ghost Embedded controllers .... 80

and he’s throwing your way now! Get ready to catch: Yarock, FFmpeg,

Roundup ....................20

ImageMagick Convert images ................74

Build a livestreaming system for your home using a Pi, webcam and save motiondetected video.

HotPicks ....................62 User groups................14

Turtles are very artistic as well as being obsessed with pizza.

Our subscriptions team is waiting for your call.

www.tuxradar.com

Don’t panic, it’s not that critical vulnerability in glibc that had Jonni “mildly perturbed”, but a nice and friendly tutorial on how to create custom themes for your blog site.

March 2015 LXF195 5

THIS ISSUE: Ubuntu Core

Ubuntu Phone

Mintbox Mini

Bodhi Linux lives

RASPBERRY PI

New Raspberry Pi Model B unleashed! Raspberry Pi 2 Model B is six times faster than the previous model.

T

he Raspberry Pi 2 Model B has been released much to many people’s surprise and a lot more people’s delight. The new model is apparently six times faster than the previous model [see p26 for loads more results] and comes with 1GB of RAM. The processor in the newest Raspberry Pi is a Broadcom BCM2836 ARM 7 quad-core processor running at 900MHz, a substantial upgrade over the previous models. Last year’s Raspberry Pi Model B+, for example, has a single-core 700MHz CPU. In a blog post (www.raspberrypi. org/raspberry-pi-2-on-sale) to announce the new model, the Raspberry Pi Foundation explained the steps it has taken to bring increased performance and add extra features, without deviating from its original mission, or alienating people who own older Raspberry Pis. And the relationship between the Raspberry Pi Foundation and Broadcom was critical to this: “Our challenge was to figure out how to get this without throwing away our investment in the platform or spoiling all those projects… Fortunately for us, Broadcom were willing to step up with a new SoC, BCM2836. This retains all the features of BCM2835, but

Eben Upton launches the Raspberry Pi 2 Model B at The Shard in London.

6 LXF195 March 2015

replaces the single 700MHz ARM11 with a 900MHz quad-core ARM Cortex-A7 complex: everything else remains the same, so there is no painful transition or reduction in stability.” The increase in specs means that the Raspberry Pi 2 Model B can boot up in less than half the time of its predecessor. It comes with a 40-pin GPIO enabling even more complex projects that include multiple sensors, connectors and expansion boards. Thankfully for those of us who have created projects on the earlier Raspberry Pi, the first 26 pins of the

“The new Pi can boot up in less than half the time of its predecessor.” new Pi 2’s GPIO are identical to the Model A and B boards. According to the Foundation, porting existing Raspberry Pi projects to the new model is as simple as updating the OS. At the unveiling of the new Pi, Raspberry Pi Founder, Eben Upton said that while the previous version of the titular PC was a “great computer”, users “had to be forgiving” when undertaking certain tasks due to its comparatively low processor power, but that the Pi 2 will be able to be used as a fully-capable PC without compromise. Upton has also blogged that the Raspberry Pi 2 will not only run ARM

www.linuxformat.com

The Raspberry Pi 2: faster, more GPIO pins, backwards compatible and the same low price.

GNU/Linux distros, including Snappy Ubuntu Core (see page, right), but also Windows 10. The Foundation has been working closely with Microsoft for the past six months to bring a Raspberry Pi 2-compatible version of Windows 10, which will be free of charge to developers. This isn’t news that everyone will celebrate, but it does reflect the success of the Pi, generally. The best news of the Raspberry Pi 2 is that not only has it gone on sale immediately – you’ll be able to buy one by the time you read this issue – but it also keeps the same low price as the previous models of around £26.42. You’ll need an updated NOOBS or Raspbian image including an ARM v7 kernel and modules, which can be downloaded from www.raspberrypi. org/downloads. Sadly, this new version doesn’t mean there’s been an official price drop for the Raspberry Pi 1 Model B and Model B+, which will continue to sell for the same price, though we should see price drops occurring once the Pi 2 picks up steam.

Newsdesk INTERNET OF THINGS

Ubuntu Core wants to power the future A Small, lean distro that will make internetconnected devices even smarter.

C

anonical has just launched its Snappy Ubuntu Core partner ecosystem, working with 22 partners to help power some of the most innovative and exciting projects which cover a wide range of upcoming devices such as robotics, drones and other smart devices. “We are inspired to support entrepreneurs and inventors focused on life-changing projects,” says Mark Shuttleworth, Founder of Ubuntu and Canonical. “From scientific breakthroughs by autonomous robotic explorers to everyday miracles like home safety and energy efficiency, our world is being transformed by smart machines that can see, hear, move, communicate and sense in unprecedented ways.” Ubuntu Core is a pared back version of the popular Linux distro, and only requires a 600MHz processor with 128MB RAM, enabling it to bring a robust and open ecosystem to upcoming smart devices. It will also provide much needed security to devices that cannot easily connect to the internet for updates and security patches. We're already seeing some pretty cool devices use Ubuntu Core, including the Erle-Copter.

Snappy Ubuntu Core is set to power some really exciting projects.

According to Victor Mayoral Vilches, CTO of Erle Robotics, the benefits of Ubuntu Core are clear: “We are delighted to reveal the Erle-Copter as the world’s first Ubuntu Core-powered drone that will stay secure automatically and can be upgraded with additional capabilities from the app store … An open platform attracts innovators and experts to collaborate and compete, we are excited to lead the way with open drones for education, research and invention.” We can't wait to see what future Ubuntu Core-powered devices emerge, and it's great to once again see Linux at the forefront of innovation, where it belongs. http://developer.ubuntu.com/en/snappy

MOBILE

Ubuntu Phone Not another 'me too' Android clone, says Canonical.

C

anonical's much anticipated début into the world of mobile operating systems is getting nearer to release, and as the launch date approaches we're getting ever more details about what we can expect when we fire up an Ubuntu Phone-powered device. Pleasantly, what we've heard so far shows that Canonical is looking to differentiate its offering compared with Android and iOS, especially in terms of apps. Unlike in Android and iOS where apps are tightly restricted and controlled by the platform owners (especially in Apple's case), and placed in a grid for users to dispassionately poke at, Canonical is trying something rather different with Ubuntu Phone. Rather than displaying separate apps in a standard grid, Ubuntu Phone is looking to integrate content and services via

Scopes, which will integrate various apps and services into an easy to use interface, so users don't have to scroll through screens and screens of icons. For example, the NearBy scope will aggregate local services centred around where you are and what you're currently doing. Going up against Android and iOS will prove to be a difficult task, however, but it's good to know that Canonical has a plan to make Ubuntu Phone stand out. http://www.ubuntu.com/phone Canonical’s Ubuntu Phone is shaping up to be much more than a fancy-looking mobile OS.

www.tuxradar.com

Newsbytes A new small-form desktop computer has been announced; the MintBox Mini. As you've probably guessed from its name, it comes with Linux Mint already installed. It will feature an AMD A4-6400T Micro processor, Radeon R3 graphics processor, 4GB RAM, two USB 3.0 ports, two USB 2.0 ports, dual HDMI ports (for up to two displays), a microSD card reader, and a headset jack. It will also have a 64GB solid state drive and comes in a mint green case that makes us feel fresh and clean. It'll cost $295 in the States and €295 in Euroland. www.fit-pc.com After years of promises and stalled attempts, Samsung has finally launched a Tizen-powered smartphone, the Samsung Z1. It won't challenge high-end smartphones like the Nexus 6, with its 4-inch WVGA screen, 1.2GHz dual-core processor, 768MB of RAM, a 3.1-megapixel camera, 4GB of storage and 3G data limitation, but it’s aimed at emerging markets. The Z1 has launched in India, setting Tizen up as an operating system for developing countries The Z1 is small, to rival Google's own attractive and Android One platform. runs Tizen. Cast out your Macbook Pros, there's a new high-end laptop in town, and one that claims it is the first that respects your freedom and privacy. The Purism Librem 15 comes with a kernel, operating system and software applications that are all free and open source. Along with some pretty neat specifications, including a 15.6-inch display in either 1080p or 4K resolutions, it comes with a Trisquel-based 64-bit Linux operating system. According to the manufacturers, the hardware used in the Librem 15 laptop was specifically selected so that no binary blobs are needed in the Linux kernel that ships with the laptop. It's not available to buy from stores at the moment, but you can pledge support to get an early version from www.crowdsupply. com/purism/librem-laptop. Bodhi Linux is alive! There had been concerns that the project was shutting down, due to some pretty big changes happening. The good news it that Bodhi Linux 3.0 is still being worked on, though there's a new timeline so we're not sure when we can expect the new release.

March 2015 LXF195 7

Comment

Cabinet Office plug-fest I was at the Open Document Format (ODF) plug-fest last month hosted by the Cabinet Office, in partnership with the OpenDoc Society. The UK Government’s policy mandating ODF for editing and sharing documents, announced in July, commits all departments to adopting the format. After years of single-vendor dominance, it was great to see the blossoming of options on-display from the ODF ecosystem at the plug-fest. There was an announcement that Google Docs will upgrade its ODF support to 1.2 and include presentation support too, which was very welcome. I also tried out Google’s ODF support afterwards and experienced the improvements over recent times. Previously, I would always choose an OOXML download, since I knew LibreOffice’s OOXML import was better than Google Doc’s ODF output. It’s great to see those filters approaching parity.

Open doors Microsoft re-iterated its support for ODF, but hedged unhelpfully on change tracking, a key feature request from Government department present, which remains an unimplemented in Microsoft office. LibreOffice was well represented with excellent ODF change-tracking support built-in, with Collabora contributing a large number of interoperability tests during the hacking to help to improve document exchange. Naturally there were other participants from many smaller projects, Apache OpenOffice, Calligra, EuroOffice and many others; a total of over 50 delegates from 30 organisations, including 17 Government representatives. The choice of vendors enabled by ODF provides an exciting side-effect: of opening the platform choice door, so making a Linux OS and LibreOffice combination a viable option in Government. Michael is a pseudo-engineer, semi-colon lover, SUSE LibreOffice hacker and amateur pundit.

10 LXF195 March 2015

Hitting the mirrors What’s behind the free software sofa?

EVOLVE OS BETA 1 The first ever beta of Evolve OS has been released, so if you want to check out this brand new Linux distro that's been built from scratch and don't mind a few bugs here and there, you now can. With a custom desktop known as Budgie and it's own package manager that's been forked from Pardus Linux, Evolve OS is shaping

Evolve OS is only in beta, but it's already looking very promising.

up to be a rather unique distribution and can be downloaded from https://evolve-os.com.

TINY CORE LINUX 6.0 The latest version of this minimalist Linux distro is known as the ‘piCore’ edition, and as the name suggests it's been designed for the Raspberry Pi. The biggest change with the latest release is the use of the 3.12.36 Linux kernel, offering more stable performance and broader hardware support. A new feature also allows Tiny Core Linux 6.0 to boot into a safe overclocking mode that shortens

boot time by 20%. Download it from http://bit.ly/TinyCoreLinux6.

It's small, compact and is a good alternative to Raspbian on the Pi.

PROXMOX 4.0 MAIL GATEWAY If you're after a new email proxy, then Proxmox 4.0 Mail Gateway could be of interest. It's based on Debian 7.8, and the latest 4.0 version of the anti-spam and antivirus protection has had all its packages updated. Proxmox protects email servers from spam, viruses, trojans and phishing and is managed through an easy, web-

based interface and has been going for over 10 years, so check it out at www.proxmox.com.

Compatible with every type of mail transfer agent, including Postfix.

4MLINUX 11.0 Another lightweight Linux distro that has been recently updated is 4MLinux and it’s on the mirrors to download. It's been built from scratch and uses a customised JWM window manager. Version 11.0 is a major release that brings GNU C Library 2.20 and GNU Compiler Collection 4.9.2, as well as a new lightweight rescure live CD 4MRescueKit. Download it from http://4mlinux.com.

www.linuxformat.com

4MLinux focuses on the four Ms: Maintenance, Multimedia, Miniserver and Mystery (cool games, basically).

Write to us at Linux Format, Future Publishing, Quay House, The Ambury, Bath, BA1 1UA or [email protected]

Wipe Dban Thanks for the recent article about erasing hard drives. I was disappointed, though, to see the omission of nwipe. Some time ago I forked the dwipe command (the wiping software in DBAN), due to the (then) poor maintenance of DBAN. Now that DBAN is near-enough proprietary software, it seems even more appropriate that nwipe be used instead. It has several additional features to DBAN: much better hardware support (as it's OS-agnostic) and is contained in several live distros if you prefer a DBAN-like experience (most notably PartedMagic). I'm no longer the maintainer, but Martijn van Brummelen (the software's Debian packager) kindly took it over recently. Andy, via email. Neil says: Thanks for the details here Andy. So you know, the background to most if not all of Sean’s articles are real-life application. I believe the erasing article was based on events from a while back, but he chose DBAN at the time. We’ll keep an eye on Nwipe and hopefully cover it in the future.

Root canal I read the article in LXF179 about rooting an Android mobile and you mentioned two main ROM sources, being TWRP and

Letter of the month

Minimalist

I

'm looking at my Christmas 2014 edition [LXF192] and I think Bothwick's Screen article [Tutorials, p74] is interesting because of what it misses. I'm a minimalist, but that's not my topic now. I'm running Slackware64 13.37 here. My Slackware came with a variety of Screen managers, and of those, I chose TWM. I've looked at the others and for me, the TWM wins by far. And looking at Screen, what it does, TWM does much better. I can start, close, move, resize, reduce to icons, retrieve windows according to my need (generally up to three). These windows can serve as Terminal windows or they can start other resources like Firefox or the Thunderbird I'm using now. In my view TWM has what I need. I removed the gimmicks from TWM's overly ambitious startup array, mine opens with one 'login' window which is the least I'll need. I think that one of these days, a good

Clockwork Mod. The device I was intending to root is a Samsung Galaxy Note (GTN7000). TWRP doesn't do a ROM for this device but CWM does. The problem was that the CWM download is a ZIP file (Superuser.zip) with the following four folders (armeabi, META-INF, mips, x86), two files install-recovery.sh and

piece on TWM would prove helpful to many of your A basic but easily viewed readers and I'd tabbed window manager, so like to see any here’s a grab from the year that I may 2000 and issue 2. have missed. And there very much is something I've missed. At age 83 I am getting some macular degeneration and some cataracts, so I like strong dark print on my screen. It turns out Bash offers a font option that you find using Ctrl+right mouse button, that offers a font option labelled 'Huge.' So I think something in your pages about those hidden options in Bash would be good; including, is there some way I could lock-in that dark font option in my Bash? Martha Adams, via email. Neil says: True to your minimalist ways my reply is: yes. For everyone else we are planning to run an Escape the GUI feature in a couple of issues.

Superuser.apk. I felt that I couldn't proceed as your instructions based on the TWRP solution as it puts an image specific file in the fastboot path. When I've looked at various forums on this they seem to be copying a ZIP file to the device's external micro SD card and booting from that. Is there any

chance that you could add a bit more detail for the alternative CWM method please? I’m using Linux Mint 17. Sorry if I'm being over cautious, but this is my first rooting attempt and I don't want to brick the device. Roger Ralph, via email. Neil says: LXF179 was back in October 2013 and you’ll find many

[email protected]

www.tuxradar.com

March 2015 LXF195 11

Mailserver

keyboard issue, if your passkey has punctuation this can change depending on the keyboard layout used. Something, I’ve discovered to my annoyance in the past, as # and £ symbols can jump around rather annoyingly. I’d be surprised if the driver wasn’t working as it’ll be a bog-standard Intel unit. Otherwise good luck with the Linux learning, it’s great fun to get stuck into and offers almost endless entertainment!

Xara Root access opens up your phone to more advanced apps and complete backup tools.

things have changed since then; usually making life far easier. The issue that always crops up is that ROMs, boot loaders and root files are model specific. As you mention it looks like the approach for the GTN7000 is to use the standard Recovery Bootloader and use this to flash the root SuperUser app using a standard shell script, all run from an SD card, which is much easier. A guide and the files are here: http://bit.ly/1CfjTED. For total control you can also pop on Cyanogenmod. This is somewhat more involved but you can find out how to do that here: http://wiki.cyanogenmod. org/w/Install_CM_for_n7000.

WiFiless There’s a recurring problem that I have with the disc that comes with your (excellent) magazine. Although the discs boot and load the distro, I am unable to connect to my Wi-Fi. I’m a Linux user, Ubuntu 14.10 and Linux Mint 17 Mate in a dual-boot setup on my Samsung NP350v5c Laptop. My only connection to the Internet is via a Mobile (3G) Wi-Fi hub running on Three. I installed Ubuntu around eight months ago and Linux Mint just over a month ago, without issue. In both cases, however, the installation discs came from a source other than yourselves. I’m fully aware that the problem


may be with me or my particular hardware setup. I am new to Linux having been away from computers for 20 years or more, other than limited use at work. When purchased it came with Microsoft Windows 8, which lasted around five weeks, the last three of them spent finding out how to circumvent the BIOS in order to boot the Ubuntu disc. Other than this very minor problem, I am having a great time learning how Linux works. On my list of things to do is learn Python. Just because I’m 65 doesn’t mean I am unable to learn new stuff, which is one of the reasons for buying your, and other, Linux magazines. Tony Maughan, via email. Neil says: Is it a case that the wireless isn’t working at all or that you can see the hub (so the drivers are correct) and it just won’t allow you to connect? In the last case it could be a US/UK

Is there or will there be a Linux version of Xara's Bitmap Editor? Ian Learmonth, via email. Neil says: Why the heck do you want Xara Bitmap Editor? We’re not actually sure that Xara has released a package called that, but certainly it doesn’t support Linux. You could try running Xara Designer 10 via Wine, though its rating of Silver is hardly reassuring. Honestly, give Gimp a spin, we even gave away a bookzine on it with LXF194 so go grab the ISO now, or better, order a back issue from http://bit.ly/ LXF193GIMP.

www.linuxformat.com

Mind games I would never want to diminish Richard Stallman, but I do have to take a tiny little bit of issue with this bit of science fiction. BSD comes from a university that was always something of a renegade within the Unix world, and was a part of the general movement that involved Richard. Who knows what may have come from them in an alternate history? It may have been years after Richard started his efforts, and also a commercial product, but it was showing a different, more capable and robust solution than the Microsoft products. It was also not the only product to try to bring the Unix environment to the huge market being generated by the IBM PC computer, and its ‘clones’, indeed Xenix development had Microsoft involved from it's earliest days. Unix was the primary direction being taken to bring a more capable environment to users. Minix was an independent solution to a related need, and we will never know what may have come of this project without the presence of Linux. There seems to have been many paths to the present world of open source, and some are almost as inspired and generous as the efforts of Richard Stallman. History would not be the same without Richard, take any person out of the

The world would have been very different without Richard Stallman.

Mailserver

story and it changes. However, in my honest opinion, we would have found (stumbled?) our way to a world not unlike this world without Richard Stallman; if only because of the massive dissatisfaction felt by so many with the limited vision of the corporate ‘machines’ that were trying to manipulate the computer industry to their financial benefit. It would very likely have been a longer path. But the impetus was there from way back. To ignore all the other contributors to what we now enjoy, and suggest that nothing would have happened without Richard, seems a bit unkind to all the other contributors. John Paterson, via email. Neil says: Absolutely, Jonni was appalled at the notion open source could possibly ever be “wiped from the surface of the earth” as it’s human nature to share knowledge. As I mentioned it was a fun academic exercise with Stallman being the embodiment of the open source movement. Being more realistic you’re right, the world probably would have rallied around BSD, but with Minix it always seemed Tanenbaum wanted it to be commercialised, though that’s wild speculation. I do think it’s important not to underestimate the need for a figurehead to rally a cause and that’s what Stallman brought to the world. Obviously not without a near unquantifiable amount of help from generous and stupidly talented developers from around the globe. We’d never want to diminish the contributions made by everyone, large or small, to the FOSS world.

Education distro, Escuelas Linux has a new release. The world needs more well-supported Linux distros.

problem trying to load the MB drivers. From the MB CD I get "Please update to the latest Linux Kernel for motherboard chipset and components support" I’ve purchased LXF187 to ‘Escape Windows’ but I am still stuck. I don’t want to load Windows XP on to a separate partition. Info on how to proceed will be greatly appreciated. Robert, via email. Neil says: Ahh you’re falling for the oldest Windows user’s trap of them all: driver updates. Unlike Windows, Linux doesn’t really need drivers at all, as they’re all built directly into the Linux kernel itself. The message is basically a default error down to you trying to run a CD built for Windows on Linux. I’ll now contradict myself and say that driver updates are possible, but not often required.

No drivers A few months ago I decided to change my old computer with Windows XP OS for a new rebuild and a later Windows OS. Having assembled a new box with an Asus Maximus V11 Hero, an i7-4770K, and Sapphire R9 280X. I then spotted your Linux Format mag and was hooked. I have loaded Ubuntu 14.10 64-bit from your CD and updated from the net. I have a

Escape Windows and leave drivers behind. Well, most of them.

The biggest exception for this ‘third-party graphics drivers’ is for your Radeon card. This is a complicated bug-bear for Linux and is down to closed-source elements that can only be installed via Nvidia or AMD. See the Settings > Additional Drivers control panel.

Sanitised In Sean Conway's article about erasing drives, [Tutorials, p68, LXF193] he says that shred is impractical for sanitising a hard drive due to the need for it to be run on individual files. While this is indeed true, what is a device if not a file? You can therefore run shred on both a device (eg /dev/sda) or partition (eg /dev/ sda1). Run this with flags -n 3 -z -v and you’ve successfully removed all files in one quick and easy command with no chance of recovery. When decommissioning a disk, boot knoppix or another live distro and you can sanitise a disk easily without need for anything more complicated. Just ensure you get the right device… Bret Giddings, via email. Neil says: says: Great advice, Bret! And you’re right devices are just treated as files as Unix demands. We’d even argue that -n 1 is fine from the latest research on modern drives, as the -z option does a second write any way with zeros. You still can’t beat Sean’s shooting option though can you?

www.tuxradar.com

Finally in English I'd like to point you to an educational Linux distro. Today we’ve released 3.0.22 of Escuelas Linux: http://bit. ly/1yrJcTm. You might not have heard about this distro before, but we have been around for a lot of years. Until recently we decided to go outside our local focus and to add English language support. Are we a small distro? That depends. In the schools we support ourselves, and we have reached 44,864 students and 2,541 teachers using voluntarily our distro. And since September 2014, Escuelas Linux has been downloaded from 36 countries via SourceForge. I hope you take a look, that would mean a lot to me as a subscriber to your beautiful magazine! Alejandro Díaz. Neil says: Always good to hear about well-supported Linux distros and thanks for supporting us English speakers! LXF

Write to us Do you have a burning Linuxrelated issue you want to discuss? Want to let us know what inventive uses you’ve been finding for your Raspberry Pi, or suggest future content for the magazine? Write to us at Linux Format, Future Publishing, Quay House, The Ambury, Bath, BA1 1UA, or email us at [email protected].


Linux user groups

United Linux! The intrepid Les Pounder brings you the latest community and LUG news.

Find and join a LUG Blackpool Makerspace Meet every Saturday, 10am to 2pm. At PC Recycler, 29 Ripon Road FY1 4DY. http://blackpool.lug.org.uk Bristol and Bath LUG Meet on the 4th Saturday of each month at the Knights Templar (near Temple Meads Station) at 12:30pm until 4pm. www.bristol.lug.org.uk

Hull LUG Meet at 8pm in Hartleys Bar, Newland Ave, 1st Tuesday every month. http://hulllug.org Lincoln LUG Meet on the third Wednesday of the month at 7:00pm, Lincoln Bowl, Washingborough Road, Lincoln, LN4 1EF. www.lincoln.lug.org.uk Liverpool LUG Meet on the first Wednesday of the month from 7pm onwards at the Liverpool Social Centre on Bold Street, Liverpool. http://liv.lug.org.uk/wiki Manchester Hackspace Open night every Wednesday at their space at 42 Edge St, in the Northern Quarter of Manchester. http://hacman.org.uk Surrey & Hampshire Hackspace Meet weekly each Thursday from 6:30pm at Games Galaxy in Farnborough. www.sh-hackspace.org.uk Tyneside LUG Meet from 12pm, first Saturday of the month at the Discovery Museum, Blandford Square, Newcastle. www.tyneside.lug.org.uk

Hap-Pi Birthday Time to don a party hat and faceplant the cake.

W

e are fast approaching the party you'll have the chance to learn third birthday of the from the best. The team behind Pi Wars Raspberry Pi (Well, strictly will also be on hand running a series of speaking that won't happen until 2016 obstacle courses for robots. Will yours as the Raspberry Pi was released on survive the trials? There will be plenty of February 29, 2012, but February 28 is a stalls from the most popular vendors in good compromise) and to celebrate the the community too, so you're Raspberry Pi Foundation has teamed guaranteed to see lots of the latest up with Michael Horne and Tim Pi-related technology on offer. Richardson, of CamJam and Pi Wars Of course, no celebration would be fame respectively, to put on a big complete without a proper birthday birthday extravaganza in the home of party and on the Saturday night there the Raspberry Pi, Cambridge. The twowill food and drink for all and a series of day long party will celebrate the great games and quizzes, all suitably projects and ideas that have been Pi-themed, of course. So grab your created thanks to this small computer tickets from http://bit.ly/ and there will be panel discussions PiBigBirthdayWeekend and get ready featuring prominent names in the for a birthday blowout! LXF Raspberry Pi Foundation, such as Eben and Liz Upton and other talks hosted by community members, such as Ryan Walmsley and Charlotte Godley. As well as great talks there will be a series of hands-on workshops for all skills levels. Have you ever wanted to hack a robot, send a Pi into space or The Raspberry Pi party is your chance to make music? Well at this birthday rub shoulders with a great community.

Image Credit: Richard Kenworthy

Community events news

Southend Raspberry Jam The next Southend-On-Sea Raspberry Jam will take place on 21 February and its schedule looks chock full of great things. There are three all-day workshops covering soldering,


Python and Scratch GPIO, alongside Sonic Pi 2 and environmental sensor hacking with Astro Pi. This great event is organised by the SOSLUG team and will take place at the Tickfield Centre from 10am until 5pm. This is a great opportunity for children, parents, adults and teachers to mix together and talk about their projects. You can find out more by visiting their EventBrite website at http://bit.ly/ SouthendRaspJam5.

Covent Garden Jam The Jam is back for 2015 and this time its teaming up with the local Coder Dojo to bring a full day of coding, hacking and learning for all the family. This event takes place on 28 February at the DragonHall in Covent Garden from 2pm until 5pm, and pairing the Jam with a Coder Dojo enables two communities to work together for a common goal. Head to the website for tickets http://bit.ly/ CoventGardenJam2.

www.linuxformat.com

Makerfaire UK 2015 We’ve mentioned this event recently, but we don’t want you to miss out on this great two-day maker event at the Life Centre, Newcastle starting 25 April. This is the largest makerfaire in the UK and there are many great projects on show. For instance, this was where Matt Venn showed off his Raspberry Pi balloon camera project in 2014. Tickets for the event are around £10 per adult. www.makerfaireuk.com

Xxxxxxxxx Reviews

All the latest software and hardware reviewed and rated by our experts

ThinkPad Yoga 11e With a cost that's in the ‘proper’ laptop ballpark David Eitelbach wonders if a flexible, durable Chromebook is worth the money? Specs CPU: 1.83GHz Intel Celeron N2930 (quadcore) GPU: Intel HD Graphics RAM: 4GB DDR3 (1,333MHz) Display: 11.6inch, 1,366 x 768 HD LED Storage: 16GB SSD (eMMC) Ports: HDMI 1.4, USB 3.0, USB 2.0, 4-in-1 card reader Comms: Intel 7260 802.11ac (dual-band), Bluetooth 4.0 Camera: 720p HD webcam Weight: 1.5kg (3.3lbs) Size: 11.81 x 8.5 x 0.87 inches (WxDxH)

T

he Yoga is one of the few Chromebooks on the market with a truly rugged exterior. Designed for use in rough-and-tumble classroom environments, it boasts rubberised edges and extra-strong hinges. At about £360 inc VAT, however, the Yoga 11e is more than twice as expensive as most Chromebooks. So is this machine's combination of flexibility and durability worth the price? The Yoga 11e sports the same minimalist aesthetic as other ThinkPads, featuring a matte black plastic chassis and few adornments other than a few silver ThinkPad logos. At 300 x 216 x 22mm and 1.5kg, the Yoga 11e is a bit heavier than the Lenovo N20p and the 13.3-inch Toshiba Chromebook 2. Still, the laptop is light enough that you would barely notice it in your bag. As with other notebooks in the Yoga series, the lid of the Yoga 11e can rotate 360 degrees to lie flush against the bottom of the laptop. This flip-screen design lets you use the device in four modes: traditional laptop mode, tablet mode (folded completely backward), stand mode (placed keys-down on the table) and tent mode. This final mode sees the Chromebook at a 270-degree angle, standing on the edges of its lid and base. Rotating the screen automatically disables the keys. While the Yoga 11e is comfortable to hold as a tablet, its utility in this mode is limited. The Chrome OS interface features dense clusters of small buttons, particularly in the upper right corner of the browser, that are tricky to touch accurately. Unlike most Chromebooks, the ThinkPad Yoga 11e was built to withstand both the rigours of the classroom and the battlefield.

Rubberised edges on the lid protect the laptop from unexpected drops, and extrathick bezel minimises damage to the LCD panel, while militaryspecification testing means the notebook can withstand extremes of pressure, humidity, vibration, temperature and airborne dust. Multitasking is where this Chromebook runs up against its limited hardware. Using Google Docs while streaming music with a dozen other tabs open, for example, the audio occasionally stutters. Worse even, periodically there are half-second delays after pressing on the arrow keys.

Solid components Thankfully, the Yoga 11e's 11.6-inch, 1,366x768 touchscreen delivers bright and crisp visuals. Colours popped when watching HD video, and text looked sharp on websites. However, the viewing angles are shallow: moving more than a foot to either side of the screen, colours begin to wash out. The touchscreen responds promptly to input, and had no problem with pinchto-zoom. The notebook's rear-facing speakers pumped out surprisingly wellbalanced and load audio. We've come to expect an outstanding typing experience when using a ThinkPad, and this doesn't disappoint. The island-style keys offer plenty of vertical travel and tactile feedback, and the textured surface makes it easy to touch type. The touchpad is large and responsive, and the scrolling is easy using two fingers, while the cursor accurately tracks finger movement across the pad. Battery life is a letdown, however. The Yoga 11e falls

www.tuxradar.com

A Chromebook that can manage more gymnastics than a Russian teenager at the Olympics.

short of competing Chromebooks. With the brightness at 30%, a dozen tabs open and streaming music, the Yoga lasted 6 hours and 21 minutes. The Lenovo Chromebook is more than twice as expensive as many Chrome OS-powered laptops. Still, it's one of very few Chromebooks that’s a true 2-in-1 laptop. Plus, the Yoga's ruggedised exterior will help it endure the wear and tear of life. LXF

Verdict Lenovo ThinkPad Yoga 11e Developer: Lenovo Web: www.lenovo.com/uk/en/ Price: £359

Features Performance Ease of use Value for money

6/10 7/10 9/10 6/10

A rugged and flexible design with an excellent keyboard. If only the Yoga 11e didn't cost over £350.

Rating 7/10 March 2015 LXF195 15

Reviews Raspberry Pi expansion

OpenPi Ever intrigued by little boxes and blueprints, Jonni Bidwell explores a Pipowered Internet of Things hub that you can redesign yourself. In brief... A Pi Compute Module-based board with a subGHz wireless tranceiver, a RTC and a thermal sensor. The board and schematics are open source so budding entrepreneurs and inventors are free to modify them.

Specs Raspberry Pi Compute Module: Broadcom 2305 SoC 700MHz, 512 MB RAM (shared with GPU), 4GB eMMC storage OS: Raspbian Ports: 2x USB 2.0, HDMI, 2x micro USB (power & programming) RTC: TI BQ3200 Temp Sensor: Ti TMP1000NA Sub-GHz Wireless: Wireless Things SRF

W

ireless Things is the new name for Ciseco, the Nottingham-based Internet of Things company responsible for the critically acclaimed EVE Alpha device. EVE is a compact device that fits snugly on top a Raspberry Pi, connecting via the GPIO pins. There are a plethora of modules that can be fitted to the EVE board: SRF or RFM12B radios, real-time clocks (RTCs), temperature sensors and more. With the OpenPi, Wireless Things hope to take this idea one step further and it’s in the middle of a crowdfunder as we write this (http:// bit.ly/KickstartOpenPi). Using the Raspberry Pi Compute Module as a base, OpenPi provides two USB ports, a RTC (battery is included), 868-915 MHz SRF transceiver (for long range serial connections) and a temperature sensor. All this is boxed up in a tidy plastic case emblazoned with the Wireless Things logo, and apertures allowing HDMI and power connections. In terms of extensibility, you get 18 (plus two power) GPIO pins (which you'll need to break out yourself), headers for SRF and USB EEPROM programming, as well as wire-by-hand pins for connection to Wireless Things' XBee module. The Compute Module has 4GB eMMC storage onboard which houses a mostly-standard Raspbian install (PuTTY is included for easy serial connections), but this can be reflashed. For the hobbyist user, it could be that a humble, standard issue Raspberry Pi will do the job, you can

Features at a glance

Starter kit

Open schematics

The kit includes a 5v power supply, an HDMI cable, wireless keyboard/ touchpad and dongle.

The inquisitive can download the EagleCAD circuit designs, the inventive can even modify them.


The OpenPi has a footprint slightly larger than the original Pi, the cut corners may or may not be an homage to Battlestar Galactica.

add the transceiver and RTC and any other peripherals easily and cheaply. It may not be a very tidy arrangement, but you can probably Macguyver yourself a suitable container using Lego, gaffer tape or a cigarette packet. A single basic OpenPi pack (comprised of just the Pi Compute module and OpenPi PCB mounted in a case) costs £49 inc VAT, so it comes down to whether having it all in a convenient box is valuable, and whether you will miss the line out and Ethernet connectors.

Open hardware What is unique about OpenPi is that all the board schematics are open source – you can download them as EagleCAD files. In so doing, OpenPi aims to make it easier to make the Compute Module a more flexible platform upon which to develop projects and products. Manufacturer research suggests that form is as important as function, and that this partly explains the hitherto lacklustre uptake of the Compute Module as a platform. They claim that while no single design will suit everyone, using the OpenPi as a reference point will greatly simplify the design process for any budding IoT entrepreneur. So the product is really targeting those with ideas and industrial aspirations.

www.linuxformat.com

Quantity discounts are available (10% for more than 100, 30% for five-figure orders) which speaks to this, but also makes the device more for educators. We found that getting access to the RTC on our prototype required loading a module and sending some bytes to the i2c bus. The thermal sensor can easily be queried with i2cget, but the raw sensor data needs to be re-ordered and scaled to give correct reading. These are things that will be addressed come launch day, it's just a question of making some wrappers. The Kickstarter ends on 4 March 2015, and a release is planned for the end of March. LXF

Verdict Wireless Things OpenPi Developer: Wireless Things Web: http://wirelessthings.net Price: £49/£99 (basic/starter kit)


8/10 7/10 7/10 7/10

A bold step conceptually, which will excite those with big ideas. But home users may prefer existing tools.

Rating 7/10

Web browser Reviews

Firefox 35 Mayank Sharma sings a smouldering rendition of Hello by Lionel Richie as he plays with the voice and video calling feature in the new Firefox. In brief... One of the most popular, feature rich crossplatform open source web browsers. See also: Chromium.

T

here’s no dearth of quality web browsers available on the Linux desktop. Despite the saturation, Firefox has managed to stay ahead of the competition with continuous improvements and innovative additions. Mozilla's latest release, Firefox 35 isn’t bursting with new features, but it does give users a chance to soak in all of the new features that were added in the previous releases. By far the biggest feature in this release is fully functional voice and video calling, dubbed Firefox Hello. Mozilla calls it “the first global communications system built directly into a browser.” Although it’s still in beta the feature works well and manages to deliver everything it promises. Mozilla’s primary intention was to enable users to communicate with each other without handing over their personal information to a third-party. Furthermore, Mozilla wanted to democratise the convenience of a free communication service, which is why you can use the feature to call any WebRTC-compatible browser, such as Chrome, Opera and Firefox, and across all desktop and mobile platforms. Mozilla first introduced the experimental WebRTC feature in Firefox 33 and it was officially launched in Firefox 34, powered by the OpenTok real-time communications platform from TokBox. In the current release, Mozilla has further refined and simplified the process of making and receiving calls.


Refined Firefox Hello

For web developers

The process for video calls has been simplified and now uses a room-based conversation model.

Over 150 devtool fixes and new CSS features, like the filter property and CSS source maps as default.

Firefox’s privacy controls extend to the Hello feature as well. You can turn off the camera or mic with a single click and turn off the ability to receive any calls.

To initiate a call, all you need to do is click the new Firefox Hello icon in the toolbar and then press the Start a conversation button. This action will generate a link to a conversation chat room that you can pass on to anyone you wish to converse with. When you click on a link to start a conversation, a pop-in window opens with your chatroom. When a person you’ve invited also clicks on the link they are added to this chatroom and the Hello icon lights up to alert you.

Call on me The Firefox Hello also enables you to create multiple conversation chat rooms, each with their own unique URL. Additionally, you can name a conversation and keep it for quick access to people you converse with regularly. And did we mention that you can do all this without setting up an account with anyone? If, however, you do get yourself a Firefox account, you can add and import existing email contacts. While adding a contact is pretty straightforward, importing your Google accounts requires you to equip the browser with your Google account’s OAuth credentials. Once you’ve added them, contacts can be edited, blocked or deleted by clicking on the arrow next to their names. If you think this feature is more of a gimmick than a real software replacement for making video calls with the likes of Skype and Facetime, then

www.tuxradar.com

we think you'll be surprised by the results. Firefox Hello is an earnest step towards replacing desktop apps with a web-based solution. It’s feature rich and in our testing, it worked across different browsers on different OSes, and even across Android devices. The quality of the calls was impressive and none of the calls were dropped. Besides Firefox Hello, there are other changes in this release as well. The release notes highlight that the browser is now much better equipped to deal with dynamic styling changes and has implemented a new public key pinning extension which should improve the authentication of encrypted connections. The release also has several improvements for web developers, such as CSS filters that are enabled by default. LXF

Verdict Firefox 35 Developer: Mozilla Foundation Web: www.firefox.com Licence: Mozilla Public Licence

Features Performance Ease of use Documentation

9/10 9/10 9/10 9/10

A more stable video calling feature and the usual lot of security updates make this a useful update to Firefox.


Reviews Remote desktop

TigerVNC 1.4.0 Mayank Sharma discovers whether the latest release of the once celebrated VNC server deserves another chance or not. In brief... A remote desktop client and server that uses the VNC protocol. See also: Vino and Krfb/Krdc.

H

alf a decade ago in 2009, the TigerVNC server was forked from the TightVNC project with the help of Red Hat. The goal of the project was to make it possible to run graphics intensive 3D and video apps in fullscreen mode on a remote computer. Soon after its debut TigerVNC was adopted by the Fedora distro as the default VNC implementation. Over the years, the project has lost steam with some developers jumping ship. TigerVNC releases became rarer. But recently the project announced a new major release and we wanted to see how it compared with its peers. Some of TigerVNC’s marquee features, such as accelerated JPEG compression, have been borrowed from other projects, and in addition to Linux, the project also runs on Windows and Mac OS X. TigerVNC includes a barebones cross-platform viewer that doesn’t include any of the fancy features you get with other clients, such as Remmina and Vinagre. There’s no list of running VNC servers on the network or the ability to bookmark connections. Instead, you get a simple window with a text box to enter the IP address of the server you want to connect to. But that doesn’t mean it lacks settings. Click on the Option button and you get plenty of settings to configure the connection. By default, the software is configured to automatically select the best settings for your connection. This is a good thing considering that most of these settings will only make sense to a


Code spring clean

Cross-platform

As well as fixing bugs, the code has been tweaked and several existing features are improved.

Both the Windows and Mac OS X clients have also been improved for a better user experience.


The VirtualGL project has published a detailed report on this release after it discovered a serious performance regression.

seasoned network admin who's used to the ways of remote desktop. You can specify things, such as colour depth and compression level, choose an authentication scheme and encryption mechanism, and more. There’s also the Xvnc server which creates its own virtual X session for remote users to connect to. On the other hand, if you wish to connect to an existing X session, there’s also the x0vncserver server. In the latest release, this server now supports the XDamage extension for tracking changes which according to the official release notes makes it “slightly less useless”.

Toothless tiger? In fact, the latest release is full of similar minor tweaks and improvements, mostly hidden away from sight. In addition to lots of bug fixes, the developers have made behind-thescenes improvements to the keyboard handling both on the server and client, and also improved TLS implementation in the project’s Java client. The server component of the new release gains support for newer versions of the X.org server and has dropped support for the 256 colour indexed map mode. One feature that didn’t make it past the beta release is IPv6 support, which was held back by the developers from the final release because it required more work and polish.

www.linuxformat.com

In our tests TigerVNC worked on expected lines. While you can connect to the TigerVNC server with other clients, it works best and doesn’t throw any unaccepted errors when coupled with its own client. Surprisingly, TigerVNC was more responsive when viewing a remote Windows session compared to another Linux distro. Furthermore, getting it to display a remote Gnome session requires more time and effort than we’d want the average desktop user to spend. All things considered, there are no stellar features that stand out in this release. in fact when you look at it, TigerVNC doesn’t offer any unique features that’d make us recommend it over others, such as Vino or krfb. LXF

Verdict TigerVNC 1.4.0 Developer: TigerVNC Developers Web: www.tigervnc.org Licence: GPL

Features Performance Ease of use Documentation

5/10 5/10 4/10 5/10

An average release of an average app that works but doesn’t offer any compelling benefits over its peers.

Rating 5/10

Linux games Reviews

Civ: Beyond Earth Ruined one planet? Don’t worry, says Russ Pitt go where no human has gone before and colonise a universe of planets and poison them as well. In brief... A new sci-fithemed entry into the Civilization series. Head an expedition and find a home beyond Earth, explore and colonise an alien planet and create a new civilisation in space.

Specs Minimum spec: OS: Steam OS, Ubuntu 14.04 CPU: Intel Core i3 or AMD A10 Mem: 4GB Gfx: Nvidia Geforce 260 HDD: 8GB

Mind Flower. Secure the Transcendence victory by harnessing the disgusting power of the planet’s consciousness.

C

ivilization: Beyond Earth begins with the very sci-fi premise of “What if?” What if you took Civilization, the classic turn-based grand strategy game, and made one of its signature endings the beginning of a whole new game? Beyond Earth has a lot of new looks: new units, new victories, a completely new tech tree (actually, it’s a web), new leaders, new civilizations and a handful of things under the hood that are also completely new. But the experience of cracking it open, watching a colony ship settle onto a completely dark map and then setting foot onto this alien world feels just like playing Civ 5 — at first. That said, if you came to Beyond Earth looking for a direct sequel or a modern update to the 1999 game, you would be wildly disappointed. Although it does have narrative elements, and certain signature aspects of Alpha Centauri have crept in, Beyond Earth is very much its own game. This is most noticeable when dealing with the planet’s indigenous creatures. Instead of Civilization’s barbarians, Beyond Earth has a variety of alien lifeforms, some more aggressive than others. The other big new thing is the orbital layer. You can build and launch satellites, and these will impart benefits to specific tiles. Now for the bad news: It’s easy to feel like Beyond Earth is just an expansion to Civ 5, albeit spacier than those that came before. For Civ 5 fans like myself, this is a loaded proposition. If you like Civ 5, then more Civ 5 will be great, but there’s no denying that even as much as we love Civ 5, we were expecting something more than Civ 5 with a sci-fi skin.

Beyond Earth offers five victory conditions: One is Emancipation, where you build an Emancipation Gate to return home to Earth, and conquer it.

Beyond Earth’s many similarities to Civ 5 mask, to its detriment, a game that is remarkably new and different, and once we were able to see past those similarities, the newness and wonder of playing in a future Civ sandbox washed over us and we were engrossed before we realised it.

More than Civ 5 in space As the Brazilians, we were aiming for a Purity affinity, but fumbled our way through the research web willy-nilly and eventually lost the game without ever realising one of our enemies had been close to victory. So we started again, this time as the Slavic Federation. We specialised in Supremacy and after a bit of research, we decided to shoot for the Contact victory, but build a strong enough civ that, should all else fail, we could at least take over the world. Beyond Earth offers five victory conditions, although two are similar, differing only in which affinity will unlock it. Contact involves discovering an alien signal and unlocking the secret of your new planet’s ‘Progenitor’ species, an ancient alien race. Domination is what it sounds like, giving you the win if you capture all of the opposing civ’s capitals. Promised Land and Emancipation are two sides of the same coin. You must research the tech to open either an Emancipation or Exodus gate back to Earth, bringing those left behind either salvation or

www.tuxradar.com

dominance. Transcendence is the Harmony victory. It involves researching alien tech to create a ‘mind flower’ that will unite your consciousness with that of the alien planet. For our third playthrough, we wanted to win without firing shot. We picked the Franco-Iberian civ and focused on the Harmony affinity. Ultimately, war never came. Although forced to kill a handful of aggressive aliens, we dominated with science and trade, with guns silent. And when our mind flower bloomed, we felt like we’d finally understood everything Beyond Earth had to offer. This is how Beyond Earth succeeds: It offers a game steeped in the traditions and mechanics of Civilization that’s nevertheless surprising and new in often unexpected ways. LXF

Verdict Civilisation: Beyond Earth Developer: Aspyr Web: GameAgent.com Price: £30

Gameplay Graphics Longevity Value for money

9/10 8/10 8/10 8/10

Although its foundation in Civ 5 makes it familiar, Beyond Earth is full of interesting surprises.


Roundup suites Roundup Office Xxxxxxxxxxxxxxxx

Every month we compare tons of stuff so you don’t have to!

Scripting languages Richard Smedley goes beyond Bash to see which scripting languages measure up to the needs and wants of a Linux system administrator.

How we tested... Comparisons, they say, are invidious. This is certainly true for programming languages, where personality and local support are, at least, of equal import to criteria such as speed, and the level of support for different paradigms. Given this, we're presenting a mixture of facts, collective opinions and our own prejudices, but it's a basis for further investigation. The key to a scripting language's usefulness to the sysadmin lies not just in how easily it helps solve problems, but in how many of the solutions have already been written, and are available to download and adapt, and preferably well-documented. We tried to work across the range of versions installed on a typical network, but insisted on Python 3. Other than that, we've tried to stay in the context of working with what you're likely to find on your network.

E Our selection Bash Perl 5 Python Ruby NewLISP

very admin loves time-saving shortcuts, and carries a selection of scripts from job to job, as well as inheriting new ones when arriving in post. The question any new admin asks is which is the best language to learn? (Followed by, where’s the coffee?) Veterans of language wars should know that the best language question rarely has a simple or definitive answer, but we thought it would be well worth comparing the most useful choices to make your Linux life easier. Most scripting languages have been around longer than you think. For


“The question any new admin asks is which is the best language to learn?” example, NewLISP was started on a Sun-4 workstation in 1991. They've borrowed from each other, and elsewhere, and accumulated a long legacy of obsolete libraries and workarounds. Perl's Regular Expressions, for instance, are now found everywhere, and in some cases better implemented elsewhere. So what matters most? How fast the script runs,

www.linuxformat.com

or how quickly you can write it? In most cases, the latter. Once up and running, support is needed both from libraries or modules to extend the language into all areas of your work, and from a large enough community to support the language, help it keep up with trends, and even to innovate it. So, which scripting language should you learn to improve your Linux life this year?

Scripting languages Roundup

The learning curve Online resources, books and good people.

T

he key questions are: how easy is the language to pick up? Are the learning resources at least adequate? Even if these two questions are answered in the positive, they still need to be backed up by a helpful community to assist you in quickly producing something useful, and help maintain that initial enthusiasm as you hit inevitable problems. To produce a backup script and test scripts in each of the languages, we started by browsing Stack Overflow. But downloading random code means no consistency between Posix (pure Bourne Shell) scripts, modern Bash, and legacy code that occasionally fails. Fortunately, www.shellcheck.net is a great tool for checking the correctness of scripts, and teaches you best practice as it corrects them. The Linux Document Project's (perhaps overly) comprehensive Advanced Bash Scripting Guide (www.tldp.org/LDP/ abs/html) is also excellent and will help you quickly gain confidence. Perl's online and built-in documentation is legendary, but we

started by running through an exercise from the classic O'Reilly admin book, Running Linux, then leapfrogged the decades to No Starch's recent Perl OneLiners by Peteris Krumins. Those who eschew the book form should try http://perlmonks.org, a source of cumulative community wisdom. Recent efforts at getting youngsters learning through Code Club (www. codingclub.co.uk) and the rest of us through PyConUK education sprints and open data hackdays have shown Python to be easily picked up by anyone. But out-of-date advice, such as the many ways of running subprocesses which persist for compatibility reasons, means careful reading is needed, and it's yet another good reason for starting with Python 3, not Python 2. Head to www.python. org/about/gettingstarted for large list of free guides and resources. Ruby is also an easy sell to learners, and before Rails, command-line apps were what it did best. David B. Copeland's book, Build Awesome Command Line Applications in Ruby will

From MOOCs to the bookshop, Python learning resources are everywhere.

save you hours of wading through online documentation, but we were able to get up and running on our test scripts with a couple of web tutorials. Last, we come to NewLISP: a challenge to programmers schooled only in non-LISP family languages, but you'll be amazed by what it manages to accomplish with just lists, functions and symbols. We dived right in with the code snippets page on http://newlisp.org, adapting to build our backup script, and were rewarded with terse, powerful code, that was easier to read than its equally compact Perl counterpart.

Verdict Bash

+++++ NewLISP

+++++ Perl 5

+++++ Python

+++++ Ruby

+++++ Python and Ruby are easier to learn, because of good docs and helpful users.

Version and compatibility Beating the wrong version blues.

T

he question here is: have I got the right version? Lets start with Bash. Every modern Linux distro ships with a version that will run your scripts and anyone else's. Bash 4, with its associative arrays, coproc (two

parallel processes communicating), and recursive matching through globbing (using ** to expand filenames) appeared six years ago. Bash 4.2 added little, and is four years old and Bash 4.3's changes were slight.

As the Unix shell dates back decades, you will find that recent Bash versions contain a few unexpected syntax changes. www.tuxradar.com

Perl is still included in the core of most distros. The latest version is 5.20 (with 5.22 soon to appear), but many stable distros ship with 5.18. No matter, you're only missing out on tiny improvements, and just about every script you'd want to write will be fine. The switch from Python 2 to 3 still catches out the unwary. Run Python 3 if you can and check the documentation if you come unstuck. Python 3.3 is our baseline for Python 3 installs and Python 3.4 didn’t add any new syntax features. Ruby version changes have caused enough problems that painless solutions have appeared, rvm enables you to run multiple versions of Ruby, and bundle keeps track of the gems you need for each script. NewLISP's stability and lack of thirdparty scripts is an advantage here. We cann't, however, guarantee every script will run on the latest versions.

Verdict Bash

+++++ newLISP

+++++ Perl 5

+++++ Python

+++++ Ruby

+++++ Ruby's version workaround is good, but Bash's lack of problems is a better result.


Roundup Scripting languages

Web native scripts Get your admin scripts moving over HTTP.

M

uch of a sysadmin's life has migrated to the web, so you'll need a scripting language that has kept pace. We examined both ease of writing our own code, and finding available solutions for doing anything from web interfaces to system stats.

What's noticeable about these languages is the difference in expressiveness and style to produce similar results. However, this is, once again, secondary to personal preference and local support for many admins. Ruby is quick and enjoyable; Python 'feels

right' probably due to it being more human readable; newLISP is astonishingly powerful. But these observations remain partisan clichés without a supportive and maintainable environment to use and develop the code for your own networks.

Bash +++++ While Bash will be no one's first choice for a web programming language, it's good to know that when your server doesn't provide for your first choice you can fall back on it thanks to bashlib. This a shell script that makes CGI programming in the Bash shell somewhat more tolerable. Your script will be full of echo statements, interspersed with your commands to produce the desired output. Security considerations mean we wouldn't recommend running this on the open Internet, but it’s worth bearing in mind that Bash works well as a prototyping language. It's easy to fill a text file with comments describing the broad structure that you want, then fill in the gaps – testing snippets interactively and pasting into www.shellcheck.net to check your code as you go. You'll soon be up and running with a proof of concept.

newLISP +++++ Code Patterns, by NewLISP creator Lutz Mueller, is available on the www.newlisp.org website and has chapters on HTTPD and CGI, as well as TCP/IP and UDP communications. If you add in the section on controlling applications, and you’ll have everything to get you started. NewLISP's built-in networking, and simple (or lack of) syntax, makes it surprisingly easy to generate HTML pages of results from, for instance, your monitoring scripts. For a ready built framework, newLISP on Rockets – which uses Bootstrap, jQuery and SQLite – combines rapid application development with good performance. NewLISP on Rockets provides several functions, from (convert-jsonto-list) via (twitter-search) to (display-post-box), which will help you add web functionality. We're impressed but we remain concerned by the small size of the community and the intermittent pace of development.

Community support Does it have a community large enough to support real work.

D

evOps, cloud deployment, testdriven development and continuous integration – the demands on a sysadmin change and evolve, but the requirement to learn something new is constant. Everyone uses Bash to some extent but, you’ll need to learn Bash plus one other. Perl was the traditional Swiss Army chainsaw of Unix admins through the ‘80s and ‘90s, gradually losing ground to Python and then Ruby over the last decade or so. Anyone who started work


in the ‘90s or earlier will be comfortable with it, so finding someone to help with your scripts is often not a problem. However, the world doesn't stand still, and many tech businesses have standardised on Python, which is used extensively at Google, for example. Much of the software necessary for modern sysadmin work is Python based although the same can be said of Ruby. Ruby benefits from being the basis of Chef and Puppet, as well as Vagrant and Travis CI, meaning a little familiarity

www.linuxformat.com

will be helpful anywhere that uses them for deployment. The web frameworks and testing tools written in Ruby have popularised the language at many of the younger web companies. NewLISP has a much smaller community supporting it, and there aren’t many ready made solutions and you may know no-one who uses it. The keenness of the online community goes some way to ameliorate this deficiency, but you have to ask who will maintain your tools when you leave a company?

Verdict Bash

+++++ NewLISP

+++++ Perl 5

+++++ Python

+++++ Ruby

+++++ Ruby has taken mindshare, thanks to some great DevOps software.

Scripting languages Roundup Perl 5 +++++ Perl was the first web CGI scripting language and has more or less kept pace with the times. It certainly has the libraries, and enough examples to learn from, but with no dominant solution you'll have to pick carefully. Catalyst, Dancer, and Mojolicious are all good web application frameworks. More likely you'll find everything you need in CPAN. You can glue together a few of the libraries – many of which are already collected together in distros – to handle a pipeline of tasks, such as retrieving XML data, converting the data to PDF files and indexing it on a web page. Perl's traditional CGI interface is still available, and despite better performing alternatives abstracted through PSGI, you may find that use CGI; is all you need to web-enable your script, and remember: 'there's always more than one way to do it'.

Python +++++ Python's Web Server Gateway Interface (WSGI), which was defined in PEP 333, abstracts away the web server interface, while WSGI libraries deal with session management, authentication and almost any other problem you’d wish to be tackled by middleware. Python also has plenty of fullstack web frameworks, such as Django, TurboGears and Pylons. Like Rails, for some purposes you may be better off coding web functionality onto an existing script. But Python's template engines will save you from generating a mess of mixed HTML and Python. Python has many other advantages, from the Google App Engine cloud with its own Python interpreter, which works with any WSGIcompatible web application framework, for testing of scalable applications to supporting a clean style of metaprogramming.

Ruby +++++ Don't imagine for one moment that Rails is a panacea for most sysadmin problems. It's not. And while Sinatra certainly makes it easy to roll out anything web-based in Ruby, even this is overkill for most purposes. That said, Rails does a good job of getting code up quickly and just doesn't drown in all that magic, generated code. Ruby is ideal for getting any script web-enabled, thanks to gems that are written by thoughtful people who have made sane decisions. Putting a web interface on our backup script, for example, was fun, but distracting as we played with several gems, eg to export reports to Google spreadsheets. Tools like nanoc, which generate static HTML from HAML, and some of the reporting gems complement the language's expressiveness, and make adding any functionality to scripts a breeze.

Programmability Verdict

Managing big scripts requires other programming paradigms.

B

efore reaching 1,000 lines of code, Bash scripts become unmanageable. Despite its procedural nature, there are attempts to make an object-orientated (OO) Bash. We don't recommend it, we think it's better to modularise. Functional programming (FP) in Bash (http://bit. ly/BashFunsh) is also impractical. Perl's bolted on OO won't be to everyone's taste, but does the job. Perl has fully functional closures, and despite syntactical issues, can be persuaded

into FP – just don't expect it to be pretty. For that you should wait for Perl 6. Python is equally happy with imperative, OO and also manages FP. Functions are first class objects but other features are lacking, even if its list comprehension is very good. Mochi, the FP language (http://bit.ly/FPMochi), uses an interpreter written in Python 3. Ruby is designed as a pure OO language, and is perhaps the best since Smalltalk. It can also be persuaded to support a functional style of

www.tuxradar.com

programming. But to get FP code out of Ruby, you’ll have to go so far from best practices that you should be using another language entirely. This brings us neatly to NewLISP, an elegant and powerful language with all the functional features at your fingertips. NewLISP uses a pseudo OO implementation in the form of functional-object oriented programming (FOOP), but this doesn't mean, however, that it can cut it for real OO programming.

Bash

+++++ NewLISP

+++++ Perl 5

+++++ Python

+++++ Ruby

++++ Python is a multi-paradigm language and the easiest to maintain.


Roundup Scripting languages

Extending the language Libraries, modules, and getting them working.

N

one of these scripting languages are as bloated with classes as, for example, Java so that you'll need to use non-core libraries (or modules as they are sometimes called) for writing many scripts. How comprehensive these are, and how easy they are to manage with your script varies greatly.

Perl continues to impress with the mind-boggling choice to be found on CPAN, but its 'there's more than one way to do it' approach can leave you easily overwhelmed. Less obvious, is the magnitude of Bash extensions created to solve problems that are perhaps not best suited to any sh implementation.

There's more than one library for that – CPAN is a useful resource for Perl.

Python has excellent library support, with rival choices considered very carefully by the community before being included in the core language. The concern to “do the right thing” is evident in every decision, yet alternate solutions remain within easy reach. At least the full adoption of the pip package manager, with Python 3.4, has ensured parity with Ruby. RubyGems provide the gem distribution format for Ruby libraries and programs, and Bundler which manages all of the gems for dependencies and correct versions. Your only problem will be finding the best guide through Ruby's proliferation of libraries. Read around carefully. NewLisp is not a large language, but it’s an expressive one, accomplishing much without the need of add-ons. What modules and libraries that there are address key needs, such as database and web connectivity. There's enough to make NewLISP a useful language for the admin, but not in comparison to the other four choices.

Verdict Bash

+++++ NewLISP

+++++ Perl 5

+++++ Python

+++++ Ruby

+++++ The CPAN's longevity and popularity marries well with good organisation.

Network security Testing and securing the network – or fixing it afterwards.

P

enetration testing and even forensic examination after an attack will fall under the remit of the hard-pressed sysadmin in smaller organisations. There are enough ready made tools available that you can roll everything you may need into a neat shell script, kept handy for different situations, but writing packet sniffers or tools for a forensic examination of your filesystem in Bash isn’t a serious option. Perl has lost some security community mindshare since the early days of Metasploit, but the tools are still there, and are actively maintained by a large user group who aren't about to jump ship to another language. Perl has tools like pWeb – a collection of tools for web application security and vulnerability testing – which is included in distros, such as Kali and Backbox. Tools such as Wireshark are a powerful aide to inspecting packets, but sometimes you'll need to throw


together your own packet sniffer. Python not only has Scapy, the packet manipulation library, but provides a socket library for you to easily read and write packets directly. Ruby's blocks (write functions on-the-fly without naming them) and other features are great for writing asynchronous network code, and its rapid prototyping NewLISP has impressive networking features, matches (and even beats) even if it lacks the pen-testing tools of the others. Python. But Ruby's biggest boon is Metasploit, which is the most-used pen-testing software. Last, NewLISP isn't well-known In terms of ready rolled tools, you among penetration testers and grey hat can mix and match as needed, but Perl, hackers, but thanks to the networking Python and Ruby all provide everything built in to the language, a function call you need to quickly examine a network and a few arguments will create raw for weaknesses or compromises packets for pen testing. Once more, on-the-fly. Note: Python is featured in NewLISP has clear potential but suffers more security-related job adverts now. from its relatively tiny user base.

www.linuxformat.com

Verdict Bash

+++++ NewLISP

+++++ Perl 5

+++++ Python

+++++ Ruby

+++++ Python edges ahead of Ruby and Perl, but all three are friends of the pen tester.

Scripting languages Roundup Scripting Languages

The verdict W

e admit it’s difficult to take the verdict out of a practical context and just declare the best language. For example, Bash isn’t a strong language, and many time-saving bits of code can be thrown together better with the other four languages, but no-one with tasks to perform at the Linux command line should avoid learning some Bash scripting. Perl is the traditional next step as it's intimately associated with the *nix command line and still found everywhere. It may suffer in comparison with newer languages, but Perl continues to offer not just the Swiss Army Chainsaw of the Linux CLI, but Perl also has a huge and very supportive community. NewLISP is a pleasant surprise. Yes it has those – Lisp is about lists – but what a compact language for the embedded space as well as the command line. Sadly, the size of the community doesn't match the power of

1st Ruby

the language, so you'll need to be prepared to back its use with plans to maintain the code yourself. Python is a powerful, multiparadigm supporting language. The Python community is large and friendly, and supports everything from education sprints to training teachers. It also backs up community efforts, supporting young learners at Code Clubs, and many other events. But useful as a community is to the sysadmin, it's often the quick and dirty hacks, readily downloadable and reusable examples, backed by an expressiveness that makes many programming challenges if not trivial, far less of a headache. Rails brought wider attention to Ruby, but Chef, Puppet and Vagrant have helped remind the admin just what can be done with the expressive and eloquent

4th newLISP

Web: www.ruby-lang.org Licence: GPLv2 or 2-clause Version: 2.2.0 Powerful, expressive and very quick to learn.

Next i anonymsissue: distrosing

+++++

Web: www.newlisp.org Licence: GPL Version: 10.6.1 So powerful, it deserves to get more use.

5th Bash

+++++

Web: www.python.org Licence: PSFL Version: 3.4.2

+++++

Web: www.gnu.org/software/bash Licence: GPLv3+ Version: 4.3.30

Multi-paradigm, encourages good practices and great community.

3rd Perl 5

scripting language that was developed by Yukihiro Matsumoto. Does Ruby edge out Python? Is Bash to be ignored? Not for the admin: as they need good knowledge of Bash to follow what’s going on with the system. And in addition to Bash, every sysadmin should know a little Perl, Python and Ruby, but have in-depth knowledge of the one that they prefer. LXF

“In addition to Bash, every Linux admin should know a little Perl, Python and Ruby.”

+++++

2nd Python

We can't help acknowledging Ruby's power and charms.

Doesn't do everything, yet remains essential.

Over to you...

+++++

Web: perl.org Licence: GPL or Artistic License Version: 5.20

We don't want to start a holy language war, but we would love to hear what you use. Email your opinions to [email protected].

Still a great friend to the sysadmin.

Also consider... While Bash falls down in some areas, traditional shell scripting could also be represented by Zsh, which has some small but useful differences, such as better access to positional variables, and the ability to extend the shell through widget functions. Nevertheless, it's not a rival to our other choices, and nor is PHP, despite its use in

command scripts by some devotees. Instead, our left-field alternative is Rebol (Relative Expression Based Object Language), whose leap towards software freedom two years ago may have come too late to ensure universal popularity. However, Rebol has an elegant design and syntax, and a useful read–eval– print loop (REPL) console.

www.tuxradar.com

Rebol's 'dialecting' (using small, efficient, domain languages for code, data and metadata) equips it for just about anything. It’s particularly good at dealing with the exchange and interpretation of information between distributed computer systems, but also powerful, terse shell scripts. If you're looking for a new language for 2015 give it a try.


Reviews Raspberry Pi 2

Raspberry Pi 2 Les Pounder salivates at the prospect of a new Pi and promptly breaks his teeth on the sweet, new raspberry-flavoured treat.

In brief... The latest single board PC from the Raspberry Pi Foundation comes with the spec boost that we were all hoping for. The Pi 2 is the latest in a line of products from the Foundation and can run a number of Linux distros.

Specs SoC: Broadcom 2836 CPU: Quad-core ARM7 800MHz GPU: Videocore IV 250MHz Mem: 1GB GPIO: 40-pin Ports: 4x USB 2.0, 100BaseT Ethernet, HDMI, MicroSD card Size: 85.60 × 56.5mm

W

hen the Raspberry Pi appeared in 2012 few could have envisaged how popular it would be. In the years after its release the Raspberry Pi has become the most popular single-board computer on the market and spawned many imitators, but none with the rich community that has grown organically around the Raspberry Pi. Since the release of the original Raspberry Pi there have been three versions of the flagship B model, starting at 256MB RAM and increasing to 512MB with the second B and B+. But in all of these models the system on a chip (SoC) has remained the trusty BCM2835 with an ARM 11 700MHz CPU. The community have done wonderful things with these resources but now the specification boost that they were waiting for has arrived. In early February, the Raspberry Pi 2 arrived and the original ARM 11 has been replaced with an ARM 7 CPU running at an improved 800MHz. But rather than stick with a single core, the new verision comes with four cores which speeds up the Raspberry Pi by as much as six times. To go with the new CPU, the amount of RAM has also been upgraded to 1GB. The rest of the hardware, however, matches that of the


B+: a 40-pin GPIO, four USB 2 ports and 10/100 Ethernet. Physically the Raspberry Pi 2 also has the same dimensions as the B+.

On the testing bench To show the improvements made to the Pi part deux, we wanted to run a few real-world benchmarks to show how powerful the new Pi actually is when

Creating a new world in Minecraft took 42 seconds on the B+, and 21 seconds on the Pi 2. Loading IDLE 3 took 13 seconds on the B+ and a mere 4 seconds on the Pi 2. Running SunSpider in the new optimised browser gave a glimpse at real-world performance. Over the suite of tests there was a 2.5 times boost in speed. Considering the complexities of multi-threading this sounds like a reasonable expectation. Even so, individual results showed a near four-fold increase on this unoptimised code. The Raspberry Pi B+ and Pi 2 both come with the same Videocore GPU as before and in our tests there was a small improvement in FPS (Frames Per Second) for the Pi 2, largely thanks to the increased RAM present on the board. Our last test was file transfer speeds via Ethernet, for this we used scp to copy a 692MB Big Buck Bunny video file to each Pi. On the B+ we saw an average of 3.8MB/s and on the Pi 2

“Booting from cold: The B+ managed it in 33 vs 17 secs for the Pi 2.” directly compared to the B+. The first test on our list is booting both Pis from cold to login prompt. The B+ managed this is in 33 seconds versus 17 seconds for the Raspberry Pi 2. We then set both Pis to boot straight to desktop and the B+ managed 42 seconds while the Pi 2 came in at 21 seconds – half the time of the B+! Once at the desktop we tested a few common applications.

www.linuxformat.com

Raspberry Pi 2 Reviews

SunSpider Benchmarks Test

Pi 2

B+

Times faster

Total

2760.9

8178

2.96

3d

550.9

1427.8

2.59

cube

157.3

473.6

3.01

167

296

1.77

raytrace

226.6

658.2

2.90

morph

The form factor may be the same as the B+, but the Pi 2 packs a punch.

we saw 4.6MB/s, which is an 0.8MB speed increase. The Raspberry Pi Foundation have released an updated Raspbian image which includes the ARM v7 kernel image necessary to use the new CPU. Applications written for the original Raspberry Pi are fully compatible with the Raspberry Pi 2 , though – building

upon the rich projects that have been written since the initial launch of the Raspberry Pi. The Raspberry Pi 2 fulfills a lot of the requests made by the community and provides a stable and well-supported platform for hackers, makers and learners to carry on with excellent projects for many years to come. LXF

access

211.9

435.9

2.06

binary-trees

27.6

69.8

2.53

fannkuch

101.5

190.1

1.87

nbody

52.8

118.7

2.25

nsieve

30

57.3

1.91

bitops

113.8

206.1

1.81

bits-in-byte

22

35.6

1.62

bitwise-and

29.1

48.2

1.66

nsieve-bits

52.8

104.1

1.97

controlflow

28.3

64.6

2.28 2.28

recursive

28.3

64.6

crypto

221.4

578.6

2.61

aes

112.4

287.6

2.56

md5

60.1

162.2

2.70

sha1

48.9

128.8

2.63

date

336.3

1269.9

3.78

format-tofte

171.5

641.9

3.74

format-xparb

164.8

628

3.81

math

158.4

394.5

2.49

cordic

43.3

99.9

2.31

partial-sums

78.7

215.7

2.74

spectral-norm

36.4

78.9

2.17

regexp

101.9

160.6

1.58

string

1038

3640

3.51 2.82

base64

63.3

178.8

fasta

156.9

409.7

2.61

tagcloud

177.8

617.7

3.47

unpack-code

514.5

2021.6

3.93

validate-input

125.5

412.2

3.28

74.68

509.58

6.8

Sysbench Prime


Verdict Raspberry Pi 2 Developer: Raspberry Pi Foundation Web: www.raspberrypi.org Price: £20


Powerful 4-core ARM v7 processor

A great new Raspbian UI

The new Broadcom BCM2836 ARM v7 quadcore processor with 1GB of RAM yields results (see the benchmarks, above) that are four times the performance of the old BCM2835 SoC.

Available since December, the new sleek Raspbian desktop runs well on the B+, but on the Pi 2 it feels like a responsive desktop that we’d expect to see on our main computers.

www.tuxradar.com

9/10 10/10 10/10 10/10

An almost perfect single-board computer that marries great hardware – that’s backward compatible – with a lively and supportive community.

Rating10/10 March 2015 LXF195 27

Raspberry Pi 2

Hands-on with the

Raspberry Pi 2 Les Pounder has early access to the first batch of Raspberry Pi 2s and takes one for a test drive. The new Raspberry Pi comes with much more power than it’s predecessors. This is thanks to the improved ARM 7 processor running four cores at 800MHz each and the generous 1GB of RAM. This increase in both CPU and RAM is a massive benefit to projects that require pure CPU grunt, such as OpenCV and Minecraft. The Raspberry Pi 2 also benefits from the design improvements made for the B+ with more USB ports thanks to the LAN9514 providing four ports over the 9512’s two. The B+ also introduced better power management and this, again, is also present in the Raspberry Pi 2. “Power consumption while performing a given task is comparable to that

of B+,” explains Eben Upton, CEO of Raspberry Pi Trading. “Obviously if you push Pi 2 hard it will consume more power as it’s doing more work. Power consumption of B+ under heavy load is roughly the same as the old Model B.” So now that we have the latest Raspberry

The easiest method to setup your microSD card is to use NOOBS (New Out Of The Box Software) To use it you will need at least an 8GB microSD card. Download NOOBS as an archive from the Raspberry Pi website and extract the contents to your 8GB microSD card which should be formatted to use a FAT32 filesystem. With NOOBS copied to your microSD card, unmount and remove the card from your computer and insert it into your Raspberry Pi 2, you should hear a gentle click once it is firmly in place. Connect up your Raspberry Pi to a monitor via the HDMI port and then attach a keyboard and mouse via the USB ports. You will also need to ensure that you have Internet access for your Pi. The easiest method is to use the

“The Raspberry Pi 2 also benefits from the design improvements made for the B+.”


Pi, lets take it for a test drive! And for this extended tutorial we will be using the latest version of Raspbian available via the Raspberry Pi website (www.raspberrypi.org/ downloads)as it comes with the kernel7.img necessary to use the ARM7 CPU.

www.linuxformat.com

Raspberry Pi 2

Ethernet port. Finally, attach the power supply to the micro USB port. Your Raspberry Pi 2 will now boot for the first time. On first boot NOOBS will ask you which OS to install, in this case we require Raspbian. Select the OS and start the installation, which will take around 10 minutes. With the install complete the Pi will reboot and start Raspbian for the first time, and the first thing that you will notice is how quick the boot process is now at 17 seconds versus 33 seconds for the B+! For the observant you’ll also note that there are now four raspberries at boot up. This denotes that the Raspberry Pi 2 has four cores, a rather nice little Easter egg that harks back to the old boot screens of Linux distributions. Once booted we are immediately presented with the raspi-config menu, this is a tool to further configure your Raspberry Pi. At this stage we will simply exit out of the menu and login as normal. The standard login details have not been changed and remain as: USERNAME: pi PASSWORD: raspberry Once logged in you need to type startx to load the desktop. You’ll see that the desktop is a little different to previous versions of Raspbian, this is due to extensive changes that were made by the Foundation back in December of 2014 and is largely the work of Simon Long, who used to work for Broadcom. In fact, it was Long who hired Eben Upton at Broadcom, but now Simon Long has joined the Raspberry Pi Foundation to work on the user interface and his first project was to develop a new desktop. The Raspberry Pi Foundation have created a very powerful single board computer, but how can we test that power? It is somewhat fitting that we can calculate Pi to as many decimal places as we wish, so lets say 10,000? To do that we need to install some software, first. Open LXterminal and type the following two lines: sudo apt-get update sudo apt-get install bc We’ve just installed a precision calculator that we can use from the terminal. So now to the test, calculating Pi to 10,000 places and timing the activity. time echo “scale=10000; 4*a(1)” | bc -l In our test it took 17 minutes 25.725s to calculate Pi to ten thousand decimal places on our stock Raspberry Pi 2. We repeated the same Pi test on a Raspberry Pi B+ and it took significantly longer at 25 minutes 5.989 seconds to perform the calculation. As you can see straight away that’s a very clear indication that the processor of the new Raspberry Pi 2 is easily much more powerful than previous models we’ve seen.

Computing Pi to 10,000 places is a fitting test for our Raspberry Pi. The command to run the test is run from LXTerminal. Here we show the time it took before we overclocked.

Our quick test, gives us a good idea of how much of a serious performance beast the Raspberry Pi 2 is out of the box, but can we tweak it into more of an animal? Well, earlier we dismissed the raspi-config menu but for this next step we need it again. In LXTerminal, and type the following: sudo raspi-config

Turbo charge your Pi Our first post install configuration is to review the memory split. This is how the memory is divided between the GPU (Graphical Processing Unit) and the main system. On the Raspberry Pi a typical setup would be around 64MB of RAM for the GPU and the remaining RAM would be allocated to the system. This is an optional config tweak but you could just leave it as is. You can easily tinker with these values, and a general rule is that a terminal will not need as much RAM as the full desktop, so for a terminal-only project you can easily get away with 6MB allocated to the GPU. For desktop applications such as Minecraft a minimum of 64MB is needed. You will be prompted to reboot your Raspberry Pi, do this and you will be returned to the login prompt. With the changes made to the memory split now let us return to the raspi-config main menu and head to the Overclock menu. Your Raspberry Pi 2 already runs at 800MHz per core, which is an improvement over the original 700MHz single core ARM 11 CPU. We spoke to Eben Upton and Gordon Hollingsworth about the new CPU and they both confirmed that it can be overclocked to around 1.1GHz per

Quick tip The default web browser for Raspbian, Midori has recently been replaced with Epiphany which has been optimised for use on the Raspberry Pi. The new browser is available via the latest Raspbian update and works really well on Raspberry Pi 2 and older Pis.

Ubuntu on Pi? The biggest surprise brought about by the new Raspberry Pi 2 is compatibility with Ubuntu for ARM 7 CPU. Before the original Raspberry Pi was released in early 2012, Ubuntu was often mentioned as a candidate for the Pi, but as Canonical didn’t support the ARM 6 architecture, which the ARM 11 CPU used in the original Pi, another distro was needed. Early on we saw Pidora, a fork of Fedora for the Pi, being used to demonstrate the power of the Pi. But Pidora

brought a full desktop experience to hardware that required a lighter distribution. After further investigations Debian, in the form of Raspbian, was seen as a suitable candidate and this remains the official distro and is used for all official tutorials and supporting documentation. But compatibility with Ubuntu doesn’t mean that the Foundation is abandoning Raspbian: “We’re not planning an official Ubuntu image,” says Eben Upon. “We’re going to benchmark

www.tuxradar.com

‘regular’ armhf Debian against Raspbian, and may switch if we see a compelling performance advantage. Our preference is to stick with Raspbian, perhaps with a few key libraries swapped out dynamically, as this will allow us to support Raspberry Pi 2 and Classic from the same image.” At the time of writing there are no ready made images for Ubuntu on Pi, but over the coming months there are sure to be many on offer for you to try. including Ubuntu Core.


Raspberry Pi 2

Quick tip Watching YouTube videos is now possible thanks to Youtube_dl. Normally YouTube videos are Flash based but there are no Flash packages for Raspbian. When watching a YouTube video in the web browser, Youtube_dl substitutes the Flash element of the web page with an HTML5 compliant video.

core. We won’t be going quite that high, but we will overclock our Raspberry Pi to a stable 900MHz using the Overclock menu. While this is a relatively safe activity it should be noted that going too far with overclocking can severely damage the CPU due to the excessive heat generated by it working harder. We asked the Raspberry Pi team and it confirmed that the core can reach temperatures of 85 degrees, and when it does it will automatically turn off the Raspberry Pi for protection. For ‘extreme’ tinkerers who want to push their Raspberry Pi 2 to the limit, now might be the time to invest in a set of heat sinks. If at anytime you wish to return the CPU to it’s normal speed, re-enter the raspi-config menu and return the values to the stock 800MHz. With our configuration changes made, and after a few reboots we have successfully ‘turbo charged’ our new Raspberry Pi. Let’s start the graphical user interface. After logging back in use the LXTerminal to type startx to return to the desktop. Now, lets see how the changes have improved our Pi to 10,000 score by repeating the test. Open LXTerminal and repeat the test code which was time echo “scale=10000; 4*a(1)” | bc -l The code will run and in our test it took 15 minutes 28.519 seconds – that’s an improvement of two minutes! The Raspberry Pi Foundation has taken great care to build upon the legacy created by the Raspberry Pi Classic: “Raspberry Pi 2 has been in development for a couple of

The raspi-config advanced menu contains configuration options that enable you to really make the Pi your own.

The Hello_Pi demos are a great way to show off your Raspberry Pi 2. You can wrap video around the teapot using any H.264 compliant video.

years,” say Upton and Hollingsworth, and that includes the time spent developing BCM2836. “The first silicon arrived at the start of last May; there will be a video on the blog of me, James and Dom in the lab at Broadcom at 1am, the day after silicon came back, with the ‘video on a teapot’ demo running from Linux on a Broadcom “Ray” dev board. The design of the Rasberry Pi 2 board started last August (2014), and we’ve had samples since October (2014). We went through three iterations of prototypes to get exactly the right performance,” says Upton.

Compatibility The performance is reflected in the choice of CPU for the Raspberry Pi 2. Rather than choose another architecture the Foundation has stuck with an ARM-based CPU that is compatible with the ARM11 found in the earlier Raspberry Pi. The quad-core ARM7 can run software written for the older Raspberry Pi: “Raspbian works out of the box, but requires a new v7 kernel, which will be included in the downloads from our website” said Eben. As for hardware compatibility, the Raspberry Pi 2 shares the same GPIO as the B+, which means that boards made for the A+ and B+ will also work with the Raspberry Pi 2 and this even includes HAT boards (Hardware Attached on Top), which contain an onboard chip that communicates with the Raspberry Pi to set up the board quickly. There are some boards, however, that are not compatible with the B+ and the Raspberry Pi 2 due to their size and design. Boards such as PiFace [see LXF180] and PiFace Control and Display – that was used to control a camera rig for the Royal Institution Christmas Lectures – cannot be

The making of Raspberry Pi 2 The Raspberry Pi Foundation are rather excited about the new Raspberry Pi 2, We spoke to the engineering team and Gordon Hollingsworth about development: “The Raspberry Pi 2 is 100% awesome. It’s as close to a typical PC as we wanted when we first started the project.” And the amount of effort that’s gone into development is impressive: “The team have put the equivalent of 20 years of work into the new Raspberry Pi and it’s processor, and that runs at


a cost of between £2-3 million.” But there’s still a lot of enthusiasm for the older Raspberry Pi, Eben Upton explains: “There are a lot of industrial customers who won’t want to switch, and, of course, we still have the Model A+. To give you an idea of the ‘stickiness’ of an old platform, we sold something like 80,000 Model Bs after the Model B+ launch.” The Foundation also has its Compute Module, which was created to embed the

www.linuxformat.com

Raspberry Pi inside industrial applications. We asked Eben if the Compute would be seeing a similar upgrade or not: “We’ll do a Compute Module 2 at some point, but probably not in the first half of 2015.” And what of the A+? Would there be a similar upgrade for it: “Nothing is currently planned as the A+ price point is quite challenging.” No upgrade for now then, but the Raspberry Pi family has grown considerably since 2012 and now features six devices.

Raspberry Pi 2

The raspi-config overclock menu comes with the required warning that pushing too far will break your Pi.

attached. But the OpenLX SP team behind these boards has released special versions for the B+ and Raspberry Pi 2.

3D graphics test Every Raspberry Pi comes with the same VideoCore IV GPU (Graphical Processing Unit) that enables the Raspberry Pi to play back high-definition video at 1080p. The new PI also comes with this GPU, also made by Broadcom just like the BCM2836 powering the new Pi. Did you know that there’s a test suite for the GPU? You can find the test suite by opening LXTerminal and typing the following: cd /opt/vc/src/hello_pi/ In there you will find a number of directories containing many different demos. But before we can use them we need to build the demos from source, and to make this easier the Foundation have provided an automated build script that will only need to run once. To run the script, in LXTerminal type ./rebuild.sh This build script will run the build process for all of the demos so it may take a few minutes, even on our new speedy Raspberry Pi. Once completed there are a number of demos that you can try out and the first on the list should be hello_teapot. To run it, in LXTerminal make sure that you are still in the hello_pi directory and type: cd hello_teapot ./hello_teapot.bin You will now see a 3D render of a teapot with video that’s been directly rendered on to its surface. To exit out of the teapot demo hold Control+C together and you will be returned to LXTerminal.

Another demo to try is hello_triangle2 and to getto that you will need to go back to the hello_pi directory and we can do that by typing. cd .. From hello_pi we can change our directory to hello_ triangle2 and run the demo by typing cd hello_triangle2 ./hello_triangle2 This demo appears to be rather static at first, but try moving the mouse around and you will see two fractals superimposed one over the other moving and reacting to your mouse movements. Apparently, you can also control the fractals to create a perfect circle. To exit out of the hello_ triangel2 demo hold Control+ C together and you’ll be returned to LXTerminal So we’ve taken a look around the new Raspberry Pi 2, seen how it performs out of the box and taken time to supercharge our Pi. On the next page we will interface Minecraft with Pibrella to create a pushbutton bomb deployment system!

Quick tip The Raspberry Pi 2 shares the same dimensions as the B+ but for those of you looking to reuse a B+ case, such as the Pibow, it’s worth noting that some surface mount components have been moved. These changes don’t affect the overall size of the board but as the Pibow uses layers to build, a new layer will be required for your Pibow.

Updating your Pi The Raspberry Pi Foundation has released many new updates to the existing Raspbian install and keeping your Raspberry Pi up to date is a really good practice. There are a few handy Terminal commands to help you, such as sudo apt-get update which will update the list of installable software. Using the following sudo apt-get upgrade

will check for the latest software, while sudo apt-get dist-upgrade is similar to upgrade, but a more intelligent updater that removes old kernels. If you would like to install the new desktop, and who wouldn’t as it is a really great piece of work by Simon Long, type the following three lines into a terminal: sudo apt-get update

www.tuxradar.com

sudo apt-get dist-upgrade sudo apt-get install raspberrypi-ui-mods Once completed, reboot your Raspberry Pi for the changes to fully take effect. Once logged back in you will be greeted with the new and improved interface. Updating is done via the terminal using the very powerful apt package management system that’s used by all Debianbased distributions.


Raspberry Pi 2

Link a Pibrella to Minecraft 1

Attach Pibrella to your Raspberry Pi

We’ll use the big red button on the Pibrella.com to set off TNT in Minecraft. The Pibrella fits over the first 26 GPIO pins. Never attach while the power is on! Use a blob of blu tack or modelling clay to prevent the underside shorting out the HDMI. Now connect the other cables as usual, but insert the power into the micro USB port.

3

sudo idle

This opens idle, a Python editor, with elevated privileges enabling us to use Pibrella with Python. Now open the example code.


This installs the software that we need to use Pibrella with Python.

Examining the code

Our code is written in Python 2 as the Minecraft module is currently only available for that version. The code is fairly easy to read. Note the line starting with # are comments. The first few lines are imports, they import extra functionality, in the form of Pibrella and Minecraft libraries, for our project. After that we use a variable called mc to store connection details to Minecraft.

6

In position

Minecraft uses x,y,z coordinates to know the position of objects in the world. We create a function called button_changed(), which locates the player and then creates a cube of TNT at coordinates near to the player. Lastly we set the function to be called when the button is pressed. Keep the window open and open Minecraft and create a new world.

Setup Pibrella

sudo apt-get update sudo apt-get install python-pip sudo pip install pibrella

4

Get the code

We’ve created a GitHub repository that contains the code for this tutorial, visit https://github.com/lesp/Pibrella-Minecraft-TNT and download a copy. Next open LXTerminal and type

5

2

With our Pi booted to the desktop, open LXTerminal and type:

Drop the bomb

With Minecraft ready and our code open, press TAB to release the mouse from Minecraft and click on Run > Run Module in idle. The idle shell will come to life and run the code. Switch back to Minecraft and go to a nice spot. Press the red Pibrella button to drop the bomb. Hit the TNT with your sword... and then RUN! Note: You can run this the original Pi, but it could crash Minecraft. LXF

www.linuxformat.com

TECHNOLOGY, TESTED

NE W TECH WE AR ABLES CHANNEL!

The best just got better! NOW LIV E !

Still showcasing the best news, reviews and features but now optimised for every screen and your satisfaction. Love tech? Then visit...

www.techradar.com twitter.com/techradar

facebook.com/techradar

Subscribe to


Read what matters to you when and where you want. Whether you want Linux Format delivered to your door, your device, or both each month, we have three great options to choose from.*

Choose your

package today! #

1 for Free Software

Print

Digital

Now o Androidn!

Every issue delivered to your door with a 4GB DVD packed full of the hottest distros, app, games, and more. PLUS exclusive access to the Linux Format subscribers-only area.

Instant access to the digital editions of the magazine on your iPad, iPhone and Android* devices. PLUS exclusive access to the Linux Format subscribers-only area, featuring complete issues & disc downloads.

ONLY £31.99

ONLY £20.49

Your subscription will then continue at £31.99 every 6 months – SAVING 17% on the shop price.

Your subscription will then continue at £20.49 every 6 months – SAVING up to 37% on the shop price.

*Only available in certain territories: http://bit.ly/LXFwhere


www.linuxformat.com

Get the complete Print + Digital


package

Now o Androidn!

A DVD packed with the best new distros and free & open source software every issue. Exclusive access to the Linux Format archive – with 1,000s of DRM-free tutorials, features, and reviews. Every new issue of the magazine in print and on iPad, iPhone, and Android* devices.

! E U L A V T BES

Never miss an issue, with delivery to your door and straight to your device. Huge savings, the best value for money, and a money-back guarantee.

ONLY £38.49 Your subscription will then continue at £38.49 every 6 months – SAVING 17% on the shop price and giving you up to a 78% discount on a digital subscription.

Two easy ways to subscribe... Online: myfavouritemagazines.co.uk/LINsubs

Or call 0844 848 2852 (please quote PRINT15, DIGITAL15, BUNDLE15) Prices and savings quoted are compared to buying full-priced UK print and digital issues. You will receive 13 issues in a year. If you are dissatisfied in any way you can write to us or call us to cancel your subscription at any time and we will refund you for all undelivered issues. Prices correct at point of print and subject to change. For full terms and conditions please visit: myfavm.ag/magterms. Offer ends 19/03/2015

www.tuxradar.com


Get into Linux

Get into

Linux Curious about Linux but not sure how to traverse this unfamiliar territory? Mayank Sharma will help you get started so you can join the growing ranks of Linux users. here’s never been a better time to get into Linux. It’s slicker than ever, easier to install than ever and all the big-name distros like Ubuntu, have just received updates. So Linux Format thinks it’s time for a fresh start with a fresh install. If you’ve never touched Linux before never fear, we’ve got a complete guide to started, while the cover disc has Fedora 21 and Ubuntu 14.10 updated and patched. These are “Live” versions

T

mainstream OSes are well worth this extra step. For starters, Linux is open source, which is to say that you can legally download a copy of Linux and install it on all your computers. It also ships with a ton of software and you can download more with a single click. Unlike proprietary OSes, a Linux flavour (or distro) is very malleable. You can swap out any of the default apps or even its entire interface and replace it with something more suitable to your needs. Choice is another hallmark of Linux with multiple options from simple components to complex suites. Furthermore, besides being compatible with all your newer hardware, Linux can also turbocharge hardware that’s past its glory days. To top it all, you can do everything you can on a Windows box. From streaming video to playing the latest games, a Linux desktop will work just as well for you as any other system.

“To top it all, you can do everything you can on a Windows box.” so you can boot them straight off the DVD and try them out instantly. One of the biggest impediments to widespread Linux adoption is that you don’t get Linux on a PC from Currys – getting Linux onto a PC is a more involved process. But the advantages a Linux desktop offers over the other


www.linuxformat.com

Ubuntu

ON THE DISC

The most commonly known Linux distribution (often abbreviated to distro), Ubuntu pays special attention to desktop usability. In the 10 years of its existence, Ubuntu has galvanised the development of Linux on the desktop and is the go-to distro for third-party developers and vendors who want to run their wares on Linux.

Fedora

ON THE DISC

Red Hat’s open source offering to the world, Fedora is known for adapting and offering new technologies and software to its users. Over the years, the distro has managed to find a clever balance between offering new features and stability to its users, which makes it popular with both new and experienced Linux campaigners.

Mageia Although it’s just had four releases to date, Mageia has a pedigree in usability that dates back to the 1990s. Mageia is a community project that’s supported by a non-profit organisation, which is managed by a board of elected contributors. The distro is known for its customised user-friendly tools for managing the installation.

Get into Linux

All you need to know to anchor Linux on your computer. he Linux installation process is involved but it isn’t actually that cumbersome. The exact installation steps are slightly different for every distribution, but in general the distro’s graphical installer will guide you through the necessary steps pretty easily. In essence, installing Linux is very similar to installing a piece of software, albeit with a few caveats: Disk partitioning Unlike a piece of software, installing Linux requires you to create a dedicated partition on your hard disk. This isn’t an issue if Linux will be the only OS on this computer. However, if you’re installing Linux alongside another OS, such as Windows, you’ll have to take steps to preserve the existing data. Many Linux distros will offer to partition the disk for you automatically, though you can create partitions yourself with ease from within Windows using the Disk Management tool (see Make Room for Linux, below). The advantage of manually partitioning your disk is that you get to decide how much space to allocate to Linux. When creating partitions remember to create two new partitions. The bigger one with at least 12GB of disk space is for the OS itself, which you’ll format as ext4. You’ll also need to create a second partition for what’s called swap space. In simple terms, the swap partition extends the amount of physical RAM on your computer. A general rule of thumb for computers with a small amount of RAM (one or two gigabytes) is to create a swap partition that’s twice as large as the amount of RAM on your computer. For computers with more RAM, it’s best to create a swap partition that’s the same size as the amount of RAM you have. Securing data During the installation process, many distros including Fedora and Ubuntu will give you an option to encrypt the Linux partition. This option gives you an added layer of security by insulating your data from unauthorised access. To enable this option you will need to supply a passphrase which will then act as the key to unlock the data.

Another important step during installation is setting up a root account. On most distros this step is part of the user creation process where you define the login credentials of your regular user account. The regular user doesn’t have any permissions to modify the system while logging in as root gives you complete control over your system. Dual boot One software you should be familiar with when installing Linux is the bootloader. It’s a small program that tells the computer where to find the different OSes on the disk. Most Linux distros use the Grub 2 bootloader. In general, you shouldn’t have to do anything here, even when installing Linux on a Windows 8 computer that uses the UEFI BIOS with Secure Boot enabled. The latest versions of most mainstream distros, including Ubuntu and Fedora install a UEFI-compatible bootloader that will work correctly out of the box. However, since different vendors implemented UEFI differently, you might not get to the Grub bootloader screen and instead end up booting straight into Windows after installing Linux. In such a case, you should consider enabling the Legacy BIOS mode wherein the UEFI firmware functions as a standard BIOS. The option to enable Legacy BIOS is under the UEFI settings screen. Testing before installation Almost every mainstream distro, including Ubuntu, Fedora and Mageia allow you to boot into a ‘live’ environment, which lets you experience the distro without disturbing the contents of your hard disk. You can use the live environment to get familiar with the distro and also verify the compatibility of your hardware with the distro. Also note that Linux distributions are distributed as ISO images. You can burn them to a CD or DVD, depending on their size, using the option to burn ISO images. You can also transfer ISO images to a USB drive. There are tools, such as UNetbootin and Yumi that will create bootable USB drives with the ISO of your distro, while Mageia recommends using the Rufus utility.

Make room for Linux: Resize a Windows partition

1

Shrink Windows

Before you can partition your disk you’ll need to squeeze your Windows partition to free up some disk space for the new partition. Head to the Disk Management tool, and right-click your main partition that is typically assigned the drive letter C. Then select the Shrink Volume option from the pop-up menu.

2

Create new partition

The Shrink dialog box shows you the total size of the drive and the maximum amount of shrink space available. You cannot squeeze more space out of a drive than the size shown here. To create a new partition, specify the size of the partition in the space provided in MB and click Shrink to start the process.

www.tuxradar.com

3

Use the partition

After the process is complete, a new partition showing the amount of free, or unallocated, space appears next to the Windows C: drive partition. You can then point your Linux distro’s installer to this free space. Remember to repeat the process and create another partition for the swap space as well.


Get into Linux

Taking baby steps with Linux. ne of the biggest difference that foxes users coming from proprietary OSes is the lack of a consistent ‘look and feel’ to the Linux desktop. The default desktops on Ubuntu, Fedora and Mageia distros all look and behave differently from each other. We say default because unlike other proprietary OSes, a Linux distro enables you to swap out the desktop and replace it with an entirely different one to better suit your workflow. The idea of the desktop as a separate entity from the operating system sounds foreign to users coming from Windows or Mac. But like all things Linux and open source, users are spoilt for choice when it comes to putting a face on top of their Linux distro.

Unity in diversity The Ubuntu distribution uses its own home-brewed Unity desktop. The most prominent component on the desktop is the vertical Launcher which functions pretty much like a taskbar. It houses icons for the frequently used apps for quick access that you can modify as per your requirements. Also, some icons have specialised right-click context menus that give you quick access to frequently used features.

The first icon on the Launcher brings up the Dash, which is Ubuntu’s take on the traditional menu-based navigation system. It features a search box at the bottom and anything you type here is used to look for matching apps, documents, music, videos, instant messages, contacts and other content. Furthermore, you can also use the Dash to install and uninstall apps and preview media files. Unity also includes the Heads Up Display (HUD), which is an innovative take on the application menus. Using HUD helps you avoid the trouble of looking for options embedded deep within nested menus. To access HUD press the Alt key from inside any app and use the Dash-like search box to perform a task. The default Unity experience is the result of extensive usability research by Canonical. But you’ll find some options to tweak the desktop from under the System Settings tool accessible via the gear & spanner icon in the Launcher. The settings are grouped into three broad categories. The Personal group houses settings for customising the look and feel of the desktop by changing the wallpaper and modifying the behaviour of the launcher. Pay attention to the Online Accounts settings which you can use to sign into several online services, such as Facebook and Google Docs, and integrate their contents with the desktop apps. For example, adding your Flickr account will integrate it with the Shotwell photo manager.

Gnome thyself Gnome is another popular desktop, and the Gnome 3 desktop contains more or less the same elements as Ubuntu’s Unity but presents them in a different way. For starters the desktop is very bare. Click on the Activities button in the top-left corner to reveal the Overview which is very similar to Unity’s Dash. In this view, you also get a Launcherlike Favourites bar for accessing frequently used apps. In the centre you get a preview of all open windows. To the right is the Workspace Switcher, which always shows the current Workspace and an additional one. If you add windows

Users of Gnome-based distros should use the Gnome Tweak Tool to tweak the behaviour of their desktop.

Resuscitate an old workhorse One Linux speciality is infusing life into machines doubling up as paperweights because they can’t keep up with the hardware demands of modern OSes. While there are many distros that are designed to power older hardware, our all-time favourite is Puppy Linux. The distro uses one of the lightest window managers (JWM) and though it might not be pretty to look at, it’ll turn that old lethargic work horse into a galloping stallion. But the main reason for Puppy’s stellar performance on hardware with limited resources is its sheer number of lightweight custom apps. The distro has graphics apps, productivity apps,


apps to playback, edit and even create multimedia. Using its custom apps, you can block website ads, grab podcasts, do internet telephony, burn optical media, and a lot more. The distro is available in different flavours. The Wary Puppy edition uses an older kernel and includes additional drivers to support peripherals like dial-up modems. There are flavours based on the recent Ubuntu releases too, such as Tahrpup based on Ubuntu 14.04 and Slacko Puppy based on Slackware Linux. These editions use a newer kernel than Wary but take advantage of Puppy’s custom apps for limited resources.

www.linuxformat.com

Puppy Linux has very helpful forum boards and loads of documentation written specifically for new users.

Get into Linux

to the second Workspace, a third will automatically be added. At the top is a search box that will match any text to apps and documents on the local computer as well as online services. Gnome includes an Online Accounts app that enables you to sign into online services, such as Google Docs and Flickr. In fact, Fedora will ask you to sign into these online services when you boot into the distro for the first time. The Gnome desktop also has several custom apps of its own that can fetch information and data from the added online accounts. For example, the Gnome Contacts apps can pull in contacts from various online sources, such as Gmail. Similarly, Gnome Documents will help you find documents from online repositories such as Google Docs. New users should also keep an eye out for the desktop’s peculiarities. For one, you won’t find the Minimise buttons on any of the windows in Gnome. When you want to switch to another app, head to the Activities Overview and launch a new window or select an existing open one. Another esoteric aspect is the lack of any desktop icons as well as the ability to create any shortcuts or place any folders on the desktop. However, Gnome’s redeeming aspect is its tweakability: you can add new features literally with a single click. Gnome supports a plethora of extensions that you can enable without any installation. Just head to the Gnome Extensions website (http://extensions.gnome.org), find the plugin you wish to enable and toggle the button to activate it on your desktop. Some of the popular extensions are designed to help ease the transition for users moving to Gnome from proprietary operating systems, such as Windows.

Furthermore, KDE ships with multiple interfaces or Views designed to make the best of the available desktop realestate. There are different Views for regular screens and netbook though you can use any View on any type of computer. To switch Views, right-click on the desktop and from the context-menu select the Default Desktop Settings option. In the window that opens up, select the View tab and checkout the different views from the Layout pull-down list. Many KDE distros place the Folder View widget on the desktop which displays the contents of a folder in a neat little box that you can place anywhere on your screen. Then there’s the Folder View which lets you place files and folders anywhere on the desktop. The Search and launch View is designed for devices with a small screen or a touchscreen. Each View has additional configurable elements. In addition to bundling the configuration options along with the individual elements, KDE also houses them all under the System Settings panel, alongside other system-wide configuration options to administer the underlying Linux distro. It might seem daunting, but you don’t need to set up or review each and every option before using the desktop. Customising KDE is an on-going process and not a one-time

“Ubuntu, Fedora and Mageia are available in multiple editions with a different desktops.” affair. The desktop is designed to grow and mutate as per your usage requirements.

Kick off with KDE Unlike the other two desktops, the layout and behaviour of the KDE desktop and the placement of its Kickoff app launcher will certainly feel familiar to users from non-Linux operating systems. But KDE is so malleable that many KDE distros look unlike each other. In many ways, KDE is the quintessential Linux desktop with its flexibility and myriad number of options. There’s literally no end to KDE’s customisation options. One of the most useful KDE features is Activities. Using this feature, you can create several context-aware activities, each with its own set of apps and desktop furniture. For example, you can create a Social activity that signs you into all your instant messaging accounts and displays updates and feeds from various social networks. Many KDE distros ship with just the default activity, called the Desktop Activity. However, you can fetch more activities from the internet and build on them to suit your workflow.

31 flavours In addition to these three chief desktop environments, there are a lot more that you can put atop your distro. There are fully fledged environments, such as Cinnamon, as well as lightweight ones, such as Xfce, LXDE and Mate. In fact, most mainstream distros, including Ubuntu, Fedora and Mageia are available in multiple editions with a different desktops. For example, the Ubuntu distro has a number of officially supported spins. There’s Kubuntu which dresses Ubuntu with KDE as well as a Gnome spin, a Xfce spin and another that uses the Mate desktop. The Fedora distro, which was once known as the premier Gnome distro, now also has a wonderful KDE flavour as well. Similarly, you can also use Mageia with the Gnome desktop as well. In fact, Mageia and Fedora, also have install-only DVD images that give the user the option to install multiple desktops.

Switch desktop environments You can also install multiple desktops on top of your distribution, such as the popular Cinnamon desktop. It’s available in the official repositories of Fedora and Mageia and you can install it via their respective graphical package managers. On Ubuntu, it’s available via the ppa:gwendal-lebihan-dev/cinnamonstable PPA. Add this PPA to your system [as explained in the main text, over on page 38] and then install the desktop environment from the Software Center. Once you’ve installed multiple desktop

environments you can easily switch to another one. To do this you just log out of the current desktop environment, and use the login manager and enter your login credentials. Before logging into the desktop, explore the buttons on the login manager. One of the buttons will reveal a drop-down list of all the installed desktops. Select the desktop environment you want to use and the login manager will log you into that desktop. This way you can try them out and choose one you like the best. Choice!

www.tuxradar.com

There are several desktop environments that you can install using your distro’s package manager.


Get into Linux

Fleshing out your chosen distribution. nlike Windows, a typical Linux desktop distribution is more ready-to-use right out of the box. Instead of shipping with basic apps, such as a vanilla text editor or a barebones drawing tool, your average Linux distro will include a fully fledged office suite and a comprehensive graphics editor. This is in addition to the basic set of apps for common tasks, such as browsing the Internet, checking email, instant messaging with your friends across various networks, organising photos, listening to music and watching videos. That said, since we all use our computers differently, you’ll likely want a piece of software that isn’t included by default. Your distro has specialised tools that’ll help you install gazillions of quality open source software without clicking through complex setup wizards. Linux distros use a collection of software tools, both graphical and command-line based, that are together referred to as a package management system. These tools help you install, remove, and upgrade software (also called packages) with ease. Individual pieces of software are grouped inside packages. In addition to the software itself, packages also include other information, such as a list of other packages or dependencies which are required for the app to function properly. Furthermore, the package management system relies on a database known as the repository to keep track of all the available packages.

U

Package Management 101 The Linux world is divided broadly into two different package formats – RPM and Deb. These are precompiled binary packages that are designed to simplify the installation process for desktop users. RPM was created by Red Hat Linux, and is used by distros such as Fedora, and Mageia while Deb is used on Debian-based systems, such as Ubuntu. Additionally, almost every major distro maintains its own set of graphical tools to enable desktop users to install, upgrade and remove app. You must also be familiar with the distro’s repository structure and how and where it houses software. Ubuntu uses the Advanced Packaging Tool or APT package

In addition to enabling repositories you can also select a different mirror for downloading software.

management system. You can use the Software & Updates tool for manipulating Ubuntu’s repositories (or repos). The tool lists repos in four different tabs. By default, the four official repos under the Ubuntu Software tab are enabled. The Main repo includes officially supported software, and the Restricted repo includes software that isn’t available under a completely free license. The two interesting repos are Universe and Multiverse repos, which include software maintained by the community and software that isn’t free, respectively. Unlike Ubuntu the Fedora distro uses the RPM package management system. The distro houses repositories under the /etc/yum.repos.d directory and the main repository is named fedora.repo. Mageia uses the urpmi package which is a wrapper for the RPM package management system. The distro has three official repos. The core repository contains open source packages, the non-free repository contains closed-source apps, and the tainted repository has packages that might infringe on patents and copyright laws in some countries. Each of these repos is further divided into four sub-repos. The release repo includes stable packages, the updates repo includes packages that have been updated since the release,

A backup primer Your distribution will include a tool to help you backup your data and you should take some time out to get familiar with it. Ubuntu, for instance, ships with the Deja Dup backup app which is designed for new users. You can also install it on top of Fedora and Mageia. No matter what backup tool you use, you should take a moment to consider what you should backup and where. Backing up the entire home directory might be convenient but is usually just an overkill. Instead you should just include the directories under your home directory such as Downloads and Documents.


Also check with important app, such as email clients, who keep downloaded emails, attachments and address books under hidden directories (prefixed with a ./) beneath the home folder. Also, keeping the backed up data on another partition of the same disk isn’t going to be of much use, since the whole disk might fail and render the backup copy useless. One solution is to keep the backup on another separate disk or external drive. Or, if you have good Internet bandwidth, the backup app might also help you store the backups on a cloud storage service.

www.linuxformat.com

Déjà Dup has a simple interface that shouldn’t intimidate even first time users.

Get into Linux

and the backports repo contains packages of new versions backported from the Cauldron repository, which will eventually become the stable repo for the next release. There’s also the testing repo which will contain software primary for QA purposes. To configure repos, launch the Mageia Control Center and head to Software management > Configure media sources for install and update. To install the official online repositories, click on Add and then either on Update sources only or Full set of sources. The first choice is the minimum to keep the distro updated, while the second allows you to install new software. These options will populate the window with a list of repos. Then toggle the Enabled checkbox next to the repo you want to fetch software from.

FedEx packages All major desktop distros include a graphical tool for managing packages. Ubuntu’s Software Center is one of the best tools for the job. You can find software by clicking on the category reflecting the type of software that you’re looking for. When you select a category, you will be shown a list of apps. There’s also a search box in the upper-right corner of the window which will look for software matching any entered keywords. Once you’ve found the software you want, click the Install button to its right. This will fetch the software as well as any required dependencies and automatically install it. All newly installed software is added to the Launcher and you can also find it from under the Dash. Fedora uses the PackageKit graphical tool that’s listed as Software in the Gnome’s Activities menu. It too lists software categories as well as a keyword-matching search box at the top to help you find software. Once you’ve found the software that you’re looking for, click on the Install button and the app will add it to your installation. The graphical package management tool for Mageia is named Drakrpm. The tool isn’t as pretty as the software centres in Ubuntu and Fedora, but is very functional and intuitive enough to get the job done. You can filter its list of available apps to show only packages with GUI, security updates, bug fix updates, and more. Applications groups are listed in the sidebar and there’s also a search box to hunt for packages based on keywords. When you find a package you wish to install, simply toggle its corresponding checkbox and click on Apply.

The Repo men The larger open source community offers a lot more packages than the ones listed in your distro’s official repos

and almost every distro has a mechanism to add and install software from these third-party repos. External repos in Ubuntu are known as a Personal Package Archive or PPA. You can add a PPA repo to your distro using the Software & Updates tool. But first you need the address of the PPA. This is listed on the PPA’s Launchpad site and will be something like ppa:example-ppa/example. Now fire up the tool and switch to the Other Software tab. Then click on the Add button and paste the address of the PPA in the window that opens. Ubuntu will then ask you to refresh the repos to enable the PPA.

Most desktop distros have an easy to use graphical package manager.

“Unlike Windows, a typical Linux desktop distro is readyto-use right out of the box.” Similarly, Mageia has a number of third-party repos as well and you’ll need their URL to add them to your distro. Once you have the URL, fire up the Mageia Control Center and head to Software management > Configure media sources for install and update and click on the Add a medium option. Enter the address of the repo in the window that pops up along with its type such as HTTP or FTP. Fedora also has a number of third-party software repos but the most popular is RPMFusion. The repo is further subdivided into two independent repos that house free and nonfree software. You can install both of these repos from within the browser itself by following the instructions on the repos website (www.rpmfusion.org/configuration).

Grabbing Google All popular Google software, such as Chrome, Earth, the audio and video plugin for Hangouts and others can be installed on Linux. But you won’t find them in the official repos of major distros because of their licensing. However, you now have all the know-how to install them with ease if we point you in the right direction. The downloads page of each supported Google app contains links to both 32-bit and 64-bit versions of the software in both RPM and Deb formats. Download the package for your

particular distro and double-click on the file to install it with the distro’s package manager. All official packages for Google apps will also install the appropriate external Google repository in your distribution to keep the software updated. Another popular proprietary software that you may want to install is Skype. Ubuntu users can simply enable Partner Repositories by visiting Software & Updates > Other Software. This will add the official Skype repos and you can then install the software from the Software Center as

www.tuxradar.com

for any other software package. Mageia, on the other hand, includes the Skype package in its non-free repo. If you’ve enabled this repo, then simply search for the get-skype package which will download Skype from its website. You can also head to the Linux download page on Skype’s website and choose your distro and architecture from the pull-down list which will download either a Deb file or an RPM file. Double-click on the file to install it with the distro’s package manager.


Get into Linux

Turn your distro into the ultimate media centre. ost Linux distros designed for desktop users are capable of handling all types of content you throw at them. But some content, especially most audio and video, is distributed in closed formats that are encumbered by patents. The distros can’t play these files straight out of the box, however, most have clearly outlined procedures to allow users to install components to play these popular non-free media formats. Ubuntu gives you the option to install the components that can play files in restricted formats, such as MP3s, during the installation itself. If you’ve already installed the distro, you should then use the package manager to install the ubunturestricted-extras package, which includes popular proprietary codecs and plugins. On Fedora these codecs are bundled in the third-party RPM Fusion repository. You’ll have to first enable the repo as mentioned earlier and then fire up a terminal and enter the following commands to fetch the codecs: su yum install gstreamer{1,}-{plugin-crystalhd,ffmpeg,plugins{good,ugly,bad{,-free,-nonfree,-freeworld,-extras}{-extras}}} ffmpeg libmpg123 lame-libs If you’re using Mageia, you’ll find the multimedia codecs under the Tainted repository, so make sure you enable it following the procedure mentioned earlier. Then launch the Welcome app from under the Tools menu and switch to the Applications tabs. From here you can install several useful and popular packages including multimedia codecs.

Where’s the bling?

Mageia’s Welcome app is a wonderful utility to setup the distro for all kinds of users.

Despite the rise of the open source WebM format, many websites still require Adobe’s Flash plugin to properly stream multimedia content. Getting the Flash plugin on Linux is tricky since Adobe is no longer developing Flash for Firefox on Linux. The only way to use the latest Adobe Flash plugin on Linux is to use Google’s Chrome browser which includes the Pepper-based Flash plug-in. That said, if you don’t want to switch to Chrome you can

You can tweak Linux’s Grub bootloader with the GrubCustomizer tool (ppa:danielrichter2007/grub-customizer).

still install the out-of-date Flash plugin and continue using the Firefox browser. Or, you can extract the newer Pepper-based Flash plugin from the Chrome browser and use it on Chrome’s open source cousin, the Chromium browser. Ubuntu users can install Flash for Firefox with the sudo apt-get install flashplugin-installer command. If you’re using Chromium, you can use the latest Pepper Flash plugin by installing the pepperflashpluginnonfree package. Fedora users can download the Firefox plugin from Adobe’s website by adding its repo. If you are running a 64-bit installation, this command: yum -y install http://linuxdownload.adobe.com/adoberelease/adobe-release-x86_64-1.0-1.noarch.rpm will download and install the correct repo. You can then import the key for the repo with: rpm --import /etc/pki/rpm-gpg/RPM-GPG-KEY-adobe-linux before installing the plugin with: yum -y install flash-plugin Mageia users can simply enable the nonfree repository which houses the Flash plugin and then install it from the Applications tab in the Welcome app.

Best multimedia apps Most distros ship with audio and video players. The popular ones are Rhythmbox which is the default on Gnome-based distros and is well integrated in Ubuntu, and KDE’s default Amarok. In addition to local tracks, both players can also stream Internet radio and podcasts. If you want more attractive looking players, fire up the package manager and look for Banshee and Clementine. Similarly, the default video player on most Gnome-based distros is Totem (now simply called Videos). If you want something with more feature you can grab MPlayer. This is essentially a command-line media player but it has a number of frontends. Gnome users can use Gnome-Mplayer and KDE users can use KMPlayer. There’s also the popular crossplatform VLC that can handle pretty much any file format.


www.linuxformat.com

Get into Linux

Full steam ahead. aming has long been considered Linux’s Achilles’s heel. Over the years we’ve had several quality open source games, but they’ve lacked the mass-appeal of popular gaming titles available on proprietary desktops. All that changed in early 2013 with Valve’s announcement of Steam for Linux client for its hugely popular game distribution service. You can install the proprietary client in your distro with ease. Ubuntu users should enable the Partner repo and then install the client from the Software Center. Fedora users should install the RPM Fusion repos and then install Steam using the package manager. Similarly, Mageia users should enable the non-free repo and then use the Mageia Welcome app to install the Steam client. But before you fire up the Steam client make sure you’ve got the proper drivers for your graphics hardware. This tends to be the major sticking point for users, as card manufacturers restrict distribution of their closed-source driver elements. Start by getting details about the make and model of your graphics card using the lspci | grep VGA command. You can get more details, such as its clock speed and capabilities, with: sudo lshw -C video Your distro will ensure you are using the most suitable open source driver as soon as you boot into the distro. The Oibaf PPA (ppa:oibaf/graphics-drivers) is popular with Ubuntu users for getting the latest bleeding edge open source drivers. However, users of Nvidia and ATI/AMD hardware should use proprietary drivers from the respective vendor for the best gaming performance. Ubuntu users should use the X-Swat PPA (ppa:ubuntu-xswat/x-updates) for the latest stable Nvidia drivers. Once enabled fetch the drivers with: sudo apt-get install nvidia-current Fedora users will find the latest GeForce drivers in the RPM Fusion repo. After adding the repo, install the driver with

yum install kmod-nvidia xorg-x11-drv-nvidia-libs kerneldevel acpid Mageia users should first enable the non-free repo and then launch the Mageia Control Center and head to Hardware > Set up the graphical server. In the window that open click on the toggle next to the Graphic Card label which will display a list of graphics card. Browse the list and select yours. If Mageia has a proprietary driver for the card, it’ll install it. Mageia’s list also includes AMD cards, but if you are using Ubuntu or Fedora, the best source for the proprietary driver is AMD’s website (http://support.amd.com/en-us/ download). This page has several dropdown menus that you can use to pinpoint the exact driver for your graphics card. Then download the suggested driver and extract it to reveal a .run script. Before you install the driver make sure you install its dependencies with: sudo apt-get install dh-make dh-modaliases execstack libc6-i386 lib32gcc1 Once that’s done, you can execute the script with: sh ./amd-driver-installer-13.35.1005-x86.x86_64.run This will launch the graphical AMD Catalyst proprietary driver installer and will also install the Catalyst Control Center GPU management software. When the installer has finished, head back to the terminal and enter /usr/bin/aticonfig --initial to configure the driver. LXF

While distros do come with open source games in their repros, Steam offers access to AAA gaming titles like Dying Light.

User admin basics When Linux is installed, it’s automatically configured for use by a single user, but you can easily add separate user accounts. The superuser, root, has complete access to the OS and its configuration; it’s intended for administrative use only. Unprivileged users can use the su and sudo programs for controlled privilege escalation. Users are grouped together into a group, and inherit the group’s access privileges. Every Linux user is a member of at least one group. Now, you can control which files and folders are accessible by a user or a group. By default, a user’s files are only accessible by that user, and system files are only accessible by the root user. In Linux files

and folders can be set up so that only specific users can view, modify, or run them. This allows you, for instance, to share a file with other users so they can read the file but not make any changes. Access to files is controlled via a system of ownership permissions. You can use the ls -l to list the permissions. When used without specifying a filename it will list the permissions of all files within the directory. The different characters in the output represent a file’s read (r), write (w), and execute (x) permissions for the owner, group, and all other users. You can alter the permissions of files and folders with the chmod command or graphically from the file manager.

www.tuxradar.com

The file manager of almost every distro allows you to tweak the group associated with a file or folder.


Wynn Netherland

Made of

Wynn

Matthew Hanson talks to Wynn Netherland about the importance of APIs, testing and the joy of Ruby.


www.linuxformat.com

Wynn Netherland Wynn Netherland has helped build the web as we know it for nearly twenty years. A prolific creator and maintainer of Ruby API wrappers, he now works on the GitHub API, as well as authoring many books. Linux Format: What are the biggest changes you’ve seen during all your time spent building the web? Wynn Netherland: Great question. It’s got more professional in many ways, it’s got less of a cowboy type of atmosphere. I remember early in my career at Compaq and later on HewlettPackard, we were kind of pioneering open source for e-commerce, as and when. So learning how to take payments online was just something that everyone was trying to figure out. The things we used to do even without RoboSource Control and things like that, the standards that are now commonplace. Even legal requirements back then. So in many ways the web has grown up, it has gotten a lot more… corporate. In some ways we’ve lost a bit of the… I want to say innocence, but a lot of the fun, playground type atmosphere that we had early on where everyone was figuring this out and it wasn’t in this walled garden, except for AOL! It seems like we’ve traded an open web in these last few years for more and more walled gardens. And for some reason as a community we’re OK with that. I’m not sure why.

Interview

LXF: I suppose in the early days younger people started off having Geocities accounts where you could make your own website, play around with HTML and even drop in a checkout if you were selling anything. It was quite easy to get started. Now people are, rather than learning how to code their own website, a lot of people are just thinking I’ll have a Wordpress blog, or a Facebook page. WN: Right, and it seems that Facebook pages are more commonplace than Wordpress blogs. Even if you pick a theme for a Wordpress blog, or any other similar service, or open source CMS tools, at some point you’ll find yourself in HTML and CSS, tweaking it somewhat. With Facebook, it’s what you see is what you get. LXF: So as the internet has changed with, as you say, more walled gardens, would you still say that the internet has changed for the better in some ways? WN: As is the way when humans are involved it’s both a positive and a negative. So interoperability, even though data is more segmented, we’re more interoperable in the way that we exchange data. So back in the late ‘90s we were still doing things like Corba and XML wasn’t really quite a thing yet. You would send HTML to the browser and just hope that it would get rendered! And now we’re kind of

settled on tools that make data interchange really easy; JSON and HTTP, these were the dial tones of the web. But who owns that data? How can I export that data and take it with me? These are some of the larger challenges. Who's looking at my data is probably one of the bigger ones now.

LXF: So creating a stable API is obviously incredibly important – but how important would you say it is? WN: It’s of the utmost importance. I think it’s still in some ways an idealistic goal. As a consumer it’s easy to say ‘Oh you broke my app! You shouldn’t change your API!’, right? We’ve got a small but vocal minority in the API community that would say if you’re versioning an API then you’re pretty much doing it wrong. Things like HyperMedia and media types. Although they help insulate change, I don’t think, short of infinite foreknowledge, that you can always remove it … Sometimes you introduce change even when you’re fixing bugs and bringing it in line with what you published

LXF: Yeah, that’s always a concern. So with your self-proclaimed passion for API user experience that ties in as well, why is it so important? WN: From a user experience standpoint, I began my career in a print design shop for a newspaper, and had been moving further down the stack as I’ve gone. And when open APIs took off a few years ago it began just by HOW THE WEB HAS CHANGED scratching my own itch and wanting to get data out of the systems I was using. I wanted to stitch together my own blog, with all the data that already works. Say you have a bug out from all the points that I was using on the web. I there, and you’ve published the API and it works just found myself frustrated at common, almost one way. Now you’ll have a whole subset of deliberate attempts to make it painful for the people who have built tools around that bug. folks that were getting [the data] out! I think it If you go and notice the bug, and fix the bug, boils down to a lack of empathy. User now you’ve broken people's tools that worked experiences are really just about empathy. We around the work around. tend to equate a user interface as a graphical user interface. We almost imply the ‘G’ now of LXF: So you need to have empathy for the GUI. The way that Unix works, it’s text in, people who are using the API, but there is a text out. That’s a user interface. It just doesn’t certain point where you have to close off and have a graphical element, for the most part. So just do your thing; sort of follow your path? APIs are the same way, even if you adapt REST, WN: Right. So the way that we approach that at there’s a way to archetype those APIs to put the GitHub API is to first quantify it. To measure yourself in the position of the consumer, so that the change that we’re going to have. Whether you can think ‘what would I like to see?’, or ‘what it’s a refactoring, or adding a new end point or would enable workflows around this tool?’ and taking an end point away, something like that. not just wrapping your database in REST.

“We’ve traded an open web in these last few years for more and more walled gardens.”

www.tuxradar.com


Wynn Netherland We’ll measure the change and quantify it. From there we figure out how to minimise the change. That allows us to handle the change. So if you have an API end point that no one’s calling, you can remove that end point and you haven’t affected anybody. Really, it’s human impact that is what it boils down to. So when we look at shutting off a particular API endpoint that is problematic to the Ops team and to the infrastructure, or it’s putting the whole site at risk – something like that. We’ll look at ways to minimise that risk and if we end up having to shut off an API method, the very first thing we’re going to do is see who’s calling it and how often and reach out to those folks and say ‘let’s look at alternate workflows to get around this’. LXF: So you mentioned GitHub. How did you get involved? WN: It was through the API wrapper. I was working with the Changelog podcast a few years ago, and one of the things we wanted to do was just showcase a lot of the open source that was happening on GitHub. It was something we called ‘Tale of the Changelog’ and it was basically a fire hose of everything that was coming out on GitHub. I didn’t like any of the Ruby wrappers that went around at the time, they mapped too closely to REST – I’m a firm believer that API wrappers should be idiomatic to the language that they’re written in. So it should feel like an extension of Ruby if you’re writing in Ruby. So I wrote a wrapper that later became octokit and then realised after they reached out to me a couple of years later to join GitHub, that GitHub was using it internally. Had I known that I’d have put a little more effort into it! Because for me it was a hobby project. The very first time I bootstrapped the GitHub source code and it started installing octokit I was ‘wow!’. So now we use it as the primary API interface for a lot of

our tools, because a lot of our tooling is written in Ruby. So it powers things like checking permissions of internal scripts, getting stats, authenticating our various internal applications. It all uses octokit internally. LXF: It must have been flattering to see your code there… WN: It was, but any open source project is a team effort, so there’s been a number of contributors over the years. Erik Michaels-Ober has probably been the biggest one on octokit. He works at Soundcloud and has a number of wrappers out there. But when the Mac team wanted to extract their API wrapper for Objective-C from the GitHub for Mac app, they wanted a name. Naming things and invalidating caches are the big computing problems, right? So, it was obvious that the best name for the Objective-C client would be octokit, so I told them: ‘You guys should have this name, and we should just make that an overall umbrella brand – for lack of a better term – and so our project became octokit and I moved it from my personal GitHub namespace into the octokit org, and so now we’ve got flavours in Objective-C and .net. People keep asking why there’s no Python. We do have one in Go, but our philosophy on releasing Octokit wrappers is that we want them to be first party, something that we’re using and testing and maintaining. That way we’re not trying to maintain something that we’re not really using, as it gets out of date. LXF: So as part of the API team at GitHub what are you main responsibilities? WN: Keeping the lights on, first and foremost. In the API team at the moment there are just two of us, there’s Jason Rudolph and myself. A small team embedded within the larger, what we call ‘.com engineering team’. So, primary

responsibility is site-wide horizontal aspects to the API. Things like authentication, rate limiting and anything that any of the API swots would use. Then we coach and cheerlead the feature teams for pull requests issues and repository access to extend and maintain their APIs, as an extension of what they’re doing on the site. The API is a smaller app embedded within the larger web app, so we share a lot of the backend code, just the workflows and the output are isolated. LXF: So you test a lot of the API behaviour as well, and in previous talks you’ve given you’ve touched upon some of the tools you use, such as JSON. Could you explain to us the testing process? WN: Sure, so our primary output format for the API is JavaScript Object Notation (JSON), so we use JSON schema. One of the shortcomings of JSON is, unlike XML, there’s no built-in schema technology that comes along with it. It’s very easy to parse and you get real primitive spec in the language of your choice, but there’s no way to know what shape the data should have. So we use JSON schema, both in the input and the output. Just recently we’ve been using it on the input and we’re getting some big gains there. Traditionally over the output we can, when we run our test suite, look at the JSON that’s coming back and validate that this object response should have these properties, and things of that sort. Now we’re starting to add that back in, allowing us to validate inputs. So when you go to create a pull request you pass us some JSON, we can run it through those same schema validations to say ‘Sorry this isn’t the right format’. And what it does for us is that it allows us to declaratively specify in the test what a valid document looks like. Then we don’t have to write one-off tests for every permutation under the sun to check if they are valid or not. LXF: Also Rack::Test is another tool you use… WN: So Rack::Test is for any Rack apps, so the Ruby world has and HTTP pipeline called Rack. So a few years ago a lot of the web frameworks got together and standardised on an inputoutput pipeline pattern that they called Rack. So now many conform to this interface where you can rack up – like with a rack of servers in a datacentre – you can rack up these middleware applications to do things in a pipeline format. So if you have things that take an HTTP response and they act on it in some way and they pass it on down to the next app to consume. So things like authentication, rate limiting, cookie handling, if you need to scrub input or output for cross site scripting vulnerabilities, things like that. We use Rack::Test to be able to test our apps at test time. LXF: It’s important to catch any potential problems at that stage, isn’t it? So can you accurately track how your API is being used?


www.linuxformat.com

Wynn Netherland Going back to your empathy comments, it’s definitely useful to see how your API is being used and how it performs. WN: We have a couple of metrics on a couple of different levels. One of those is just sheer responses and the status of those responses at particular endpoints. We know the Events endpoint which is our activity streaming API. It gets a certain amount of volume and we know that these are the types of applications that people are building on top of GitHub. So that’s one level just to see the sheer volume, but volume doesn’t always tell the whole story. There’s also, I guess because not every method has the same value from a workflow standpoint, the things that really excite me are the, I guess more in-depth tasks, around creating pull requests and merging pull requests and doing everything that you would do in the website. Instead of the API being simply a data store for a mashup to use as a database, I really get excited when I see applications building entire workflows on top of it, their cloning a repository, they’re adding features and then they’re using the API to submit those pull requests. So you’ll see a lot of tools that build code review type applications on top of GitHub that mirror what we’re doing with the website, but fit their individual workflows better.

WEB VS NATIVE

coming from. Or the new code path is 100% in line, but it’s slower, or faster. Science allows us to compare those things and roll them out with confidence and make it data driven, evidencebased, instead of just kind of putting your finger in the air and saying ‘I hope it works’! LXF: And I suppose it helps you identify patterns, then? WN: Yes, especially when it’s things that involve security that just can’t afford to be wrong. LXF: How long have you been using Science? WN: Science has been around… this particular definition of it [laughs]… I think it’s about a year old. It hasn’t been open source that long. I first spoke about it back in March 2014, and it’s really catching on throughout the team as a codebase gets long in the tooth and we’re trying to figure out how to make things a little faster and a little bit more modern in certain places. LXF: So how important are open source tools for creating and testing APIs? WN: I think they’re indispensable. Outside of a Microsoft ecosystem I don’t know how you would do it otherwise. The very first slide of my talk [at OSCON last year] is about how I want to build a resilient API in open source, and for me that actually starts with them [the audience] some tea mugs and some Shell tools… I only have 40 minutes for the talk so I have to fast forward a little bit! But it really involves everything from the web frameworks you choose, what servers they’re going to run on, the instrumentation stack of how you’re going to collect metrics and gather any sort of feedback on how your website performs. Your testing frameworks… every one of these aspects is a decision matrix, really, figuring out which tools you’re going to use. It’s almost an embarrassment of riches in the open source community. Because you have choices. It’s a tyranny of choice, and often it’s nice to come into a project, like I did with GitHub, and a lot of those decisions have already been made, you just get to play in somebody else’s sandbox.

“It’s a false dichotomy because even native apps are using JSON and APIs.” LXF: Another open source project that you’ve touched on in your talk is Science. Could you explain what that is? WN: This is from John Barnette and Rick Bradley, two GitHubbers. So we had a particular problem inside of the .com codebase where as good as the tests suite is, and it’s a phenomenal test suite just based on the experience of projects I’ve worked on, there are certain aspects of your applications, a soft underbelly, that you really don’t trust your tests 100%. It could be because your framework’s using metaprogramming techniques that are difficult and are a dynamic language to track down at design time. So if there’s something that’s critical that can’t change but needs refactoring, Science is an application… a project, rather… that tells you test alternate code paths. If you have a method that is read only, it’s not mutating, and it returns a value you now, just like the scientific method, you can run both code paths side by side and compare the results and then publish those results into some charts and graphs. So you can see side by side, OK this is the original code path, this is the new code path. They differ about 10% of the time, and we need to figure out where that variance is

LXF: So going back to the APIs, there’s now a huge variety of devices that we can now use to access the internet and online services. Has that provided challenges for creating APIs? WN: It definitely introduced more load on APIs. I saw a definition the other day where, when it comes to at least the mobile space, when people say ‘web versus native’, it’s almost a false dichotomy because even native apps are using JSON and APIs behind the scenes. For the most part those are hitting the same applications, they’re just rendering them in a different

www.tuxradar.com

computer-friendly format, right? So it’s definitely exploded the amount of applications that are wired up to the web, and it’s not just the browsers that are having that fun anymore. But it’s also kind of changed the way we think about native apps. I remember one of the ways I got into writing my second or third API wrapper was for the now defunct [location-based social media] service called Gowalla. At the time it was up against Foursquare. Where with Foursquare you would be competing with everyone else to be mayor of a certain geographic location, with Gowalla it was more of your personal backpack. So every state of the United States, or every country in the world that you visited, you got a different badge. You could also leave virtual objects for other people. It just had a totally different vibe, it was really art driven and I just loved it. I knew some friends who worked there and I kept bugging them to come out with an API and they kept saying they were working on it, but keeping it really close to the vest. Well, then their mobile app came out, so I wrote a Gowalla API wrapper by sniffing out their mobile API, by running their app through a proxy server. It also turned me on to a lot of the things that I’d poked on in their API that were sound decisions that had made those patterns that you see in other APIs now, and a lot of those have landed in the GitHub API as well. LXF: So when you’re trying to figure out how different people are going to be accessing certain services or apps, to test that is probably quite difficult as well? WN: So one of the things we’ve done since change is expensive, and usually not welcome as we mentioned before! We’ve started tinkering with the notion that beta API endpoints, and normally the ways these happen is you’ll have somebody will release an API and after it gets some adoption and they figure out what they want to rev on it and stuff, they’ll introduce the second version of the API – and GitHub even did this in the past where they had the V2 and a whole different namespace of


Wynn Netherland URLS and then V3. When we moved to V3 we decided we were going to start versioning endpoints based on output, not based on path. Which meant a certain HTTP header that you send in, you ask for a particular version of the resource. And here lately over the last 18 months of so we’ve started using that as essentially beta passwords for new features. We’ll release a beta version of an API and right there in the media type header we’ll have the word ‘preview’, which is basically your opt-in contract to say ‘I understand I’m not supposed to build anything in production on top of this API endpoint’, but it allows us to get some realworld feedback. It’s subject to change at any moment, we make no warranty period of these beta APIs, but we’ve done this a couple of times and gotten some really nice real-world feedback from folks that have used some particular feature in ways that we haven’t anticipated. So we’ve made changes before we release the final product and the final product has been much better for it due to the changes.

key to the Ruby community. Ruby is just… I was a .net developer before I came into the Ruby community, and for me personally nothing makes me as happy when I’m writing Ruby. There’s certain times when you need more performance than Ruby just can’t give you, because it’s a dynamic language, although you can push it pretty hard. Go, I like a lot but it’s not as approachable as Ruby for the most part. Which is interesting because I’ve written a lot of JavaScript too, but there’s just something about JavaScript’s syntax, with Function word everywhere that tends to read not as close to, I guess, human speech, as Ruby does … You can do interesting things with its metaprogramming capabilities to write specific languages … instead of writing around dozens of assertions, I can roll that up into like a five-line DSL, that I can write those assertions behind. It makes it very clear and very declarative about what it’s going to do without having to be as verbose as a lot of languages have been.

LXF: And that’s a lot friendlier to newcomers as well. They don’t just have to wait until the end to see the result… WN: It’s very easy to follow, especially if you’re SOFTWARE EATING THE WORLD new to Ruby … some of its faults are that there can be three ways to do something, and you might not know which is the idiomatic way to do something in Ruby. So but now if we discover that a particular resource producing it does take a bit more time than consuming it, but it’s very easy to follow. doesn’t work well for a particular class of users, we can give them a different media type output LXF: So GitHub has recently reported that it that does. For instance, there’s fields that have has eight million users, which makes it the to be removed when certain features are largest code host in the world, what do you removed from an application, and in the Python think has helped make it so successful? or Ruby space, removing a field from a JSON, WN: Open source and getting influential even setting it to null, is no big deal. For static developers and influential projects into GitHub clients they don’t like this as much. Compile has just been phenomenal. I remember in the time issues. So in those cases we can offer a early days, back in the beta period, when you different media type on that endpoint and allow had to sell someone on Git first before you them to fetch a different version of that resource that meets their needs. could sell them on GitHub. Just doing that at lunch there was a gentleman sitting next to me, LXF: You mentioned at GitHub Ruby is the they’ve just moved to Subversion with Git. language of choice? So the whole table just erupted into ‘Oh you WN: Yeah, I think the vast majority of our need to try this and that, and that’s how it tooling is in Ruby. We’re doing more and more in helped me’, so when you see those light bulb Go, with media uploads and things like that, just moments and you see entire projects moving because it’s well-suited for concurrent uploads, into GitHub, it’s just a tidal wave of activity and things like that. Of course, if you’re doing a Mac community that goes into GitHub and that goes app you’ll be using Objective-C and now Swift, back out. The big thing is there’s no longer a those teams are now evaluating Swift and gatekeeper for most projects. The Canonical moving over to that. Of course, the Windows repository is whichever one is most active, and team is going to be in .net. So we’re always so if your project lies dormant, there’s nothing going to be a polyglot shop, but when it comes stopping the community from forking it and to web apps, the vast majority are in Ruby. now this project is the Canonical one, because that’s where all the activity is. Maintainers really LXF: What is it about Ruby that’s so useful? have to make an effort to stay on top of the WN: Well, ‘optimising for happiness’ is kind of communities. Or delegate. But we’re still finite LXF: So has the take up been encouraging? WN: It has. Not only from a beta standpoint,

“It’s getting to a point where there’s this ubiquity that’s easy to underestimate.”


www.linuxformat.com

humans so as activity has exploded, you’ll see a new pattern emerge where someone commits a pull request and it looks good and it looks solid, the next thing you do, which in previous years would be unheard of, you’d go ahead and give them a commit, so they can then help maintain that project, because there’s just no way you’re going to maintain all the open source – you can’t keep up with your tooling from that standpoint. LXF: I suppose positive word of mouth in the open source community helps. WN: I’m still blown away by if you go out of the echo chamber of typical tech conferences – especially if you go into design circles – where more and more of the HTML 5, and some other technologies, that whole community is now intersecting with traditional open source development communities. Or you go into enterprise-flavoured conferences, we’ve got far less penetration, but you see it’s still advancing at the same rate, so it’s very encouraging to get immersed in those communities and spend a few days with folks that really had that light bulb moment of ‘Oh wow, this could revamp things’. LXF: What does the future hold for GitHub? WN: Revolutionise the way people write software. I guess the saying inside the company is that software is eating the world. So if you think about how far software goes into every facet of life. From the devices that you hold in your hand to even the scarier parts, like hospitals. It’s not the needles that scare me, it’s the software! Because I know how a lot of that is written! [Laughs] But software is everywhere, right? And it’s getting to a point where there’s this ubiquity that’s easy to underestimate, and so the more software is around us, the more tooling is going to have to evolve to keep up with just every industry out there. I think GitHub is going to be ground zero for that activity. LXF

Helping you live better & work smarter

LIFEHACKER UK IS THE EXPERT GUIDE FOR ANYONE LOOKING TO GET THINGS DONE Thousands of tips to improve your home & workplace Get more from your smartphone, tablet & computer Be more efficient and increase your productivity

www.lifehacker.co.uk

twitter.com/lifehackeruk

facebook.com/lifehackeruk

Riak - NoSQL

Mihalis Tsoukalos explains everything you wanted to know about NoSQL, but were too afraid to ask and discover why admins love this high-speed system. he world of databases moves slowly (but perhaps not when transacting), so when a revolution hits it can take decades for the repercussions to be felt. Coined back in 1998, NoSQL are databases that started life not using the then standard query language SQL. But more revolutionary the actual database designs themselves, moved away from the standard relational model altogether for speed and ease of design. As you might expect by the name, even though NoSQL databases were originally designed not to use SQL, they can now, instead they use various different query languages. While these originally might have appeared in 1998, NoSQL didn’t gained prevalence until the late Noughties when it was adopted as a rallying Twitter hashtag for a group of non-relational distributed database projects that were after something small and unique.

T

If you are wondering whether or not it’s worth considering NoSQL databases, you should be aware that according to DB-Engines Ranking (https://db-engines.com/en/ ranking), MongoDB, a popular NoSQL database, is currently the fifth most popular after Oracle, MySQL, Microsoft SQL Server and PostgreSQL – and even Oracle has a NoSQL version of its famous database. The problem with relational databases is that in order to store complex information you

The next logical step is to use many machines to run your database, but that also creates a problem because relational databases were originally designed to run as single-node systems. So, large companies, such as Google and Amazon, developed their own database systems, Bigtable and Dynamo respectively, that were quite different from traditional relational database systems, and which inspired the NoSQL movement. It’s quite difficult to define what a NoSQL database is but you can identify a few common characteristics among NoSQL databases: they are non-relational; open source (although not always); schemaless; easy to be distributed on many machines (again, not always) and trying to serve data from the 21st century web culture. So, NoSQL databases are designed for the web and don’t support joins, complex transactions and other features of the SQL language. Their terminology is also a little different, but lets dive into the details.

“NoSQL DBs are designed for the web and don’t support joins and complex transactions…”


have to deconstruct it into bits and fields, and store it in lots of different tables. Likewise, in order to restore the data, you have to retrieve all those bits and fields and put them back together. Neither of those two tasks is efficient particularly if you have a big and busy website that’s storing and querying data all the time.

www.linuxformat.com

NoSQL - Riak

Examples of column-family NoSQL databases include Cassandra and Apache HBase. Graph This model is totally different from the other three as it is based on the Graph structure. As a logical consequence Graph NoSQL databases handle hierarchies and relationships between things very well; doing similar things with a relational database would be an extremely challenging and slow task. Neo4j is a graph NoSQL database. For this article, we’ll be using Riak as our NoSQL database test case.

Installing Riak Every Riak server has a web interface. In this case, we’re accessing the server statistics, using http://localhost:10018/stats/ The port number and IP are defined in riak.conf.

The main advantage of a NoSQL database is that they are suited to and efficient for big data and real-time web applications. They also offer easy scalability, and enable you to implement high availability painlessly. They are also generally easier to administer, set up and run, and they can store complex objects. Additionally, it’s easier to develop applications for and with NoSQL databases. Their schema can change easily without any downtime because, in reality, they have no schema. Most of them, with the exception of Oracle NoSQL, are open source projects. Key disadvantages of NoSQL databases include the fact that they require a totally new way of thinking and that you still need a DBA on large and/or critical projects. If your company needs to use both SQL and NoSQL databases, you will have two entirely different systems to program and administer and therefore will need even more people. Being relatively new, they are not as mature as relational databases; therefore choosing a NoSQL database for a critical problem may not always be the safest solution, but this will not be a problem in a couple of years. The last disadvantage is the fact that although they look like they have no schema, you will need to assume an implicit schema in order to do some serious work with your data. This isn’t unexpected because as long as you are

working with data, you cannot get away with having a schema, even an informal one. There are several kinds of NoSQL database each of them being good in one or more areas but not all. You can categorise NoSQL databases according to their data model: Document This is a very common data model. It thinks of the database as a big storage for documents where each document is a multipart data structure that’s usually represented in forms of JSON. You can still store documents in any format you want. MongoDB, CouchDB and RavenDB are representative document NoSQL databases. Key-Value This is also a common data model that’s similar to the hash map data structure, where you have a key and you ask the database to return the value stored for that particular key. The value can be anything from a single number to a whole document. The database knows nothing about the stored data. Examples of key-value NoSQL databases include Riak, Redis and Project Voldemort. Column-family This is a rather complex data model. You have a ‘row key’ that enables you to store and access multiple column families. Each column family is a combination of columns that fit together. Row keys must be unique within a column family. The data model might be more complicated than the others but it results in faster retrieval times.

The first thing you should know before installing Riak is that you need Erlang [See Tutorials, p88, LXF194] on your system. The best way to install Riak is by compiling it from source because you have better control and a totally autonomous build of Riak. Follow the next steps: $ wget http://s3.amazonaws.com/downloads. basho.com/riak/2.0/2.0.2/riak-2.0.2.tar.gz $ tar zxvf riak-2.0.2.tar.gz $ cd riak-2.0.2 $ make rel Alternatively, you can get the Riak source code from GitHub and compile it as before: $ git clone git://github.com/basho/riak.git $ cd riak $ make rel Both ways should work without any particular problems; we used the first way to compile Riak. After successfully compiling Riak, you can find its main binary files inside the ./rel/riak/bin directory. In the same directory that you build Riak, you can run make devrel and get eight ready to run Riak databases that we will use as example servers.

Generating a Riak cluster with five nodes is pretty easy, see p53 for details.

Map and Reduce MapReduce is an advanced querying technique and a tool for data aggregation used in NoSQL databases. It’s an alternative technique for querying a database that differs from the usual declarative querying techniques. You give instructions to the database on how to find the data you are looking for and MapReduce tries to find the data. (See the top of p52 for a simple

example of how MapReduce works.) Using MapReduce can be very tricky sometimes. Nevertheless, it enables you to create queries that would have been extremely challenging to create using SQL. Once you understand the MapReduce process and practice it, you will find it both very reliable and handy. The MapReduce solution

www.tuxradar.com

takes more implementation time but it can expand better than an SQL solution. It provides some flexibility that’s not currently available in the aggregation pipeline. The tricky thing is deciding whether or not the MapReduce technique is appropriate for the specific problem you are trying to solve. This kind of knowledge comes with experience!


Riak - NoSQL

This is the main reason to get the Riak source code and compile it for yourself. Before we continue with the rest of the article, we need to introduce you to some terms. First, a Riak node is analogous to a physical server. A Riak cluster is a 160-bit integer space which is divided into equallysized partitions. Partitions are, in turn, the spaces into which a Riak cluster is divided. Each vnode in Riak is responsible for a partition. Vnodes coordinate requests for the partitions they control. A Riak cluster can have many nodes that reside on the same or different physical machines. A ring is a 160-bit integer space equally divided into partitions, and a bucket is a namespace for data stored in Riak. Internally, Riak computes a 160-bit binary hash of each bucket/key pair and maps this value to a position on an ordered ring of all such values. As you will see later in this article, any client interface to Riak interacts with objects in terms of the bucket and key in which a value is stored, as well as the bucket type that is used to set the properties of the bucket. The default mode of operation for Riak is to work as a cluster consisting of multiple nodes. Riak nodes are not clones of one another by default. You can start three example Riak database servers – you don’t have to start all eight Riak servers – by executing the next commands: $ ./dev/dev1/bin/riak start $ ./dev/dev2/bin/riak start $ ./dev/dev3/bin/riak start $ ./dev/dev1/bin/riak start Node is already running! Each Riak server offers a web interface (see top of p50 for an example of what you will see after connecting to a Riak server). The port number and the server IP address are defined inside the riak.conf file. This is a plain text file that you can edit. The following command reveals the IP and the port number that each Riak server listens to: $ grep listener.http.internal `find ./dev -name riak.conf` ./dev/dev2/etc/riak.conf:listener.http.internal = 127.0.0.1:10028

A MapReduce example. It may look simplistic but MapReduce is a very powerful technique. Attempting the same with SQL would be extremely difficult.

./dev/dev1/etc/riak.conf:listener.http.internal = 127.0.0.1:10018 ./dev/dev3/etc/riak.conf:listener.http.internal = 127.0.0.1:10038 And so on… Every node in Riak has a name associated with it. You can change the name by changing the nodename variable of the riak.conf file. The first server (dev1) uses port number 10018, the second Riak server (dev2) uses port number 10028 and the third (dev3) uses port number 10038. Riak versions prior to 2.0 used a configuration file called app. config which has been replaced by riak.conf. The easiest way of finding out if a Riak node is up or down is via the curl command and the web interface of the node to ping it: $ curl http://localhost:10018/ping OK $ curl http://localhost:10038/ping curl: (7) Failed to connect to localhost port 10038: Connection refused Alternatively, you can use the following: $ ./dev/dev1/bin/riak ping pong $ ./dev/dev6/bin/riak ping Node ‘[email protected]’ not responding to pings. The advantage of the ‘curl way’ is that you can run it from a remote machine – provided that the Riak server also listens to an external IP address – without having to login to the

machine that runs the Riak node. You can stop the dev1 Riak server by executing the ./dev/ dev1/bin/riak stop command. Riak uses epmd – the Erlang Port Mapper server – which plays a crucial part in the whole Riak operation. The epmd process starts automatically by the erl command if the node is to be distributed and there’s no running instance present. The epmd process enables Riak nodes to find each other. It’s an extremely lightweight and harmless process that can continue to run even after all Riak nodes have stopped. You may kill it manually after you stop all Riak nodes, but this isn’t compulsory. The following command lists all names registered with the currently running epmd process: $ epmd -names epmd: up and running on port 4369 with data: name dev3 at port 49136 name dev1 at port 55224 name dev2 at port 48829

Storing and retrieving data You can connect to Riak dev1 server and store a document using the web interface: $ curl -v -X PUT http://127.0.0.1:10018/riak/ LXF/test -H “Content-Type: text/html” -d “
This is a test.
”

Riak benchmarking Basho (http://basho.com) offers a benchmarking tool for Riak written in Erlang. You can get and install it with: $ git clone git://github.com/basho/basho_ bench.git $ cd basho_bench $ make You should then run ./basho_bench myconfig.config to get the tool collecting data, and either create a myconfig.config yourself or


modify an existing config. Existing files reside in the examples directory. We used the examples/basho_bench_ets.config file as a starting point and added the {riakclient_nodes, [‘[email protected], ‘[email protected]’]}.” line. Basho Bench creates one Stats process and workers based on what’s defined in the concurrent configuration setting in myconfig. config file. As soon as these processes are created and initialised, Basho Bench sends a

www.linuxformat.com

run command to all worker processes and this initiates the testing. The Stats process is notified every time an operation completes. It also gets the elapsed time of the completed operation and stores it in a histogram. All the results are inside the tests directory. The latest results can be found using the ./tests/current/ soft link. To generate a graph against the current results, run make results. [See the bottom of p53 for a sample output.]

NoSQL - Riak

What is actually stored in the /riak/LXF test location is what follows the -d option. When you successfully insert a new value, Riak will return a 204 HTTP code. As you already know, Riak is a key-value store, therefore in order to retrieve a value you need to provide a key to Riak. You can connect to Riak dev1 server and ask the previously stored document by going to the http://127.0.0.1: 10018/riak/LXF/test URL. Every URL follows the http://SERVER:PORT/riak/BUCKET/ KEY pattern. The following command returns the list of available buckets: $ curl -i ‘http://127.0.0.1:10018/ riak?buckets=true’ HTTP/1.1 200 OK Vary: Accept-Encoding Server: MochiWeb/1.1 WebMachine/1.10.5 (jokes are better explained) Date: Fri, 19 Dec 2014 21:13:37 GMT Content-Type: application/json Content-Length: 33 {“buckets”:[“LXF”,”linuxformat”]} The following command returns the list of keys in a bucket: $ curl ‘http://127.0.0.1:10018/buckets/LXF/ keys?keys=true’ {“keys”:[“test2”,”test”,”test3”]} Most of the times, you are going to use a script written in a programming language to access a Riak database. The following is a Python script that connects to a Riak database, stores and retrieves a document: import riak # Connect to the cluster client = riak.RiakClient(pb_port=10017, protocol=’pbc’) # The name of the bucket bucket = client.bucket(‘python’) # “myData” is the name of the Key that will be used aRecord = bucket.new(‘myData’, data={ ‘Name’: “Mihalis”, ‘Surname’: “Tsoukalos” }) # Save the record aRecord.store() # Define the key for the record to retrieve

myRecord = bucket.get(‘myData’) # Retrieve the record! dictRecord = myRecord.data # Now print it to see if all this actually worked. print dictRecord $ python myRiak.py {u’Surname’: u’Tsoukalos’, u’Name’: u’Mihalis’} The pb_port value of 10017 is defined in the ./dev/dev1/etc/riak.conf file using the listener.protobuf.internal parameter. This is the Protocol Buffers port that is used for connecting to the Riak Cluster. Due to the flexibility in the way that a NoSQL database stores data, inserting, querying and updating a NoSQL database is more complex than a database that uses SQL.

Generating a Riak cluster Creating and manipulating clusters in Riak is relatively easy with the help of the riak-admin command. If you try to add a node that’s not already running to a cluster, you will fail with the following error message: $ dev/dev2/bin/riak-admin cluster join [email protected] Node is not running! $ ./dev/dev2/bin/riak start $ dev/dev2/bin/riak-admin cluster join [email protected] Success: staged join request for ‘[email protected]’ to ‘[email protected]’ $ dev/dev2/bin/riak-admin cluster join

Riak offers a benchmarking tool called Basho Bench. The graph is produced with R.

[email protected] Failed: This node is already a member of a cluster Similarly, if you try to join a node to itself, you will get an error message: $ dev/dev1/bin/riak-admin cluster join [email protected] Failed: This node cannot join itself in a cluster The following command shows the members of an existing cluster: $ dev/dev2/bin/riak-admin status | grep members ring_members : [‘[email protected]’,’d [email protected]’] $ dev/dev1/bin/riak-admin status | grep members ring_members : [‘[email protected]’,’d [email protected]’] $ dev/dev3/bin/riak-admin status | grep members Node is not running! Another useful command that shows the status of the nodes is the following: $ ./dev/dev1/bin/riak-admin member-status The joining status is a temporary status and will become valid when all changes that are waiting in a queue will be applied and committed. If you want to force changes, you should execute the riak-admin cluster commit command. If you run the riak-admin member-status command again you will see the new status of the dev3 node, and the riak-admin cluster plan command displays the changes that are about to be applied. For a node to actually leave the cluster (see bottom of p51, to see what an interaction with a cluster of five nodes looks like), you must first review the changes using the riak-admin cluster plan command and then commit them with riak-admin cluster commit. So far, you won’t have seen any security when interacting with a Riak database. Nevertheless, Riak supports users and passwords. You can find a lot more information on how Riak deals with authentication and authorisation at http://bit.ly/RiakDocsAuthz. LXF

Data consistency Data consistency in databases is critical. ACID (Atomicity, Consistency, Isolation and Durability) is a set of properties that guarantee that database transactions perform reliably. Atomicity means that when you do something to change a database, the change should work or fail as a whole. Isolation means that if other things are taking place at the same time on the same data, they should not be able to see halffinished data. Durability refers to the guarantee

that once the user has been notified of the success of a transaction, the transaction will persist, and won’t be undone even if the hardware or the software crashes afterwards. Graph databases perform ACID transactions by default, which is a good thing. On the other hand, not every problem needs ‘perfect’ ACID compliance. MongoDB is ACID-compliant at the single document level, but it doesn’t support multiple-document updates that can be rolled

www.tuxradar.com

back. Sometimes, you may be OK with losing a transaction or having your DB in an inconsistent state temporarily in exchange for speed. You should carefully check the characteristics of a NoSQL database and decide if it fits your needs. Nevertheless, if data consistency is absolutely critical, you can always implement it in code if it’s not fully supported by your NoSQL DB. Keep in mind that this might be non-trivial especially on distributed environments.


Fedora 21

Fedora 21

Jonni Bidwell dons his felt hat (but leaves his whip at home) to uncover the treasures in the new distro. e tried Fedora 20 roughly a year ago [Reviews, p19, LXF180] and now we're trying out the latest release of the venerable distribution, which is made of three distinct distributions. By the time you read this Fedora 21 will have been released into the wild and will be roaming the internets. Additionally, at the time of writing, thanks to the release being postponed several times while the kinks were ironed out, your reviewer has to make do with the latest release

W


candidate (being at the mercy of deadlines, and also the bar at office parties). As such we won't mention several minor bugs that we

Schrodinger's Cat, Spherical Cow, Beefy Miracle – but if all you care about is quirky names you can always count on Ubuntu. Fedora, like Debian, has always been targeted towards the intermediate and above tiers. This doesn't mean it's hard to use, but it does mean there's a distinct absence of wizardy interfaces for doing stuff that's best done at the command line. Fedora has always enabled SELinux by default, but unless you have specific requirements it won't get in your way. It didn't always used to be this way, incidentally.

“a distinct absence of wizardy interfaces for stuff that’s best done at the command line.” know will be fixed when the official release makes it out the door. Breaking from tradition, this release doesn't have a funky codename like its predecessors – Heisenbug,

www.linuxformat.com

Fedora 21

hile we could spend the entire article, detailing all the version updates and comprehensively enumerating all the packages with which a default install of this new Fedora ships, we shan't. For most users, everything will be sufficiently contemporary, excepting the usual gripes about official proprietary driver packages, for which there exist ways and means by which to ameliorate the situation. In general, if you do need up-to-the-minute releases, then use a rolling release distribution, such as Fedora's Rawhide. Fedora is pretty much the last major distro still to use the much maligned Gnome Shell as the default desktop, but don't worry if it's not your cup of tea. There are plenty of other desktops available to install, and with a big name distro such as this, things are sufficiently wellpackaged as to make the process all but trivial. The hardest part will be making a cup of tea to enjoy while the requisite packages are downloaded and installed.

W

Three is the magic number So what exactly is worth mentioning? Well for starters this isn't really one distribution, but three: Fedora has been split into Cloud, Server and Workstation incarnations. Lots of work has been done to make the Cloud image as small as possible, in particular the kernel package is now a meta-package, with any modules irrelevant for a minimal cloud image annexed off to a separate package called kernel-modules. The kernel-core package contains, besides the kernel image itself, just those modules required to work as a, for example, an EC2 or OpenStack instance. So the Cloud image relies only on kernel-core package while the Server and Workstation images depend on the meta-package (which installs both the -core and -modules packages). The cloud release also features Red Hat's Project Atomic framework, which enables you to set up a so-called Atomic Host, being a lightweight platform whose sole purpose is to host Docker containers. Besides Fedora 21, Atomic hosts are also available built from CentOS and Red Hat Enterprise Linux. This is an ambitious project (in keeping with Red

The weather program didn’t recognise Bath UK, We’re sure Bathurst, Canada is lovely, but also glad to be here, rather than there.

Hat's often touted adage "to lead and not follow") and is in direct competition with the already established CoreOS distribution (which incidentally is planning a container schism with the introduction of a rival format entitled App Container). Project Atomic does away with traditional package management entirely – the underlying OS is updated, using a new technology called OSTree, in much the same way as a Git repository, which makes it easy to perform incremental updates, and to roll things back when they fail. On the server side of things, we have several new technologies to streamline provisioning and administration. Rolekit provides a deployment system so that servers can be quickly equipped to perform a certain function, or role, if you will. Each role can be managed from the same interface, providing simplicity through consistency. The Cockpit is a web-based server management interface, which will be particularly appealing to those new to administering Linux. Much like the Webmin control panel, Cockpit enables you to

start and stop services, inspect log files and perform all manner of perfunctory housekeeping without having to type anything (but where's the fun in that?). Cockpit can also manage remote servers as well and best of all it doesn't force itself upon you – services started from Cockpit can be stopped via the command-line, and vice versa. Perhaps the most ambitious feature of Fedora Server 21 is the introduction of OpenLMI, a Linux Management Infrastructure which aims to abstract away the myriad tools, syntaxes and idiosyncrasies which sysadmins have hitherto had to learn and subsequently keep up with. OpenLMI steps in with a standardised API accessible from C/C++, Python, Java or good old-fashioned command line tools, allowing routine tasks to be carried out with ease. OpenLMI provides a clientserver interface, and so lends itself to remote management and monitoring. A web interface is available too and as such, there's a fair amount of overlap with the aforementioned Cockpit. However, these tools won't tread on

Upgrading If you're already a Fedora user, and want to upgrade your system swiftly and efficiently without munging your carefully-curated package selections, then good news: You can use the FedUp tool to do precisely this. Make sure everything is up to date, then install FedUp: $ sudo yum update $ sudo yum install fedup

While it’s possible to use an ISO as FedUp's upgrade source (using the --iso option), it's simpler to just pull everything from the internets, since in general packages on the image will be superseded fairly quickly, using $ sudo fedup-cli --network 21 --product=nonproduct will pull all of the required packages, instigate a

www.tuxradar.com

reboot and then install them without any further user intervention required. If you want to your install to have everything, for example, the Workstation release has, then you can instead supply --product=workstation. Bear in mind that this will install Gnome in addition to any other desktops you're running, so you may want to tidy things up later.


Fedora 21

each other's toes, and it's nice to have options. OpenLMI has the potential to be much bigger than just Fedora, it uses Distributed Management Task Force standards to shuffle data back and forth, and works in tandem with (rather than replacing) standard system programs. Those already familiar with the arcanum probably have no use for OpenLMI (other than something at which to scoff derisively) but it will certainly ease the learning curve for those getting into the game. And so to the Desktop, sorry Workstation, edition, in which we discover... nothing particularly surprising. The install process is straightforward, Btrfs is not the default filesystem (Ext4 retains this honour and it probably will do so until Fedora 23) but it is exactly two clicks away if you are hungry for COW (Copy-on-write) [see Get The Tech Of 2015, p34, LXF194]. If you just go with the defaults, you'll end up with three partitions: boot, root and swap, and the latter two are

managed via LVM. While the system is installing you can set up users and set the root password, which is slightly more fun than watching the progress bar, but you'll still have time for the LXF-mandated cup of tea. The initial install occupies about 4GB, which is pretty standard nowadays.

Assisting developers The default desktop is Gnome 3.14, and comes with a Dreamworks-esque cloudy background, in the royal blue hue that has for many years been Fedora's colour. While Gnome 3.15 was released at the end of November, we wouldn't expect it to be included in this release. There are, of course, ways and means of upgrading, though. Application-wise, there is LibreOffice 4.3, Firefox 33, Shotwell, Rhythmbox, Transmission, all of which you will probably be familiar with. Perhaps less familiar will be DevAssistant, a Fedora project which started nearly two years

Installing the 173 packages that constitute Plasma 5.1 is easy with DNF.

ago. DevAssistant aims to take the hassle of out of programming projects – setting up development environments, installing dependencies and publishing your code – so that the programmer can concentrate on actually programming rather than fiddling. There are also some of the Gnome core apps, including the new cloudy Documents application (as well as the trusty Evince, it's desktop-bound counterpart), the not-exactly Google-threatening Maps, some kind of Instagram/Snapchat clone of dubious worth called Cheese, the venerable Brasero discburning application and a handy weather application. Gnome also supports DLNA media streaming, file sharing via WebDAV and screen sharing via the Vino VNC client. Gone is the old PackageKit frontend, having been replaced with the applicationfocused Software, which is certainly more tolerable than Ubuntu's Software Center. Around 50% of Fedora applications ship with the required AppData files (which provide XML-wrapped screenshots, descriptions, translations and ratings) required to display nicely in the application. Red Hat are encouraging upstream developers (rather than its own packaging team) to supply AppData files – the newly redesigned Gnome Software is no longer the preserve of Fedora users, with support now included for Arch, Debian and OpenSUSE. Further, as was mentioned in [Package Management: How Does That Work Then?, p51, LXF186], AppData is part of a movement for a panDesktop approach (which is being led by Fedora) to Software Centre type applications, so it will eventually be useful on KDE and other desktops too. Fedora also ships a (non-default) Gnome on Wayland session, which is refreshingly stable. Obviously not all applications support Wayland at this time, but these gracefully fall back to the Xwayland compatibility layer. While it's nice to know that Wayland is making progress, it (or at least the ecosystem it intends to support) is still not quite there yet, so this should be considered a preview more than anything else.

Spinning and wrangling Although you can customise the official release to your hearts content, you may prefer to start from one of a number of Fedora-endorsed 'Spins'. These are releases tailored for particular preferences or workloads. So you could try the KDE Spin (which uses the Plasma 4 Desktop), or any of the other desktop Spins, which include but are not limited to: Xfce, LXDE and MateCompiz. Besides desktops, there are Spins customised for scientists, robotics aficionados, gamers and the tinfoil hat brigade (a security-


focused release with forensics, intrusiondetection and network-sniffing tools). You can even make your own remixes, and if you feel it's worthwhile, submit it to Fedora for official endorsement, whereupon it will become a Spin. The process is well-documented at https://fedoraproject.org/wiki/Spins_ Process, but the general idea is to start from a pre-made kickstart configuration, which may come from an official release or Spin, and add (or subtract) whatever you see fit. Once you've

www.linuxformat.com

tidied everything up so that it conforms to Fedora's guidelines, you can submit it to the Spin Wrangler, who will decide if your Spin Page is complete and accurate, if so then your work travels up the river to the Spins SIG (Special Interest Group) who will decide if your work is worthy of passing to the Board who ultimately hold the rubber stamp. It seems like a lot of politics and bureaucracy, but you can't just let any old Thom, Rickert or Harriet sullying the Fedora brand with their half-baked wares.

Fedora 21

We did a grand old feature on package management [Again, see p51, LXF186] but poor ol’ Yum hardly got a mention. While it is in many ways comparable to Debian's Apt, it does have a few unique features. Yum automatically updates a local SQLite database of available packages. This means that there's no need for a separate upgrade command, so you should never run into the situation where the system attempts to install a package older than what is available on the repositories. Another notable feature is its use of delta-RPMs, which can save a huge amount of bandwidth by downloading diffs, where available, rather than complete packages. Yum is written in Python and is extensible via a plugin system, the diff functionality originated in a plugin called Presto, but it was so good that it ended up being mainlined into Yum back in Fedora 19. You list the officially endorsed plugins with: $ yum search yum-plugin This will reveal such plugin delights as fastestmirror (which will do exactly what it says on the tin), local (which maintains a local repository of all installed packages – useful if you're installing the same packages on several machines) and merge-conf (for dealing with changes to configuration files). Naturally, there are many more unofficial plugins available to use, which you are encouraged to explore.

New package manager Another notable feature of Yum is that it will, in Fedora 22, be replaced by DNF (some claim this is short for Dandified Yum, but the official response is that it doesn't stand for anything). Don't worry though, the syntax is much the same, and you can even install the dnf-yum package to provide a wrapper for DNF, so that

it is called instead of Yum. DNF has been shipped since Fedora 18, but only since Fedora 20 has it been considered ready for everyday use. DNF is a fork of Yum, with a much tidier codebase thanks in part to shipping out much of the functionality to a backend library called Hawkey. The main reasons for another package manager (besides the seemingly ineluctable desire to fork) are: Extensibility While plugins can be written for Yum, the API is not well-documented, and everything has to be done in Python. DNF (more correctly Hawkey) provides a clean API which can be accessed through C as well as Python, and hence plays much nicer with frontends such as Gnome Software. Dependency resolution While everyday users will not notice this, it is possible to lead Yum into a situation of unresolvable dependencies. DNF uses OpenSUSE's advanced libsolv, is much more adept at getting out of these dependency dilemmas. Speed From a consumer experience point of view, DNF also takes the bold step of synchronising metadata as a background service via a cron job, which means that: First, you'll see random spikes in your internet traffic and second, installing new packages

Gnome Software: the new design is like someone took the ‘in-your-face’ out of the Ubuntu Software Center.

streamline the process of packaging your software into an easily accessible repository. Prior to Copr, making unofficial Fedora packages involved using either Fedora's Koji Build System or the OpenSuse Build Service. Both of these are comprehensive and as such are rather heavyweight. Furthermore, since Koji is also how official Fedora packages get made (they don't come from storks and cornfields), anyone using it has to comply with its draconian and lengthy guidelines. So Copr is certainly not a replacement for Koji, in many ways, it is its opposite: an easy to use, lightweight and feature-full build system. For example, if you want to run Plasma 5.1, the latest KDE desktop, you first enable Daniel Vrátil's Copr repository and then let DNF take care of the rest: $ sudo dnf copr enable dvratil/plasma-5 $ sudo dnf install plasma5 You can install the KDE 5 Application suite, by adding the repository kde-applications-5 as above. While the underlying KDE 5 Frameworks libraries are already in the standard Fedora repos, Plasma and the applications are still considered unstable and need to be annexed since they conflict with all the stable KDE4 gubbins. To summarise, Fedora 21 is a grand old distribution triumvirate, which builds on a great legacy. The newly anointed trinity approach marks a bold step, though we tend to see it as a needed one, as each edition is tweaked for the relevant target audience: Server's do not need desktop environments, and workstations do not need to run LDAP servers. Some say the idea of a distribution is dying, with the future being all clouds and containers – be that as it may, these three distros, for now at least, are very much alive and kicking. LXF

“DevAssistant aims to take the hassle of out of programming projects.” and upgrading the system will feature about half as many progress bars than before. Besides being a generally slicker package manager, DNF also makes it even easier to use the new Copr repositories. These are analogous to Ubuntu's PPAs – helping to

Boxes makes setting up virtual machines easy. Also, we note that the default background is a crescent moon and a small boy with a fishing rod away from a lawsuit.

www.tuxradar.com


Dr Brown’s Administeria

Dr Brown’s Dr Chris Brown

Administeria

The Doctor provides Linux training, authoring and consultancy. He finds his PhD in particle physics to be of no help in this work at all.

Esoteric system administration goodness from the impenetrable bowels of the server room.

The 3M machine

M

y first exposure to a graphical desktop (and to BSD UNIX) was a Sun-2 workstation. I had one in my office. At that time (around 1985), the talk was of a '3M' machine – one million bytes of memory, one million instructions per second, and one million pixels on the screen. Sometimes a fourth M was suggested – the cost should not exceed one megacent ($10,000). My Sun-2 just about scraped in on all four M's. Fast forward to a current low-end laptop and memory is up by a factor of 4,000 (to 4GB). MIPS figures are hard to compare because today's processors have multiple cores and more complex instruction pipelines; however, for a humble Intel Core i3, 2,000 MIPS is not an unrealistic claim. And the price (around $250) is down by a factor of forty. But when we come to pixels, one million is still the going rate. The reason, of course, is that this is already quite close to the limits of visual acuity. My trusty Stanley tape tells me I'm sitting half a metre from my laptop screen, which is 340mm wide. Doing the math, I discover that my screen occupies 38 degrees (2,280 minutes) of my field of view. Now a human eye with '20/20' vision can discriminate two pixels separated by one minute of arc – so I should be able to resolve about 2,000 pixels across my screen. Significantly more than that, and I just won't see them. Many humans can do better than 20/20, and there might (just) be a case for so-called '4K' screens. But it would take a truly inspired marketing effort to persuade me to go any higher.

The Sun never sets Although the Sun finally set beneath an Oracleshaped horizon several years ago, its rays still light the Linux landscape.

I

'm in an historical frame of mind this month, pondering where some of the technology in Linux has come from. And you can't go far down that road without coming across Sun Microsystems. Sun is probably best known for its hardware; in particular its high-end multi-core servers based on its SPARC architecture and targeted at large-scale data centres. However, Sun has also contributed a huge amount to the world of open source software. Bill Joy (the only one of Sun's founders I have actually met) wrote the vi editor and the C-shell and contributed heavily to BSD Unix, especially the TCP/IP stack. Sun was the developer of NFS (Network File System), which is still the mainstay of filesharing across Unix and Linux networks. It gave us NIS and NIS+. Sun was the employer of James Gosling, who developed Java, and Sun eventually released it under the GPL licence. The company was also instrumental in introducing PAM (Pluggable Authentication Modules) in 1995; PAM remains central to all

The Apple Lisa Apple’s Lisa, introduced in 1983, took the ideas of the Xerox Alto and produced the very first personal computer system with a graphical desktop to be sold commercially. With just seven applications (lisawrite, lisadraw, lisacalc,


lisagraph, lisaproject, lisalist and lisaterminal) there was no doubt about its intended audience, but it wasn’t a commercial success for Apple, being (in hindsight) a stepping stone towards the Macintosh.

www.linuxformat.com

Linux authentication. Through acquisitions of other companies, Sun bought other important technologies, including MySQL, Open Office and VirtualBox. Sun also gave us ZFS, arguably the best file system out there. Though not lying directly on the ancestral family line of Linux, OpenSolaris (2005) represented a milestone in the open-sourcing of Unix. It was a blessing for those of us who wanted a free 'real' Unix that could run on ordinary PCs. Richard Stallman commented: "I think Sun has contributed more than any other company to the free software community in the form of software. It shows leadership. It's an example I hope others will follow". Unfortunately, after Oracle acquired Sun in 2010, it decided to discontinue open development of the core software, instead releasing a proprietary distribution called Solaris Express. Immediately prior to the closure of the Solaris source, the community forked the Illumos and OpenIndiana projects to keep the dream alive. (Roughly speaking, Illumos is the kernel project; OpenIndiana builds a distro on top of it.) Sun ceased to exist as a company after it was bought by Oracle, leaving many of us nervous about the future of many open-source products of which Sun had been the custodian.


The mouse's tale We trace the history of the computer mouse back to 1968 and the extraordinary demo by Douglas Engelbart.

H

ere's a question for you. Who invented windows, mice, menus and so on? No Googling please… What's that? Microsoft you say? Well, no. Microsoft is a marketeer par excellence, but (at the risk of opening the floodgates of flame) it has actually invented rather little. Well, perhaps Apple then? That's a little closer to the truth. Apple was certainly the first to put the technology in front of large numbers of users, initially with a machine called the Lisa which dates from 1983 and later with the MAC Classic in 1990. The Lisa sold for about $10,000 which put it within reach of some professional business users, and the Classic brought the price down to $1,000, making it truly a 'personal computer'. But the desktop metaphor, with its windows, icons, menus and mice, did not originate there. It turns out that Steve Jobs and his team at Apple were much inspired by a visit to the Palo Alto Research Center of Xerox in 1979 where they saw a machine called the Xerox Alto, which is generally regarded as the first computer to provide a graphical desktop. (By the way, Xerox PARC also invented Ethernet, which was just as important as the GUI stuff, but not relevant to our story here.) But we can trace the lineage of the mouse back even further than that, because the folks at Xerox PARC were, in turn, influenced by the work of a team headed by Douglas Engelbart. Engelbart worked at the Stanford Research Institute, and the name of his group is interesting – it was called the Augmentation Research Center (ARC). Engelbart was interested in moving the computer out of a pure numbercrunching role and making it "an instrument for helping humans to operate within the domain of complex information structures." I wrote a short story about what might happen if we let this go too far [See Administeria, p56 LXF192]. And central to our theme here, he is generally regarded as the inventor of the computer mouse. Let me shift gear a little. It may seem strange to say this, having grown up through a lifetime of intense technological development, but I sometimes feel disappointed at

One of Engelbart’s prototype mice. Apple fans note there’s only one button!

Video superposition of Engelbart and his computer screen during the Mother of All Demos in 1968.

having missed out on some of the key historical moments: to have turned on that first electric light bulb; to have witnessed the first transatlantic telegraph in operation; to have tuned in my crystal set and listened for the first time to that faint voice plucked by pure magic from the ether. And there's one other: to have attended the Mother of All Demos (as it has been called retrospectively) which Engelbart and his ARC colleagues presented on December 9, 1968 in San Francisco. The entire demo (minus the bits lost during reel changes!) is available on YouTube. The image quality is terrible, but if you can put up with that, you'll see demonstrated a folding editor presenting various hierarchical views of a document, the use of a mouse to select text, cut and paste, and a simple form of hyperlink. We even see Engelbart using the computer to show a hierarchical view of the presentation itself – and to put that into context, it's 22 years before PowerPoint came along. Engelbart has his co-worker turn the mouse upside down to demonstrate how it works; at that time it had two rotating wheels mounted at right angles. (It was Bill English who built the first mice and he didn't come up with the ball mouse until 1972. Optical tracking mice appeared much later, around 1980. They were much better because you didn't have to keep dismantling them to dislodge the hairballs and bits of Bombay Mix.) In the demo, Engelbart explains the concept of a cursor (he calls it a "tracking spot") that follows the mouse on the screen. He also showed a five-point touch device that lets you enter characters by playing “chords” with your fingers. The networking involved in the demo is impressive too – Engelbart is interacting in real-time with a computer that's actually 30 miles away in Menlo Park. Engelbart died in July 2013, and as when Dennis Ritchie passed away, the press were (for the most part) entirely unaware of a missed opportunity to write an obituary for a genius. Nowadays, with multitouch screens the order of the day, users have to do little more than make an artistically-inspired hand gesture in the general direction of the computer to make their wishes clear. But it is worth remembering that the whole thing began some 50 years ago when Douglas Engelbart placed two little wheels in a box at right angles, turned it upside down, and called it a mouse.

www.tuxradar.com



ClamAntiVirus Scanning for malware on Linux file servers and mail gateways can help prevent infected files from reaching Windows in the first place.

I

f you’re running Windows, virus protection is something you really can’t do without and most users, I suspect, fork out for a decent antivirus subscription. Though not a complete stranger to malware, Linux suffers to a much smaller extent. Analysts will generally point to two factors: First, the security model for Linux is stronger, making it much harder to write viruses. Second, the low market share of Linux makes it a less profitable target. And to these two I would add a third: Linux users are more cautious and streetwise when it comes to security. For example, they may choose to only install digitally signed software from their distributor’s repos. I don’t really know how true each of these three is; but one thing I think is clear: Linux doesn’t represent the ‘low hanging fruit’ that malware writers prefer to target. So is there any case at all for running antivirus software on Linux? Well, there probably isn’t much of a reason for performing on-access scanning of every executable file, but if your Linux system is carrying files that are destined to be consumed on Windows machines (for example if it’s a mail server or a Samba file server) then there’s certainly a case for scanning that content for viruses before it ever reaches the Windows systems. Clam AntiVirus is an open source (GPL 2 licensed) antivirus toolkit available for multiple platforms including Linux, Solaris, FreeBSD and Windows. ClamAV was owned by Sourcefire, which was acquired by Cisco in October 2013. It’s now actively maintained by the Talos Group, the Security Research and Intelligence group within Cisco. The website is http://clamav.net, and several mailing list archives are available at http://lists.clamav.net. The heart of the antivirus engine is a shared library; built on top of this is a command-line scanner, and a multi-threaded scanning daemon. One of its main uses is scanning incoming messages on mail gateways. I decided to take a look at ClamAV on Ubuntu 14.04. Installation is trivially easy, because it’s in the Ubuntu repos: $ sudo apt-get install clamav clamav-docs Dependency resolution also brings in packages clamavbase, clamav-freshclam and libclamav6. The main clamav package includes three executables: clamscan, sigtool and clambc; along with their

ClamAV can delve inside a wide variety of file formats in search of viruses.


www.linuxformat.com

corresponding man pages. Let’s dive in and try running a scan from the command line, using clamscan: $ sudo clamscan -i -r Training/ Training/XXX/tmp.tar: Trojan.Linux.RST.b FOUND Training/XXX/psybnc/whiper: Linux.RST.B-1 FOUND Training/XXX/psybnc/pico: Linux.RST.B-1 FOUND ----------- SCAN SUMMARY ----------Known viruses: 3720060 Engine version: 0.98.5 Scanned directories: 476 Scanned files: 3570 Infected files: 3 Data scanned: 1771.99 MB Data read: 1168.35 MB (ratio 1.52:1) Time: 277.079 sec (4 m 37 s) Here, -r is the recursive option and -i says to display a report only for infected files. I have edited the path names down to ‘XXX’ so they fit in our printed column, but the report itself is genuine. The three files listed here alarmed me at first, until I realised they were actually sample files from a Linux security course I used to teach. As you’ll notice, clamscan’s default behaviour is just to print the names of the infected files, but you can also ask it to ‘quarantine’ them by moving them into a specified directory, or to delete them. Just think about what we just did: one short command to install the tool and initialise the virus database, another short command to run the scan, and no suggestion of any money changing hands. Don’t you just love Linux?

Agent Clamd, licence to scan There is also graphical front-end called clamtk. It is basically just a wrapper around clamscan and doesn’t add any functionality, but it does look a little more like a Windowsbased scanner. ClamAV also makes its scanning engine available as a daemon (clamd), which is distributed as a separate package (clamav-daemon). By default, this daemon listens on a Unix-domain socket, so it can only be accessed from the local machine. However, you can configure it to listen on a TCP port instead, which opens up the possibility of running clamd as an ‘agent’ on all your machines, and controlling the scans from one central location. There is a client tool called clamdscan which will connect to clamd; however as far as I can tell there’s no way to tell it to connect to a remote daemon. For this you’ll need to hunt down a python program called clamdscan.py. The clamd daemon logs its actions to the file specified by the LogFile directive in its config file, /etc/clamav/clamd.conf. If you want to write a scanning tool from scratch, there’s a library (libclamav) which provides the actual scanning engine. The library is well-documented in the ClamAV user manual. The user manual includes mention of “on-access” scanning using the dazuko module: this is a third-party kernel module that intercepts file access calls and passes the file information to a user-space application. The idea is to support on-access virus scanning, file access logging or other external security tools. However, the official website at


http://dazuko.org has a forlorn look; the most recent release is almost four years ago and the project is currently unmaintained. There is, however, a FUSE-based user space filesystem for Linux called clamfs which provides on-access AV file-scanning through clamd (see Clamfs box, below).

The Virus Signature Database The quality of a virus-scanning tool depends on two factors: how up-to-date its virus definition database is, and the range of file formats it can peer inside. The virus definitions for ClamAV are primarily stored in two files (daily.cld and main.cvd) in the directory /var/lib/ clamav. The program freshclam can be run to manually update this database: $ sudo freshclam ClamAV update process started at Wed Dec 31 06:30:56 2014 main.cvd is up to date (version: 55, sigs: 2424225, f-level: 60, builder: neo) Downloading daily-19859.cdiff [100%] Downloading daily-19860.cdiff [100%] daily.cld updated (version: 19860, sigs: 1299145, f-level: 63, builder: neo) bytecode.cvd is up to date (version: 244, sigs: 44, f-level: 63, builder: dgoddard) Database updated (3723414 signatures) from db.local. clamav.net (IP: 193.1.193.64) You’ll notice that freshclam is downloading only the diffs, not the entire database. (That got downloaded as part of the original installation.) Freshclam can also be run as a daemon (freshclam -d) to update the database automatically. In fact, installing ClamAV onto Ubuntu automatically configures the freshclam daemon to run once an hour, with no further configuration or action on my part. A log of freshclam’s activity is maintained in /var/ log/clamav/freshclam, and a quick examination of this file shows that updates to the virus database are being added (and downloaded to my machine) roughly four times a day. The command sigtool can be used to examine virus signatures. For example, to list all the signatures in the daily. cld file: $ sigtool --list-sigs daily.cld Trojan.Hupigon-9863 Trojan.IRCBot-1971 Trojan.SdBot-8227 ... plus another 3.7 million ... Or to find specific signatures bases on a regex match, we might try something like this: $ sigtool --find-sigs=”Linux.*Worm” daily.cld As you might expect for an open-source project, you can contribute by submitting your own virus signatures if you find

something that ClamAV doesn’t already recognise. Just email them to [email protected]. Typically, virus signatures won’t just show up ‘in the clear’ in the files on your system. More likely, they’ll be lurking in files that have been packaged in some way (zipped archives for example, or MSI files built for the Microsoft installer), or have been compressed using something like Gzip. Windows PE (Portable Executable) files – a very common target for viruses – are commonly compressed with tools, such as UPX or Petite, or are deliberately obfuscated to hide them from the eyes of virus scanners using something like Yoda’s Crypter. To help ClamAV find viruses in as many places as possible, it can delve inside a truly enormous range of file formats ranging from TAR, CPIO, GZIP and BZIP2 archives to PDFs, mailbox files, Windows Cabinet files, and a host of compressed or obfuscated PE formats. The package clamav-testfiles includes a total of 44 files of different types which have all been ‘infected’ with the virus signature ClamAV-Test-File and can be used as a basis for testing. Not surprisingly, clamscan found all of them. (I tried copying these files across to a Windows system and scanned them with Norton 360 premier edition. It found no threats, presumably because the ClamAV-Test-File virus signature is not in its database.) The library also includes a ‘data loss protection’ module which can detect credit card numbers from the major credit card issuers, as well as U.S. social security numbers inside text files, though I haven’t tested this.

The original Trojan horse, which gave rise to the quote “beware of geeks bearing GIFs”. Oh yes, I did.

Filtering mail As mentioned, one of the key uses of ClamAV is scanning the messages received by a mail gateway. This is made much easier by a technology called milter (mail filter). Basically, milter is a set of hooks that can be used by mail transfer agents, such as Postfix and Sendmail to interact with an external virus scanner or other filter program at various points during mail message delivery. The package clamavmilter provides the necessary filter; installing the package automatically configures Sendmail to use it, but for detailed instructions on setting this up under Ubuntu, see http://bit. ly/UbuntuMailFiltering. In fact, clamav-milter is just one of many available milters – see www.milter.org for a list. LXF

Clamfs Clamfs is another tool built around ClamAV. It implements a user-space filesystem in which all your file accesses are vetted by clamfs, which in turn connects to the clamd daemon to perform a virus scan on the file. I got this to work, despite the absence of proper documentation. Here are the steps: 1 Install the clamfs package: $ sudo apt-get install clamfs

2 Copy the sample config file somewhere sensible, and unzip it: $ cd /etc/clamav $ sudo cp /usr/share/doc/clamfs/clamfssample.xml.gz clamfs.xml.gz $ sudo gunzip clamfs.xml.gz 3 Tweak the config file. The only line I actually changed was the node, to define the pathname of the file system I wanted to mount, and the

www.tuxradar.com

directory where I wanted to mount it. 4. Start the clamfs daemon, specifying the path to the config file: $ sudo clamfs /etc/clamav/clamfs.xml At this point I have a mount in place of type fuse.clamfs; any files I access under this mount point will be scanned when I access them. If a signature match is found, the operation is not permitted, and an email is sent to root.


The best new open source software on the planet Alexander Tolstoy handpicks the finest chunks of open source gold to be melted down and hammered into shape for this month’s HotPicks showcase.

Yarock FFmpeg Dropbox Uploader KWave Linuxbrew SuperTuxKart

Gloobus-preview MPS-Youtube Aualé Hollywood Boomaga

Music player

Yarock Version: 1.0.0 Web: https://launchpad.net/yarock

E

ven though there are plenty of Linux music players, new ones just keep coming out. Yarock is another new kid on the block, recently grown to a 1.0.0 version after almost five years as a series of 0.x releases. The player's most distinctive feature, as mentioned on the official website, is an "easy and pretty music collection browser based on cover art". Besides that, Yarock is also a Qt4based application with a stylish exterior, which was revamped for the 1.0.0 release with new UI icons and some layout changes. These changes include the playback controls being relocated at the bottom, and the category pane

on the left side of player's window now housing more items. The player's interface has three parts. The left pane offers subcategories (Home, Music, Playlist and Radio browsers, and also local folders) while the central part shows where you are, and the right pane is used for exploring and possibly changing track metadata. The details may not thrill veteran users up to this point, but

A smart and fast music player written in pure C++ and bound to Qt4 and the Phonon multimedia framework.

“Sorting your music collection in Yarock has rich options.”

Exploring the Yarock interface Navigation Just as you find in a web browser, Yarock’s upper panel shows where you are and houses various optional buttons that prompt you to take some decision.

Feature list The left pane always shows the various sources of music and the different browsing modes available.

Main view The current selection is usually displayed here as cover art. This is the Settings view showing off the great customisation available for lyrics, scrobbling and more.


Reserved for a track

Now playing controls

When something is playing, you find more details here. This pane is also for arranging items into playlists.

Play, pause, stop, skip back or forth here. Extra buttons on the sides stand for volume control, equaliser and shuffle modes.

www.linuxformat.com

Yarock's strengths are revealed when you start using the player for enjoying your music collection. Yarock offers fast indexing of files into the SQlite3 database and enables you to manually fetch missing covers from the web. You don't need to drag a song into a playlist – though you can do it – as Yarock plays your tracks directly from collection. Sorting your music collection in Yarock has rich options: by albums, artists, songs, genres, years and even folders. Additionally, when you've listened to a few tracks, Yarock can generate smart playlist based on the playback history. Could Yarock become your player of choice? Well, it depends. The application is robust, and doesn't have either KDE or Gnome dependencies, and provides command line interface for all you Bash fans. It's suitable if you're looking for a player with simple play queue, favourites, tag editor, volume normaliser and other essentials. By the way, Yarock is desktopindependent – its supported formats list is defined by your Phonon backend. In most cases this would pass you down to the GStreamer backend, so take care of the codecs that you have.

LXFHotPicks Multimedia libraries and programs set

FFmpeg Version: 2.5.2 Web: http://ffmpeg.org

T

he FFmpeg project is a mature and well-recognised one in free software circles that produces libraries and programs for handling multimedia data. FFmpeg is the backend of many popular media players, such as VLC, MPlayer, Handbrake Converter and many more. FFmpeg is even used by YouTube on the server side and in Google Chrome locally to handle HTML5 video and audio data. The new version brings many improvements and now supports UDP Lite protocol, which enables playback of broken network bits by attempting to restore missing data by local FFmpeg decoder. FFmpeg has also started to support animated WebP and APNG images, as well as multithreading (thanks to merging with ffmpeg-mt fork), new demuxers and muxers, including HEVC/H.265 RTP container and support for fragmented MPEG-

DASH streams. This feature means that a video stream is divided into fragments of certain size, and if network bandwidth occasionally shrinks, an FFmpeg-based player can automatically change quality to a lower bitrate, without interrupting playback. FFmpeg also features nearly all improvements of libav, the concurrent fork, again thanks to solid backporting work. Of course, it isn’t the only choice for enabling multimedia support in Linux, but it does boast very high code quality. The project analyses all incoming patches using its own regression test suite called FATE (which stands for FFmpeg Automated Testing Environment), so that if any patch

If you expect your player to use the latest FFmpeg, mind the installation prefix and define in ./configure command.

“FFmpeg also features nearly all improvements of libav.”

created slows things down more than 0.1%, it gets dismissed. The project website offers very detailed descriptions about every major library that comes with the whole bundle, and you can also download the source code there. In order to compile the code, you'll need some prerequisites, such as Yasmans standard devel-stack for building software on your machine. Some FFmpeg components are optional, and the ./configure script will report which of them will be included in build. The new features and additions to the 2.5.x series, make this a must-have.

Dropbox CLI app

DB Uploader Version: 0.14 Web: http://bit.ly/Dropbox-Uploader

S

ome time ago, the Dropbox API was changed in a way that made it impossible to perform basic actions with files and folders from the command line. For example, for the sake of better security, public links (of shared files in the Public folder) were changed so they could only be created after communicating with the cloud, which returned the link with a unique server-side calculated hash value. There's also currently no service menu for the likes of Dolphin, Thunar and other file managers. Of course, users could have built those menus themselves if they could communicate with Dropbox cloud directly. Thanks to Andrea Fabrizi, this is now possible (again) with the use of his gorgeous Dropbox Uploader. This is a Bash script, which restores the CLI interaction with Dropbox by adding itself as a third-party app and connecting to the service.

Download the master ZIP archive from the project's Github page, extract it and run ./dropbox_uploader.sh. The first time the script is launched, it will guide you through the setup process. There's very good built-in documentation with all necessary details. The script will explain the steps to register the script as a third-party app, to which Dropbox grants access for working with shared files. After you paste the newly created app key and secret to the script's prompt, the initial setup is finished. The uploader doesn't require any extra authentication and thus doesn't store any sensitive data. It works with the official Dropbox API

The first time your run the script, a helpful wizard will guide your through initial setup.

“This is a Bash script, which restores the CLI interaction with Dropbox.” www.tuxradar.com

and enables copying, moving, deleting and renaming files within your account, as well as uploading, downloading and sharing files. For instance, getting a public link for a file is as simple as: ./dropbox_uploader.sh share Public/ My_file.odt The application uses the ~/Dropbox folder as a root directory, you don't need to provide full path. If a file exists, it’s synced and you'll receive a valid link from Dropbox followed by a success notification. The uploader script doesn't depend on the official client either, so it's even more cross-platform.


LXFHotPicks Quick preview tool

Gloobus-preview Version: 0.4.5 Web: http://bit.ly/GloobusPreview

T

his is a well-known quick file previewer, designed primarily for GTK-based desktop environments, but suitable for almost any desktop. The application supports images, documents (PDF, ODF and ODS, etc), audio (MP3, WAV and OGG), video (AVI, OGG, MKV and FLV, etc), folders, archives, fonts, plain text files and more. In Gnome, Unity or Cinnamon, Gloobus-preview integrates with the Nautilus or Nemo file manager (you press Space to preview), while in KDE you can manually create a simple servicemenu, which would open a file in quick previewer using: gloobus-preview %f command. A previous version of Gloobuspreview dates back to late 2009, and since then the application has lost compatibility with modern Linux distributions. However, a new developer, György Balló, joined the project recently

and brought some major updates to the code. He's ported the user interface to Pygobject3 and GTK3; the media framework to GStreamer 1.0; and media keys to GDBus. The list of supported file formats is also greatly enhanced. The old icns engine has been superseded with ImageMagick, and XPS format support has been added and the overall office plugin capabilities are improved by using ssconvert (derived from Gnumeric), while bsdtar improves archives support. Gloobus-preview isn’t a loner in its class. Various competing projects exist, from Gnome Sushi and Nemo Preview to Klook, but Gloobus-preview has the widest list of supported file types and

The preview feels very nice in any environment and works like a lightweight multi-purpose reader and player.

“Has the widest list of supported file types and looks very polished.”

also looks a very polished product and works smoothly. Installing the new version in Ubuntu and its derivatives is quite simple thanks to the dedicated PPA (http:// bit.ly/WebUpd8PPA). On other systems you may want to either convert Ubuntu's Deb package using Alien (it worked wonderfully for OpenSUSE) or build the previewer from source. In the latter case make sure you're working with the latest branch (use the URL in the strap), not the outdated 4.1 one.

Youtube CLI app

MPS-Youtube Version: 0.2.1 Web: http://bit.ly/MPS-Youtube

W

ith MPS-Youtube we continue our series of command-line tools that enable easy control of your favourite apps and services. MPS-Youtube was planned as an audio player and track downloader for YouTube, but it was soon equipped with a video playback feature. The project is based on MPS, a terminal-based program to search, stream and download music. As the name suggests, MPS is bound with YouTube as a source of music and videos, and the app has decent support of YouTube features, too. It can retrieve metadata (view count, duration, rating, author, thumbnail and keywords), create and save local playlists, search and import YouTube playlists, view comments for videos and more. All this is possible through the Pafy library (http://bit.ly/PafyLib), which is already included with MPS-Youtube.


The application is packaged as a Python module (both 2.7 and 3.x series are supported) and can be easily installed using Python's pip catalogue. Make sure you have python-pip (or similarly named) package in your system and issue the following: sudo pip install mps-youtube Once installed, launch it with the mpsyt command. By default MPSYoutube doesn't have video playback enabled, so let's fix it: set show_video true And another prerequisite for enabling video search feature is (only music is used for searching by default): set search_music false

This CLI-based app could be a replacement for your browser, offering both Flash plugin and YouTube extensions.

“MPS-Youtuber: Search, stream and download music and video.” www.linuxformat.com

Now you can search for a YouTube video, using a dot or slash sign followed by a search string. For example: .beatles cover The application will return a table of search results, with two columns: a number and a corresponding name. Enter the desired number and press Enter to start watching. You can change the default video player, MPlayer, to another, for instance, MPV with: set player mpv To return to the search results, close the video window and press Ctrl+C in the shell. You can easily download best audio or best video version of a track by using da # or dv # respectively.

LXFHotPicks Sound editor

KWave Version: 0.8.99-2 Web: http://bit.ly/KwaveApp

K

Wave is a sound wave editor, built with an Qt4 interface and thus fits nicely with KDE and any other desktop environment for that matter. Handling wave sound data in Linux is considered a niche occupation, and there’s a lack of high-quality editors. Audacity seems to be the most powerful solution, but sometimes you need a more lightweight application, which can offer a bit more than the legacy kde3-krecord. KWave meets these conditions just fine. It began life back in 2004 and nowadays is being updated every few months. KWave can work with Pulseaudio, ALSA and even OSS sound system, which can be set in Settings > Playback. After you've done that you can start recording by pressing the red button. A new window with recording settings will appear, and you can change the sound quality, pre-define recording timeslot, input device and more.

The new recording will appear in a new tab in KWave. Under the FX menu you'll also find some essential effects, such as Normalise, Low and Band pass filters, Pitch Shift and Notch filter. Under the Calculate menu you can add extra noise to your recording, or convert the selected wave area into silence. The wave stripe can be zoomed in and out, plus you can select part of it and export to a file. KWave supports multiple track handling with all necessary features for making multitrack composition. You can then export your work into OGG, FLAC, WAV or even MP3. The new version fixes a lot of past stability issues, adds new translations

If you are looking for an Audacity equivalent for KDE, you’ve probably found the best match.

“KWave can work with Pulseaudio, ALSA and even OSS.”

and some new command line arguments. For instance, it’s possible to quickly load the desired record using something like: kwave --iconic --disable-splashscreen test.wav KWave also supports remote usage through methods, such as kwave:plugin%3Aexecute?normalize, kwave:save KWave is included in the Ubuntu apps directory, though sometimes it's not updated there to the latest version. The official KWave page (see above) offers the latest version for most of RPM-based distros, and the source code.

Package manager

Linuxbrew Version: rolling dev Web: http://brew.sh/linuxbrew

A

nyone who might have tried Mac OS X will be familiar with the Homebrew package system. While package management is a bit unusual for Apple users, it’s a keystone system feature in Linux. However, Homebrew greatly differs both from Macports (BSD style) and Apt and Yum etc for Linux. Linuxbrew is a fork of Homebrew for Linux, and it shares nearly all the peculiar features of its parent, so lets cover the basics. Linuxbrew is a Ruby script that enables access to an online repos of software, particularly system tools, Ruby on Rails packages and related extras. If you're not a Ruby developer, this will open a whole new world of software at http://braumeister.org. To install the Linuxbrew client you'll first need to check that you have some basic programming stuff installed (curl, git, m4, ruby, texinfo, bzip2-devel, curl-

devel, expat-devel, ncurses-devel and zlib-devel). When you're ready, issue the installation command git clone https://github.com/ Homebrew/linuxbrew.git ~/.linuxbrew and add the following to the .bashrc file: export PATH="$HOME/.linuxbrew/ bin:$PATH" export MANPATH="$HOME/. linuxbrew/share/man:$MANPATH" export INFOPATH="$HOME/. linuxbrew/share/info:$INFOPATH" After that installing any package is as simple as: brew install package_name Linuxbrew doesn't need root authentication or sudo because it puts

Browse repositories online or use the brew search command to find the right package.

“If you’re not a Ruby developer, this will open a new world of software.” www.tuxradar.com

all installed files in your home folder at ~/.linuxbrew. The Linuxbrew solution might not seem very germane for outdated or aging LTS Linux distros, which no longer provide fresh versions of development tools. But the choice is far richer than just development armoury. Homebrew houses many repositories, including games with many popular titles (Freeciv, SuperTux etc). The application database is also the same for both Homebrew and Linuxbrew, so it’s very ease to sync your development stack between OS X and Linux, if needed.


LXFHotPicks HotGames Entertainment apps Race simulator

SuperTuxKart Version: 0.8.2b Web: http://bit.ly/SuperTK

S

uperTuxKart is a kart racing game originally started in 2004 as a free and open clone of Nintendo's Mario Kart. Since then the game has absorbed new features and technologies, as well as lovely details from many proprietary kart simulators, particularly Moorhuhn Kart. The game features many popular mascots from the open source world, including Tux, Beastie. Wilber, Pidgin, Konqi and the XFCE Rat. Each can be selected as your driver, while others will be your competitors. The game referee is Thunderbird, by the way. The last stable version of SupertuxKart is 0.8.1, which dates back to November 2013, and there wouldn't be much sense in writing about it here in HotPicks unless

some important changes hadn't landed recently in the game's Git repository. And indeed, they are ground-breaking. The latest unstable 0.8.2b version uses the new Antarctica graphics engine, which supports 100% dynamic light/shadows, volumetric fog, depth of field, different model maps (glossmap/ specular map/normal map/emit map) and requires an OpenGL 3.1-capable video chip to unleash all these benefits. There's also a new shader-based rendering engine, physics, racing tracks, Wii support, online accounts support and a long-awaited multiplayer mode.

This beloved open source kart simulator has returned with renewed vigour and lots of graphics enhancements.

“There's a new shaderbased rendering engine, physics and tracks.”

The game's settings enable you to change the graphics quality, but keep in mind that in order to use the top-grade settings (with dynamic lighting etc) you've got to have the appropriate graphics. Intel's HD3000 is an official minimum requirement, but the best results can only be achieved with Nvidia or AMD high-end cards using proprietary graphics drivers.

Board game

Aualé Version: 1.0 Web: www.joansala.com/auale

A

ualé is a strategy board game with a long history. The name is a Catalan translation of Mancala, which is a bigger family of board games from Africa and Middle East. The main principle is to collect seeds from one hole on your side of the board and then distribute (sow) them into the holes counterclockwise; putting one seed into each hole. In certain conditions you can grab seeds from your opponent’s holes and store them in your barn. Aualé has each player start with six holes, and there’s a total of 48 game pieces, or seeds. According to the rules, you can grab seeds for your barn when a hole belonging to your opponent contains two or three seeds. The player who collects 24 seeds first wins the game – and Aualé turns out to be quite hard to


play, even against the computer, which provides very basic artificial intelligence. The gameplay requires you to constantly re-evaluate risks and possibilities and, of course, accurately count the number of seeds in each hole. Both players try to reach a situation where an opponent runs out of seeds on his side and thus doesn’t have enough space to manoeuvre. In this case it is quite easy to put the squeeze on the opponent and win. Apparently the long game tends to require well thought out positional moves, where players concentrate seeds in a few holes and redistribute them carefully.

After getting energised with racing, we spend a couple of hours with seeds and holes. This really calms us down.

“Aualé turns out to be quite hard to play, even against the computer.” www.linuxformat.com

Aualé lets you choose the strength of the computer player. You can also save a match and continue it later, and roll back and forth recent moves. The game’s website provides binaries for many popular Linux distros, both Deb and RPM-based. But the inside of a package shows that Aualé is a noarch game, consisting of Python scripts and bindings and some platformindependent JAR pieces. If you like logic games, Aualé is a strong choice.

LXFHotPicks Activity simulator

Hollywood Version: 1.8 Web: http://bit.ly/HWTechMelodrama

T

here's a rare category of applications that are entertaining, delivering pure fun and joy, but at the same time are nearly useless from a practical perspective. Dustin Kirkland and Case Cook, who both work at Canonical, found a few spare hours to bring us such a totally useless package, but it's utterly brilliant one for the Ubuntu console: it's called Hollywood Technical Melodrama. But what is it? This package turns your Ubuntu console into a Hollywood technical melodrama hacker interface, blending the 'real-life hacking' seen in classics like the 1995 Hackers movie, and the kind of rolling text and graphics that you tend to see on huge screens, whenever a generic top secret control centre is required in an action movies. We doubt that the NSA or GCHQ bother sliding vast marble floors back to reveal huge maps every time they run a mission (the healthy and safety issues

alone would be a nightmare), but of course, most of what's displayed on the vast screens in such movie scenes makes little or no sense, but who cares if it manages to convey the right mood of covert and usually, evil enterprise. The Hollywood package supports all recent versions starting from 12.04, as well as other compatible *buntu flavours and derivatives. If you're not on Ubuntu, you can still download the tar. gz package manually (http://bit. ly/1Aig3x2), extract it and run. The applications contains no binaries, but has a large number of dependencies, some of which are familiar command-line tools. The Hollywood Technical Melodrama relies

This screenshot is reserved for the movie about Mir development.

“Delivering pure fun and joy, but at the same time nearly useless...”

on a Byobu, a text-based window manager and terminal multiplexor, which creates a tmux session and splits the main window into parts. Then different commands output are shown there. For extra pizzazz, the host script changes the Byobu layout a few times a minute. In order to run Hollywood smoothly, you'll need htop, MPlayer, mlocate, ccze and some other utilities. To guess what's missing, simply examine the splits for errors and complaints. Fully charged Hollywood should flicker with a Matrix effect, colourised text, logs and even an MP4 file, played with the ASCII video output. Of course, CPU load will be quite high, so take care of battery drain when showing off your leet hacking skillz to everyone in your local coffee shop..

Virtual printer

Boomaga Version: 0.6.2 Web: www.boomaga.org

T

his software will be a great relief for anyone who ever wants to print out a booklet in Linux. Normally, most Linux text processors and graphics applications allow standard print settings with page selection and access to printer driver options, and it's more than enough unless your task is more specific. For example, you may want to place two pages in a single-paper sheet and create a two-up saddle stitch booklet, where the first sheet will have the first and the last page, the next will have the second and the last but one page and so on. And all this should be printed double-sided. In the old days only Pagemaker (and later InDesign) offered the Booklet plug-in. Sometimes this feature was included in the printer driver for some models. Boomaga (BOOklet MAnager) does this job gracefully. If you run it as a

standalone application, you'll have to prepare your PDF file beforehand, as it's the only supported file format. However, Boomaga also adds itself to the list of your printers, so you can use it from any application, which supports printing. After choosing to print to Boomaga, the program will automatically generate a PDF and show it loaded in the Boomaga main window. Here you can sort pages the way you need them to be placed on paper. The layout can be oneup (plain order), two-, four- and eight-up or booklet (for a saddle-stitch after folding sheets in two). After choosing your layout variant you can select the actual physical printer and access its

Rearrange pages and print out a book, as a real DTP professional.

“A great relief for anyone who wants to print out a booklet in Linux.” www.tuxradar.com

options by pressing the Configure button. Options vary depending on the printer model you have. If your printer doesn't have duplex printing feature, you may want to enable it in the Boomaga Layout settings. In this case the application will ask you to turn over the pages halfway through printing your document. The applications places all printing jobs in its own queue, enabling you to export each job as PDF (very helpful in pre-printing), renaming jobs, rotating entire jobs and more. The project's official download page (above) contains instructions for installing Boomaga. Luckily, packages for most mainstream Linux distributions are available. LXF


Back issues Missed one?


Issue 194 February 2015

Issue 193 January 2015

Issue 192 Christmas 2014

Product code: LXFDB0194



In the magazinee

In the magazine

In the magazine

What’s the Next Big Thing in Linux OS tech? We show the hot stuff you’ll want to try. Bored of your default desktop? ? Take your pick of our alternatives. Plus, cake for everyone! Firefox celebrates 10 years.

Create a multimedia hub for your home and stream films, music and photos around the house. Try out next-gen filesystems for a RAID array, mod Minetest and simplify your firewalls and so much more!

More power! Charge up your distro with essential tips and tools. Build a robot and a monstrous 24TB NAS box. Plus: system recovery, Linux certs and our pick of the most productive desktops.

LXFDVD highlights Fedora 21 Workstation, Manjaro, ALT Linux, 4MLinux and more!

LXFDVD highlights Ubuntu 14.10, OpenSUSE 13.2 and XBMCbuntu 13.2 and more.

Issue 191 December 2014

Issue 190 November 2014

Issue 189 October 2014




In the magazine

In the magazine

In the magazine

Take your Raspberry Pi mastery to the next level with our hot hacks. Learn how to contain everything with Docker and plug in to professional audio production using JACK. Plus: Develop with PHP.

Origin of the distro – LXF tracks the source of your favourite distro and picks the best of each genus. Plus: we chart Ubuntu’s bumpy history as it celebrates 10 years. Also, Pi alternatives and the best web browsers.

Discover how to solve your Linux problems with a huge troubleshooting guide. Plus: Chromebook roundup, run Linux on almost any Android device, cryptography explained and Minecraft hacking.

LXFDVD highlights Hot Pi distros! Kali Linux, Jasper, RetroPie, Pi MusicBox and more.

LXFDVD highlights Tails 1.1 Live DVD, Deepin 2014 and 3 essential rescue distros.

LXFDVD highlights Ubuntu 14.10 Remix (pick from 5 desktops), ROSA, Rescatux.

LXFDVD highlights Tails 1.1 Live DVD, Deepin 2014 and 3 essential rescue distros.

To order, visit www.myfavouritemagazines.co.uk Select Computer from the all Magazines list and then select Linux Format.

Or call the back issues hotline on 0844 848 2852 or +44 1604 251045 for overseas orders.

Quote the issue code shown above and have your credit or debit card details ready

GET OUR DIGITAL EDITION! SUBSCRIBE TODAY AND GET 2 FREE ISSUES*

Available on your device now

*Free Trial not available on Zinio.

FULLY REVISED & UPDATED FOR 2015

180 PACKED PAGES! THE COMPLETE GUIDE TO SETTING UP AND TRADING ONLINE

DISCOVER HOW TO: CREATE A COMPANY • SELL PRODUCTS ONLINE • MARKET ON SOCIAL MEDIA

Available at aall go good oo od dn newsagents ewsagent nts orr vvisit

Tutorial Xxxx Minecraft/Pi Unleash the artistic turtle

inside you to make trippy trees

Minecraft/Pi: Turtle graphics Jonni Bidwell once again fires up Minecraft on the Raspberry Pi, and plays with the graphical power that resides within a humble turtle.

Our expert Jonni Bidwell learned that in North America the world ‘turtle’ refers to Tortoises, Terrapins and Turtles — the whole chelonid family. Whereas in the UK, the meaning is much more specific.

Y

Quick tip You can also use LibreLogo to do turtle graphics in LibreOffice. Go to View>Toolbars>Logo. Yet another avenue of unneeded procrastination.

ou may be old enough, while still possessing the necessary memorial faculties, to recall programming Turtle Graphics in a language which called itself Logo. Actually, there have been several languages calling themselves thus, the first one appeared in the late 60s. Be that as it may, through the captivating medium of controlling a pen-wielding turtle (robotic or screen-abstracted), users could make shapes and patterns using diversely coloured lines by issuing simple directional commands. The turtle had a head (or at least a tail) and these commands (forward, backwards, left 20 degrees etc) were all taken to be relative to our critter's current bearing, not based on an absolute screen-defined compass. Students were encouraged to learn by ‘body-syntonic’ reasoning – imagining themselves as the obedient reptile in order to better understand the effect of any given command. Once the basics had been mastered, one could then learn all about functions and parameter passing, so that you could draw a regular ununpentagon (which would be a 115-sided figure if it was a real thing) of any given side length with a single command. Python comes with its own turtle graphics module, which uses the lightweight Tk toolkit to handle the graphics. If you're


www.linuxformat.com

using Raspbian you already will have it installed, and if you've not experienced the joy of being a turtle that draws, then you should definitely have a play with it – check the docs at https://docs.python.org/2/library/turtle.html . However, for this tutorial we're bringing Turtle graphics to Minecraft, so we're going to make our own Turtle implementation. One in which our chelonian plotter can move in not two but three dimensions. Did I say "make our own"? That's not really true, the work has all been done by our good friend Martin O' Hanlon, do check out his website www.stuffaboutcode.com for all manner of Minecraft: Pi Edition projects. We've already 'borrowed' Martin's stuff before, so you might have already seen his fantastically handy MinecraftDrawing class in previous tutorials, [see Tutorials, p84, LXF186]. The class provides functions for drawing arbitrary points, lines, faces, circles and spheres all over the Minecraft world, and forms the basis for our mcpiturtle project. Before we begin, it's worth explaining some of the complications that arise from letting our turtle move in three dimensions. In the 2D case, we really have only one angle with which to concern ourself: how many degrees to turn left or right from the current heading. Moving to three dimensions grants us another heading we can adjust: the elevatory angle. Thus our turtle will have four different rotation functions, as well as the two translation ones (forward() and backward()).

The default representation for our turtle is a diamond block that’s because our turtle is hard.

Minecraft/Pi Tutorial

Inspired spirals Turtle graphics make easy work of drawing some shapes which are quite hard to construct objectively. Consider a spiral in two dimensions: Each coordinate is given by a parametric equation, of the form x=a.t.cos(t), y=a.t.sin(t) where t is a parameter that varies from 0 to infinity (or until you get bored of drawing spirals). But turtles don't give a damn about trigonometry, if they want to trace out a spiral, they need only follow the following recipe: for step in range(100): t.forward(step // 2) t.right(45) Okay, so that's not quite as curvey as a spiral ought to be but it's a much less complicated method. Here's another approach to try, which constructs a spiral from the outside in by rotating and scaling squares: def square(length):

for k in range(4): t.forward(length) t.left(90) def sqr_spiral(): length = 50 for j in range(100): square(t, length) t.left(10) length = int(0.9 * length) Calling the second function, with everything set up correctly, results in the much more impressive result shown. By adding another dimension things can get even more impressive: extruding a circular path (a degenerate case of a spiral, parametrically speaking) results in a helix, so having two oppositely sensed circles on top of each other will result in a double helix. Thus by pairing four colours you could fill the Minecraft world with DNA.

We will keep track of the two different headings via the turtle object's heading and verticalheading properties. It can be confusing at first, but once you've played around a little it will become second nature. Besides the six commands for relative rotation and movement, our mcpiturtle object accepts commands for setting absolute position and headings: setx() etc for setting individual co-ordinates, setposition() for setting all three, setheading() and setverticalheading() for changing the angles. While these weren't part of the original language (they don't make sense if you're a robot liberated from screendefined co-ordinates), they can be very useful for getting to where you need to be in Minecraft world. We also have penup() and pendown() to control whether or not our turtle will draw anything, and penblock() to control the block with which the turtle will draw (the default is black wool). We can also choose the speed of our turtle, by calling speed() with a number from 0 to 10. With 10 being very fast and 1 being very slow, and setting the speed to 0 will draw lines instantly.

keep the y-coordinate constant, so that turtle will fly above valleys and tunnel through mountains. Once you've got Minecraft and mcpiturtle.py installed [see bottom of p73] we can begin to explore the power of the turtle. With a Minecraft world loaded, open up an LXTerminal, cd to the directory where you copied mcpiturtle.py, and start Python (you can also use the IDLE development environment if you prefer). Import our module and set up the connection to the game server with: import mcpiturtle.py import mcpi.minecraft as minecraft import mcpi.block as block

Turtle wrangling Being a turtle in a 3D world full of blocky hills, lakes and other features comes with its own challenges. Suppose for a second you are that turtle and are told to walk forward 50 blocks, but right in front of you is a great mountain. As an obedient creature, how ought you proceed? You could follow the level of the ground, up and possibly over the mountain, or you could ignore the scenery and tunnel mercilessly through it. A similar dilemma applies if you stand before a gaping chasm. These two alternative means of tackling a varying ground level are chosen by setting the turtle (you can stop imagining its you now if you want) to walk() or fly() mode. Walk will respect the ground level, so that the y-coordinate of the turtle and any tracks it lays will vary. Fly (the default) will

These psychedelic trees can be yours with just a few lines of code.

Get print and digital subs See www.myfavouritemagazines.co.uk/linsubs www.tuxradar.com


Tutorial Minecraft/Pi

Quick tip If you're turtle is travelling in the negative y direction (what the vulgar call 'down') then the diamond block which represents it will erase it's previous mark. You are free to fix this bug by caching and restoring blocks as the turtle moves.

mc = minecraft.Minecraft.create() And instantiate a new turtle object at Steve's current position: >>> pos = mc.player.getTilePos() >>> t = mcpiturtle.MinecraftTurtle(mc,pos) Behold, a block of diamond appears at your current position. This is actually our turtle, if you want to change its appearance then you can modify the turtleblock attribute. For example, to change it into a block of stone: >>> t.turtleblock = block.Block(block.STONE.id) Once you're satisfied with the appearance of your turtle, we can begin some artistic, or otherwise, endeavour. This code will draw the edges of a square-based pyramid: >>> for j in range(4): >>> t.forward(20) >>> t.right(90) >>> t.right(135) >>> t.up(52) >>> t.forward(23) >>> t.down(104) >>> t.forward(23) >>> t.up(52) >>> t.right(135) >>> t.forward(20) >>> t.right(135) >>> t.up(52) >>> t.forward(23) >>> t.down(104) >>> t.forward(23) Note that our turtle draws underneath its current position. This gives rise to an interesting bug (see the Quick tip, left). For now, enjoy this example wherein we have created a wireframe pyramid. If you fill in the gaps (or get several thousand slaves to do it for you) then you would have a fitting tribute to Pharaoh Stevehotep of Minecraftland (or something like the illustration pictured, below). All very well, but fairly tedious (and you didn't even have to work out the geometry), and not really as awe-inspiring as the one at Giza, (even though the proportions are close, the Great Pyramid has a base angle of just under 52 degrees). Here's a slightly more colourful example, this time we'll put our drawing commands

“Steve, what do you see?”. “Things! Wonderful things!”. Or maybe ancient curses and diseases, all made of blocks. [Image credit raspberrypi-spy.co.uk]

inside a function. You can either enter these commands into the Python interpreter directly, or save them as a file, say ngon.py, in the same directory as mcpiturtle.py. import random def ngon(t, n, length): angle = 360./n for j in range(n): colour = random.randint(0, 15) t.penblock(35, colour) t.forward(length) t.right(angle) Now import the ngon module, if you saved it externally, move your turtle out of the way of your pyramid and from the same interpreter run: >>> import ngon >>> ngon.ngon(t, 9, 10) This will draw a regular 9-gon (a nonagon) of side-length 10 blocks. Through the random module we change the turtle's draw colour. Wool has blockType 35 (we don't have access to the block class so we can't use the more descriptive names seen earlier) and the second line of code sets the blockData parameter, which for wool blocks controls the colour.

Lignum fractus We've covered fractals before, most recently using Python to generate von Koch snowflakes in Gimp [see Tutorials, p84, LXF190]. There we learned all about recursion, and how a function which calls itself can lead to pretty pictures. For this tutorial our focus shall be more arboreal than meteorological, in that we shall produce some glorious trees, making this a carbon-neutral tutorial. Making 2D fractal trees with turtle graphics is a pretty straightforward process – start with a straight line (the trunk), turn right through a set angle and repeat the procedure for a smaller line, then turn left a set angle and repeat. In Python the general shape of this code (excluding initial setup) would be: def tree2d(branchlen, t): if branchlen > 2: t.forward(branchlen) t.right(20) tree2d(branchlen - 2, t) t.left(40) tree2d(branchlen -2, t) # then return to original heading/position t.right(20) t.backward(branchlen) Then calling branchlen(20) would slowly but surely grow a tree-like object. The if statement stops the function being an infinite recursion – each time we draw smaller and smaller branches, but the function will stop doing anything when called with a branch length of 2 or less. Hence our function call produces a tree with a trunk of length 20, and up above twigs of length 4. Generalising this to three dimensions is quite straightforward – instead of having two recursive calls to our function, we shall instead have four. One for each compass direction, and each one deviating from its parent's vertical heading by 20 degrees. We can also add some flair by drawing the shorter (higher up) branches in random colours,

Never miss another issue Subscribe to the #1 source for Linux on page 34. 72 LXF195 March 2015

www.linuxformat.com

Minecraft/Pi Tutorial

Next issu Build youe: own rout r er

speaking to the spring blossom that will be appearing Any Day Now. Larger branches (and the trunk) will be made of wood, just like in real life. Finally, since this can be quite a slow algorithm (calling tree(20) results in 4,096 recursive calls) we reset the turtles original position directly using t.setposition() per function call, rather than using t.backward() and doing all the inter-block calculations and animations unnecessarily. And without further ado, here is our 3D tree function: def tree(branchLen,t): if branchLen > 6: if branchLen > 10: t.penblock(block.WOOD) else: t.penblock(block.WOOL.id, random.randint(0,15)) x,y,z = t.position.x, t.position.y, t.position.z t.forward(branchLen) t.up(20) tree(branchLen-2, t) t.right(90)

tree(branchLen-2, t) t.left(180) tree(branchLen-2, t) t.down(40) t.right(90) tree(branchLen-2, t) t.up(20) t.setposition(x, y, z) You'll find all this together with all the required setup in the file mcpi3dfractaltree.py on this month’s LXFDVD. Importing this file (while a Minecraft world is loaded) will result in animation being disabled (by calling speed(0)) and a rather funky looking tree automatically being generated in your current position. You can easily alter the tree if it's not to your tastes. If you wanted to expand the 3D trees project a little, why not add small random numbers to the angles – after all, nature isn't all right angles and perfect symmetry. Also, trees (in the northern hemisphere) tend to lean southwards, so your code could favour one direction in order to recreate this phototropism. And that concludes another exciting instalment of our Minecraft: Pi Edition series. LXF

Helices ain’t no thang, if you’re a turtle. Next week: Build your own Minecraft DNA sequencer [Ed – er, no].

Quick tip There are all manner of other Minecraft Turtle examples on Martin's GitHub: http://bit.ly/ MineCraftTurtle.

Installing Minecraft:Pi Edition and the turtle module We've covered this before, but since packages are now available for Raspbian it's a bit easier now. First, if you downloaded Raspbian after September 2014, then you'll already have Minecraft installed system-wide and don't need to do anything. Otherwise a simple $ sudo apt-get update $ sudo apt-get install minecraft-pi will get the job done. Prior to this packaging, it was necessary to install the program to your

home folder and copy the Python API files from there to your project's folder. This approach will still work, but it's messier than the new way. You'll find the mcpiturtle.py file on the LXFDVD, this will work without modification if you've installed the package as above. It will also work the old way if you happen to have the API files in a directory called mcpi and the mcpiturtle.py file one level above this directory. Otherwise you'll need to modify the

www.tuxradar.com

first two import lines, according to where the API files are located. You can test this out by starting Minecraft ($ minecraft) and entering a world. In a separate terminal window, cd to wherever you've copied mcpiturtle.py, the run it with $ python ./mcpiturtle.py This will run the tests in the if __name__ == "__main__": stanza, which include: drawing a pentagon, changing pen colours, and general meandering.


Tutorial Xxxx ImageMagick Use the convert and

mogrify commands to alter images

ImageMagick: Convert images Neil Bothwick says forget Gimp, put your mouse away and learn how to manipulate images from the command line.

Our expert Neil Bothwick has a great deal of experience with booting up, as he has a computer in every room, but not as much with rebooting since he made the switch from Windows to Linux.

If you find the man pages a little terse, the ImageMagick website has more details and command examples.

A

sk a group of Linux users to name a graphics processing package and the vast majority will mention Gimp (GNU Image Manipulation Program) first. You may get a few votes for other programs but you are unlikely to get many for one of the most used graphic packages on Linux, ImageMagick. This is a suite of command line image processing programs, and what is the point of such a thing, you may be asking? Surely editing graphical images is best handled by a graphical application? If you want to retouch or otherwise modify a single image, you would be right, but have you ever tried converting a directory of images to a different format,or resizing them? That's where the ImageMagick programs come into their own, and the reason why this is one of the most used image processors on Linux, even if it is behind the scenes. The most used command in the ImageMagick toolbox is convert, and format conversion is the simplest of its operations. Using convert pic1.jpg pic1.png does exactly what you would expect it to, converts an image from JPEG to PNG. Well, it makes a new image file, in place modification is something we will look at shortly. We didn't need to specify the type of input and output, convert works


www.linuxformat.com

the input format out for itself (that's part of the magic) and uses the file extension to choose the output format. You can change this by prefixing the output file with an image type and a colon, like this: convert pic1.jpg bmp:pic1.gif This would create a BMP image with a GIF extension, which may seem rather useless until you consider that, in common with many other programs, convert accepts a filename of - to represent standard input or output. So if you want to send image data to another program in PPM format, you could do: convert pic1.jpg ppm:- | another_program Some output formats also accept extra options, such as JPEG quality: convert pic2.png -quality 75 pic2.jpg The convert man page and online documentation give the full range of such options.

Resizing images After format conversion, resizing is one of the more popular image processing tasks, and convert does this too. convert pic3.png -resize 50% pic3_small.png This is the simplest resize operation, a single scaling factor is applied to the image. You can scale by different factors on the X and Y axes with -resize 40x60%, even though there is only one % sign, it applies to both numbers. If a single numeric value is given then that is taken to be the width in pixels of the new image; the height is calculated to preserve the aspect ration of the image.

ImageMagick Tutorial convert pic3.png -resize 1000 pic1000.png convert pic3.png -resize x1000 pic1000.png in the second example, the value is for the height – think of it as widthxheight with the width value missing. If both width and height are given, the resulting image size is the largest that will fit into that area while preserving the aspect ratio. If you want to set absolute dimensions and ignore the aspect ratio, add a ! to the size: convert pic4.png -resize 1000x600 pic5.png convert pic4.png -resize 1000x600! pic5.png The first line preserves aspect ratio, the second creates an image exactly 1,000x600 pixels. If you are creating thumbnails of larger images, you may want to avoid enlarging any smaller images in the list, which usually just creates a bigger but fuzzier image, so add > like so: convert pic4.png -resize 1000x600\> pic5.png The > tells convert to only resize the image if the original is larger than the given size. Note: that you will need to escape or quote the > when running this in a shell or it will be interpreted as a redirection operator. Operations can be combined in a single call, and are applied in order, so you can resize and convert to a different format at the same time: convert pic6.png -resize 30% -quality 75 pic6a.jpg

Add some text An interesting option is the ability to add a text caption to an image, for example: convert pic7.png -gravity SouthEast -family Times -pointsize 30 -annotate +10+10 "Some text" caption7.png The key component here is annotate, which is followed by the x,y position of the text and then the text itself. The -gravity setting is also important, it describes where coordinates are measured from. In this example the text is drawn 10 pixels in from the bottom right-hand corner of the image. The -family and -pointsize options control the font and size of the text. It’s often preferable to use -family rather than -font because -family will fall back to the closest match if the exact font is not available, instead of failing with an error. ImageMagick options apply from the point at which they appear in the command line, so you must specify the gravity and font settings before -annotate or they will not apply (unless you have a subsequent text operation, in which case they will apply there). This could be used in conjunction with the date command to caption images with their timestamp in the following way: convert pic7.png -gravity SouthEast -annotate +10+10 "$(date -r QR.png +%d-%b-%y\ %H:%M)" new7.png

Convert creates a new file but sometimes you may need to modify a file in place, ImageMagick provides the mogrify command for this. This works like convert, except mogrify overwrites the input file with the changes, compare these two commands: convert pic8.png -resize 50% pic9_small.png mogrify -resize 50% pic9.png As you are destroying the original, you need to be sure of the operation you are performing before running it, so it's wise to use convert to test it out first. However, mogrify has another trick up its sleeve. With convert, you need to specify both the input and output file, so each invocation can only be run on a single file. On the other hand, mogrify only ever needs one file name, so if you feed it several, it runs the command on each file in turn, so mogrify -resize 50% *.png will process all matching files in the current directory. We said that mogrify overwrites the original file, but there is an important exception to this; if the command is used to change the file format, a new file is created with the relevant extension and the original is left untouched, even if other operations are also performed, using mogrify -resize 25% -quality 70 -format jpg *.png will create JPEG thumbnails of all PNG files found in the current directory. We have only scratched the surface of what is possible with the ImageMagick package, the commands we have covered have many more options and there are other commands. For instance, Identify does what it says, tells you about the format and size of a file, while the import command can be used to grab images from the X server, and screenshots of individual windows or the entire screen. The man pages are rather dry to read, but there are plenty of explanations and examples on the official ImageMagick website at www.imagemagick.org. LXF

Use identify to get more information about image files of any type.

Batch processing While mogrify can work with multiple files, convert works on a single file at a time, so how do you process a directory full of files? The easy way is with the shell's for operator. for i in *.png do convert "$i" -resize 500x500 "thumbnails/${i%. png}.jpg" done This runs the commands between do and done once for each file that matches the pattern.

Each time, $i is replaced with the name of a file. The ${i%.png} part is replaced with the name of the file with the .png removed from the end (% removes the following string from the name) and we then add the extension .jpg, so what is actually run is: convert "pic1.png "-resize 500x500 "thumbnails/ pic1.jpg" convert "pic2.png" -resize 500x500 "thumbnails/ pic2.jpg" ...

We enclose the file names in quotes in case there are any that contain spaces, which would otherwise confuse the shell. This works well if all the files are in a single directory, if they are in subdirectories it is sometimes easier to use find: find photos -name '*.jpg' -exec convert "{}"-resize 25% "thumbnails/{}" \; Each file that is found is passed to the command following -exec, where {} is replaced by the file name. Once again, use quotes to avoid problems with spaces.



Tutorial Xxxx Detect movement, record a Motion

livestream and back it up on a server

Motion: Detect and record Kent Elchuk demonstrates how to build a livestreaming system, using a Raspberry Pi and a webcam and how to save motion-detected video.

Our expert Kent Elchuk is an avid web developer, Linux enthusiast and the creator of Cloner and Sitemakin CMS. He likes to experiment with new ways to make the web a greater experience for everyone.

W

e'll assume you have none of the required packages to follow this tutorial on video surveillance and video recording. You will use Motion which is the heart of this article. Aside from that, you will require Apache (or Nginx) and PHP. Although this tutorial is geared towards using a Raspberry Pi, you can use another PC setup if you prefer. Do note, that if you go the Apache and PHP route, everything will work very easily without having to make extra changes to the server and PHP. If you do decide to go with Nginx instead of Apache you will need to make some extra changes: such as installing PHP-FPM; changing the root folder path for web page files; and editing the following files: /etc/nginx/sites-available/

Use simple password protected authentication to keep files secret.


www.linuxformat.com

default, /etc/nginx/sites-enabled/default and /etc/ php5/fpm/php.ini. Now, for the synopsis of each package. Motion will be used to record video after movement is triggered. The video clips will be written to a folder as Flash SWF files. However, Motion still allows you to see the location even without movement, much like a regular security camera. Once you have those files, you may want to be able to sort through them effectively. Here is where the web server and PHP play their role. With the Apache or Nginx server, you can serve these files over the web. Realistically, many files will be accumulated and you may want to create a loop with PHP in order to output each file into a link that can display the video in a popup. In which case a free video popup application, such as Shadowbox can be used. Lucky for you, the code included on the LXFDVD contains the files that do all that stuff. With all that covered, you'll have a setup that can deliver your videos. This tutorial will show you various options and their how-to counterparts. Since a camera like this could be used in your home as a security camera, you may want to password protect any web pages or the folder where you keep the videos. Now, if someone did happen to break into your premises and decide to steal or wreck your Raspberry Pi, we'll also guide you through a backup plan that can be used to move your video files to a foreign web server that the robber won't have a clue exists.

Getting things to work Since this article is about Motion, let's install this first: sudo apt-get update sudo apt-get install motion Now that one installation is out of the way, let's add the rest, which includes Apache sudo apt-get install apache2 and PHP: sudo apt-get install php5 libapache2-mod-php5 php5-mcrypt Let's move on and make some basic procedures and tests to see everything is working as it should. The main files which you will customise are /etc/motion/motion.conf and /etc/ default/motion. Open up motion.conf with your favourite editor. By default, you'll note that the parameters shown below are the opposite of the default values. For example, daemon off becomes daemon on: daemon on webcam_localhost off control_localhost off

Motion Tutorial Save the changes and open up the /etc/default/motion file and make the following changes: start_motion_daemon=yes Now, let's fine tune some options. Three changes that are needed are: the frame rate, quality and minimum amount of frames to trigger the motion to record: framerate 30 quality 90 minimum_motion_frames 5 Without changing this setting, two frames per second looks way too jerky and will miss out a lot of action, so we change the frame rate from 2 to 30 frames per second. The second change is obvious since it's a quality upgrade. The third change sets the minimum amount of frames of motion that need to be detected. By default, the value is 1. The problem with a number this low is that you can end up with unwanted recordings from things such as lights flicking. Keep in mind that you have many options and can look deeper into the features. A good place to start is on the offical website (http://bit.ly/MotionConfigFileOptions). Some of the other features you might want to consider are: taking a picture at a desired interval, such as every second, every minute or every hour. This feature makes it easy to host a live weather cam, for instance, or to determine if someone is sitting on your couch.

Configuring Motion Changing all parameters to suit your specific needs is very easy and the motion.conf file will often have nice, selfexplanatory comments while the website and man page have more information to offer. Obviously, this service doesn't do much without a working, compatible webcam and a list of webcams worth trying with the Raspberry Pi can be found at http://elinux.org/RPi_USB_Webcams. Using a plug and play webcam makes life easy for this task, and one cheap, readily available webcam that works is the Logitech C170. Note: If you are using the Raspberry Pi, the Raspberry Pi cam won't work with Motion. To tell if the USB webcam connects OK, run the command lsusb. At this point, you will likely have a working webcam, a working web server and an adequate Motion configuration. This is good, but you'll also need to create a folder for the images and set ownership for Motion. By default, Motion drops the images and SWF files into the /tmp/motion folder. It won't create the folder, therefore, you will need to: cd /tmp mkdir motion

Alert! Man on sofa. Capturing live video feed.

chown motion:motion motion Now, let's see how everything works. To start with, you can get Motion up and running with the command service motion start and you can always restart it with the command service motion restart The first thing you'll need to do to test that everything works fine is to see if you can view the default web page. Since your Pi will have a unique network address, you can just type it in the browser. For example, if the Pi is connected to a router with the IP of 192.168.0.1, your device could have an IP like 192.168.0.106. Thus, the URL would also be http://192.168.0.106. If you have success returning the default web page, you will see a message stating that everything is working properly. If not, you will get a typical browser error which makes it obvious that something is not quite right. Now with a working server, let's move on and get down to viewing and recording video. You can test the video in your browser by typing your network IP and port. By default, the Motion webcam port will be 8081. Therefore, if you type http://192.168.0.106:8081 in the browser, you should see your video stream. A simple setup like this can be beneficial and have many uses aside from security: such as keeping an eye on a newborn while you’re working in another room. Since all should be well at this point with the Motion service running, you can now go in front of your webcam and jump around a bit. Actually, hand waving will suffice but a few jumping jacks aren't going to hurt. After that, you should be able to browse the new motion folder and see a bunch of JPEG files and at least one SWF file.

Quick tip When you're logged in via SSH and need to edit files, using vim to find a string is easy. To do this, all you need is a / followed by the string name. Just type n to move on to the next one.

Using multiple webcams Need more than one webcam? No problem. Motion enables you to add more easily. You need to open up /etc/motion/motion.conf and set up the threads. If you go to the bottom of the file, you see various lines that are commented out followed by the word thread. As you can see, the default location for these new files are in /usr/ local/etc folder. To keep it simple, you can change the thread folder to the /etc/motion folder. This way, you keep all the editing in one simple location. Now,

the first thread would resemble the line below: thread /etc/motion/thread1.conf Once you’ve set up your threads since you are using multiple webcams, you can create the thread files. Thus, your first thread would be called thread1.conf, followed by thread2.conf, and so on. The code that you add to these threads just needs to be a few lines. The code samples below display two threads. As you can see, each thread has its own videodevice parameter, custom text that appears on the left-

hand side of the video stream and image folder and port number. Here' s thread1.conf videodevice /dev/video0 text_left Camera #1 target_dir /var/www/images webcam_port 8081 followed by thread2.conf: videodevice /dev/video1 text_left camera #2 target_dir /var/www/images_cam2 webcam_port 8082



Tutorial Motion

Displaying stored images and Flash video.

Now, that you have a motion-detecting device that you can view from your local network, you can move on and make adjustments so that you can see and view your webcam from outside your local network. For some of you, this may be the setup you desire; especially if your webcam and Raspberry Pi are well hidden and unlikely to be tampered with. However, in addition to storing data in your own home, we will explain how to back up the files to another server for safe keeping, just in case your SD card or hard drive fails, or someone decides to steal, break or ruin your webcam (or webcams) in your home.

With that said, we will move on and create a simple way to organise the saved files that recorded the movement. The first detail that will need to be changed is to save the images and SWF files into a folder within the web directory. The root web folder is located at /var/www/html or /var/www. At this point, a light should go off since you have already made several changes to the Motion setup. Reopen the /etc/motion/motion.conf file and change the target directory. By default, the target directory is located at /tmp/ motion. The new target is /var/www/images: target_dir /var/www/images

Making it web friendly

Viewing recorded video and images

The whole idea here is to record, store and manage video from a website or by using your IP address that was given to you from your ISP. To find out your IP head to http:// whatismyipaddress.com. In order to broadcast video out, you'll need to set up port forwarding on your router to enable your network IP to use port 8081. While you are at it, you may as well do the same for port 80 since this same network IP will be used to display web pages from computers outside your local network; such as your friend across town or a loved one overseas. After you have made the previous changes to your router's settings, try typing http://my_ipaddress_from_isp and http://my_ipaddress_from_isp:8081. You should get the same results as you did when you were checking your local IP. The next step is to clean this up and view organised data view web URLs; such as http://my_ipaddress_from_isp/ video.php or to view it from another website using an iframe. In order to show the webcam from the page video.php, you just need to use an img tag with the network IP and port. Have a look at the code below, which shows this in complete detail and displays it with the default width and height as specified in the motion.conf file: Now, let's imagine a scenario where you want to stream this livecam from another website using an iframe. Well, all you have to do is make an iframe from another page on the different server. The simple one-liner is shown below. <iframe style="width:320px; height:240px;" src="http://isp_ ipaddress/video.php"> The next set of code will explain how to display the files that have been saved after motion is detected and recorded.

After making changes to motion.conf, type the command: sudo service motion reload so that it will now attempt to write any new files to the /var/www/images folder. Now, you can easily access the files created by the Motion service and display them on the web just like any other typical web page. Although the path has been changed in motion.conf, the images folder hasn't been created yet. So, make it now. The folder will be located within the www or html folder. If it sounds like we’re repeating ourselves here, because it means you have been paying attention and are aware that the Apache root web folder can be in one of two paths: cd /var/www mkdir images By default, the www directory will be owned by the root user and root group. You will want to make some changes; such as all files will be owned by pi and the group will be www-data. To change this use: cd /var chown -R pi:www-data www So, what we are up against now is to make this images folder writable by the Motion service. As of right now, the other files have adequate ownership and permissions, but, the images folder does not. Well, let's change that right now. The code snippet below has three commands. The first command will add the user motion to the www-data group. In case you are wondering, www-data is an existing user and group for the Apache server. The second command gives the images folder permissions to the Motion user and www-data group, and the final command makes the folder writable so that the images and SWF files can magically appear in the images folder: usermod -a -G www-data motion chown motion:www-data images chmod 777 images It's within the www folder where you can create a file called shadowbox.php that will be used to display content on the web. This file has all the code you need to display a thumbnail from each recorded SWF video, the video itself and the first JPEG image of motion. The coding process to display the content goes like this: The images directory gets scanned and an array of files is created and sorted in descending order. These files are multiple JPEG files and a single SWF file for each event. All files have a name that starts with the event, followed by the date and finally followed by the sequence. The date sequence is year, month, day, hour, minutes and seconds. After that, Motion just adds the sequence starting from 01 only for the JPEG files.


www.linuxformat.com

Motion Tutorial In order to keeps things simple and uncluttered, the coding for the shadowbox.php file will only display a single image and a single SWF file for each event. This simple script also uses shadowbox to create popups of the first JPEG image and flash SWF file for each event. Now, you can see all the latest detected motion, down to the first recorded by Motion. This file gives all of the results and here is where you may want to customise the output. If you want to keep these web pages password protected with Apache, you can open up the file /etc/apache2/sitesavailable/default and make some minor changes. If you look for the line , you can add three simple lines to it. The code sample is shown below: AuthType Basic AuthName "Private Documentation Repository" AuthUserFile /var/www/.htpasswd Require valid-user After that, you navigate to the /var/www folder and create a blank file called .htpasswd, and create the username and password with the simple command displayed below. You will be prompted for a password twice. You simply add it followed by Enter: sudo htpasswd /var/www/.htpasswd add_username_here Since the files can pile up pretty quickly and your disk space can readily disappear, you may want to create a purging system or back them up to another drive. Some backup plans are discussed next.

Backup plans One tip for determining your backup plan is to watch how much space you routinely tend to use and develop a plan based on those conditions. Simple. For example, if you go through 1GB a week and you have a 8GB card you may want to TAR the images folders, SCP the file to a remote server and remove all files that are more than one-week old. Since the files contain the year, month and date, it's a rather easy process to delete the ones that have expired. The file called purge.php is a cleanup file that is included on the LXFDVD this month. This file removes every file that's more than a couple of days old. I will explain the code in a little more detail in a moment. First off, the images folder is scanned and all of the files become an array. That array of files then iterates through a foreach loop. A few built-in PHP functions, such as strstr(), preg_replace(), substr_replace(), substr(), date() and unlink() are used to translate all the file names into actual date timestamps that can be used for comparison.

Once a timestamp is made from the filename, it goes through a simple if() statement and is compared against a time that is set to two days ago from the current time. This part is really easy to change since you just need to change the number 2 to your desired amount of days in the past. Once this criteria is met, the file is deleted with the unlink() function. Since this system is only using files without a database, it's rather elementary to move all of these files to your backup location, and since this is copying and moving files, two methods come to mind. One is using a package such as rsync and the other is a simple method of compressing the desired files and folders with ZIP or TAR and shipping them to their new destination with SCP. An simple example of SCP is shown below: scp -P 22 /var/www/images.tar [email protected]:/home/pi/ images.tar So there we have it. You've just created your own video surveillance and motion recording system that has several options to suit your needs or that you can customise. Although we've made a rough skeleton and files for you to monitor your video, record files and make backups, you can take this farther if you want. Some simple suggestions would be are to add a responsive template to both the video.php and shadowbox.php files, and polish up the content with a little CSS magic. On top of that, you could set up webcams at other sources and have them viewable by the public or friends, depending upon what you want to achieve. Have fun! LXF

All the files in the images folder are named with an event, followed by date and sequence.

Quick tip You may want to include the time in the file name so backups will not be overwritten. In addition to that, you may want to run a cron job that does this procedure on a regular basis.

Nginx and Motion Nginx doesn't ship ready to go out-of-the-box for Motion as Apache does. In fact, after you install Nginx there's a series of required steps that you must do before you have a smooth operation. In the case of Raspberry Pi, you can adjust the worker_processes value from 4 to 1. You can change this in the /etc/nginx/nginx.conf file. This is recommended since the Pi only has a single CPU core. After that, you will want to change the default web folder since the default points to /usr/ share/nginx/www. To change this, you open the file called /etc/nginx/sites-enabled/sitesenabled/default. The change is shown below so the web folder is /var/www:

#root /usr/share/nginx/www; root /var/www; After the previous step, you can quickly install fastcgi. The command is below: apt-get install php5-fpm After that, you’ll need to open the file /etc/ nginx/sites-available/default and change a few lines of code so it resembles the content below. Basically, you’ll just need to remove a few comments: location ~ \.php$ { fastcgi_split_path_info ^(.+\.php)(/.+)$; # NOTE: You should have "cgi.fix_pathinfo = 0;" in php.ini # With php5-cgi alone:

www.tuxradar.com

# fastcgi_pass 127.0.0.1:9000; # With php5-fpm: fastcgi_pass unix:/var/run/php5-fpm.sock; fastcgi_index index.php; include fastcgi_params; } We’re almost there. Now you’ll have to open the file /etc/php5/fpm/php.ini and remove another comment so that it looks like the line of code below: cgi.fix_pathinfo=1 Finally, make sure to restart Nginx after making all of the changes. The command /etc/init.d/nginx restart will do the trick.


Tutorial Xxxx blogging platform Build a Ghost

custom theme using Handlebars.js

Ghost: Create custom themes Steven Wu explains how to get started and build your own theme for the popular open source blogging platform Ghost.

Our expert Steven Wu is a freelance Magento & Wordpress web developer. Follow him on Twitter at @designtodevelop to learn about his current projects.

W Quick tip All the files you need for this tutorial can be found on Steven's GitHub at http://bit.ly/ BuildAGhostTheme.

e covered the fundamentals of Ghost previously [See Tutorials, p74, LXF183], but in case you missed it: Ghost is a free open source blogging platform. Successfully funded through Kickstarter in May 2013, it surpassed its original request of only £25,000 achieving over £196,000 in funding. Started by John O’Nolan, Ghost has a unique purpose in providing bloggers with a simple interface that enables them to write and publish their content without the hassle or distraction by the sheer complexity of development from traditional platforms. Ghost has been beautifully designed from the ground up. Its clean and simplified UI enables you to quickly browse through the archive so you spend less time managing your blog and more time blogging. It has a smart writing screen using Markdown with a real-time preview on the right-hand screen and simple drag and drop functionality to add images into place. Ghost has three main core principles: First, it’s developed for users rather than developers, unlike many blogging and CMS platforms out there. Second, the platform has a MIT licence so you can do what you like with this platform with


www.linuxformat.com

few limitations. Third, it’s made for love. Ghost is a non-profit organisation, which means its motivations are to support bloggers rather than satisfying investors. In this tutorial we will show you how to install and set up Ghost locally and build your first Ghost theme. To begin building our Ghost theme, start within the Ghost installation folder [see top, p81 for the installation guide]. Under content/themes, create a new theme directory called mytheme, or something more imaginative – make sure it's in lowercase without any spaces (hyphens are acceptable). This will be the directory that houses our theme codebase. Within this directory, create the following files and folders: assets/ css/ normalize.css screen.css images/ js/ fonts/ partials/ header.hbs default.hbs index.hbs post.hbs Both index.hbs and post.hbs are the only files required for a valid theme. Without any of these you’ll receive an error. Now in the Ghost dashboard, navigate to Settings > General. Under Theme, select the new theme you just created called mytheme. If it’s missing, you’ll need to go to the terminal and restart Ghost. Click Save to activate this theme. You won’t see anything in the frontend yet. This is because we have yet to add any markup in our theme.

Using Handlebars Ghost makes use of a templating language called Handlebars. js (http://handlebarsjs.com), and its predefined expressions make it simple to build and maintain Ghost themes. Handlebars separates the templates from the raw HTML for you. Bear in mind that, with Handlebars, you can’t write functions or hold variables. Handlebars is developed simply to display content where the expressions are outputted. Handlebars expressions are wrapped with curly brackets and look like this: {{author.name}}. This basically looks up the author.name property and outputs it. Let’s get our hands dirty and start creating our theme. Open up the default.hbs file in your favourite text editor. This is the base template and includes all the basic ,

Ghost Tutorial Installing Ghost Ghost is a lightweight web application. It only takes a matter of minutes to install locally and get up and running. Ghost is a JavaScript application built upon Node.js. You can install the latter from your distribution's repositories – they will almost certainly contain a version from the required latest stable 0.10.x series. Debian/ Ubuntu/Mint users can get this with: $ sudo apt-get install nodejs If you want/need a later version then you can head over to http://nodejs.org and download it, or if you prefer a tidier install check out the distro-specific guidelines at http://bit.ly/

InstallingNodejs. Ghost itself hasn't been officially packaged on most distros, so go to https://ghost.org/download to get the latest version and uncompress it. To install Ghost in the terminal, run: $ cd /path/to/downloads/ghost $ npm install --production A javascript package manager that comes with Node.js called npm will now install all the necessary dependencies to the production and development environment. Once this is completed you can now start Ghost in development mode with:

, tags that will be used throughout your Ghost website. In this template we input our HTML doctype, basic meta tags and head and body tags. (See default.hbs in tutorial files on the GitHub here: http://bit.ly/BuildAGhostTheme.) You’ll notice expression tags: {{! Responsive Meta Tags }}. Any expressions preceded by an exclamation point within the curly brackets are comments and will not be printed in the final source code. The {{ghost_head}} helper is used to output any system scripts, styles and meta tags. The {{ghost_foot}} helper is used to output scripts at the bottom of the document. The Handlebars expression {{{body}}} is an important one. This is where all your content will be displayed, which extends the default template. The following {{body_class}} is used to automatically generate CSS class names for targeting specific pages:
{{{body}}}
{{ghost_foot}} Index.hbs Now in our index.hbs (see this source code in the project files on the GitHub), we use the Handlebars expression {{!<

You can use the {{date}} helper to output the published date and use the format option to control the date format.

$ npm start In your web browser, navigate over to your newly installed Ghost blog at http://127.0.0.1:2368, and you can register your administration login by going to http://127.0.0.1:2368/ghost. When you visit this URL, you’ll notice Ghost notifies you of the send email and Ghost URL being used. Configure this by changing the base URL. Open up config. js in your text editor and change the url parameters in the Development or Production stanzas. Once you’ve completed development, shut down by pressing Ctrl+C in the terminal.

default}} at the very head of this document to reference to our previous base template. This template will be used for our homepage. We will want to style each blog post within the foreach helper. By using the opening {{#foreach posts}} and closing {{/foreach}} loop, anything inside this will display each post with the markup. To display the content of each blog post we use a Handlebars expressions {{content}}. We can also limit the word count by using the parameter words="100". You’ll notice all of the class names are proceeded with mytheme_. This is a recommended practice when building a Ghost theme. Ghost will automatically assign particular class names and more specifically IDs to certain elements in your theme. You’ll want to prevent clashes and consider your scope of class names.

Partials Typically we can insert our header markup just below the {{!< default}}, but the advantage of Handlebars templates are hierarchical support, whereby one template can extend another. This includes the use of partials. This helps to eliminate repetition of code and encourage reusability. We can separate our header into a partial template. Within the partials directory open up the header.hbs. In Ghost dashboard Settings, you can upload your own blog logo and blog cover image. We’ll use an if statement to check whether a blog cover image exists. If so, we’ll output it as a background image: {{#if @blog.cover}} style="background-image: url({{@blog.cover}})" {{/if}} This time, we’ll check if a blog logo is available. {{#if @blog.logo}} {{/if}} The @blog global data accessor has access to global settings in Ghost we can output in our theme. Now let’s dive into the fun part of styling our theme. In our theme, we’ve linked to normalize.css for our HTML5-ready CSS reset. Within the screen.css is where we’ll input all our custom theme styles. We’ll then add some global styles, followed by styling our header and set a max-width to prevent our layout from expanding over certain pixel size:

If you missed last issue Head over to http://bit.ly/MFMissues now! www.tuxradar.com


Tutorial Ghost

: Next issuee Keep timP with NT

.mytheme_page { max-width: 980px; margin: 0 auto; } .mytheme_header { padding:20px 0; text-shadow: 2px 2px 2px rgba(26, 26, 26, 0.1); text-align: center; color: #2f3727; } Now style each blog post within its article container: main article { margin: 30px 0; border-left:1px solid #DBDBDB; border-right:1px solid #DBDBDB; background-color: #FFFFFF; } .mytheme_post_content { padding: 0 20px; } .mytheme_post_title { margin: 0 0 20px 0; padding: 10px; font-size: 2em; letter-spacing: 2px; text-align: center; text-shadow: 2px 2px 2px rgba(26, 26, 26, 0.2); color: #FFFFFF; background-color: #8ACD36; } .mytheme_post_title a { text-decoration: none; color: #FFFFFF; } .mytheme_main_img img { width: 100%; max-width: 100%; border: 0;

Select the newly created theme in the dashboard to activate it. You may need to restart Ghost in the terminal to pick up the theme.

After installing Ghost, you’ll see your development URL and a warning about setting up an email service.

} Place the date on the left-hand side and the Read More button opposite. Give this link the presentation of a button: .mytheme_post_info { overflow: auto; padding: 0 20px; background-color: #98C148; } .mytheme_date { float: left; padding-top: 20px; color: #FFFFFF; } .button { float: right; padding: 20px 0; } .button a { padding: 5px; transition: ease .3s; text-decoration: none; color: #FFFFFF; background-color: #39A9DA; } .button a:hover { background-color: #199ED9; } We touched upon the key aspects of developing a Ghost theme. Hopefully this insight has opened up possibilities to creating your own personalised theme. You’ll see that building a Ghost theme is very simple to picking up with the predefined Handlebars expressions. Don’t forget to download the tutorial’s accompanying files from the GitHub: http://bit.ly/ BuildAGhostTheme. Also check out the documentation for more information to help you build and customise your own theme by visiting: http://docs.ghost.org/themes. LXF

Preparing post.hbs Having only set up the homepage template, the post.hbs controls the display of each single blog post page. Again on this template (see the post. hbs file in the project folder, which you can download from GitHub here: http://bit.ly/ BuildAGhostTheme) we use the same {{!< default}} and {{> header}} Handlebars expressions. This time we use the opening {{#post}} and {{/post}} expressions to display a


single blog post. While in the single post context we have access to the author data, including the author name and bio. We can display the author details by simply adding the code displayed below:

Written By:

{{author.name}}

{{author.bio}}

www.linuxformat.com

We can now apply some CSS styling to our single blog post. It’s recommended that you actually place all your single-post styles in a separate CSS file. It’s important to do this is because, in the near future, Ghost will be unveiling a new split screen feature, which will load all the custom theme styles in the admin interface.

IT INSIGHTS FOR BUSINESS

THE ULTIMATE DESTINATION FOR BUSINESS TECHNOLOGY ADVICE Up-to-the-minute tech business news In-depth hardware and software reviews Analysis of the key issues affecting your business

www.techradarpro.com twitter.com/techradarpro

facebook.com/techradar

Python 3

Python: Dive into version 3

Jonni Bidwell investigates probably one of the least loved and disregarded sequels in the whole history of programming languages.

Our expert Jonni Bidwell is sad that all his print statements are now at least two characters longer, but happy that Python 3 is better than the third (and second) Matrix movie.

W

ay back in December 2008 Python 3.0 (alternately Py3k or Python 3000) was released, yet here we are, seven years later, and most people are still not using it. For the most part, this isn't because Python programmers and distribution maintainers are a bunch of laggards, and the situation is very different to, for example, people's failure/refusal to upgrade (destroy?) Windows XP machines. For one thing Python 2.7, while certainly the end of the 2.x line, is still regularly maintained, and probably will continue to be until 2020. Furthermore since many of the major Python projects (also many, many minor ones) haven't been given the 3 treatment, anyone relying on them is forced to stick with 2.7. Early on, a couple of big projects – NumPy and Django – did make the shift, and the hope was that other projects would follow suit, leading to an avalanche effect. Unfortunately, this didn't happen and most Python code you find out there will fail under Python 3. With a few exceptions, Python 2.7 is forwards-compatible with 3.x, so in many cases it's possible to come up with code that will work in both, but still programmers stick to the old ways. Indeed, even in this


www.linuxformat.com

esteemed publication, this author, whether by habit, ignorance or affection for the past, continues to provide code that is entirely incompatible with Python 3. We won't do that in this article. We promise. So let's start with what might have been your first ever Python program: print 'Hello world' Guess what? It doesn't work in Python 3 (didn't you just promise...). The reason it doesn't work is that print in Python 2 was a statement, in Python 3 print is a function, and functions are, without exception, called with brackets. Remember that functions don't need to return anything (those that don't are called void functions), so print is now a void function which, in its simplest form, takes a string as input, displays that string as text to stdout, and returns nothing. In a sense you can pretend print is a function in Python 2, since you can call it with brackets, but a decision was made to offer its own special syntax and a bracketless shorthand. As an aside, this is rather like the honour one receives in mathematics when something named after its creator is no longer capitalised, eg abelian groups. But these kind of exceptions are not a part of the Python canon ("Special cases aren't special enough to break the rules"), so its brackets all the way. On a deeper level, having a function-proper print does allow more flexibility for programmers – as a built-in function it can be replaced, which might be useful if you're into defying convention or making some kind of Unicode-detecting/defying wrapper function. In sum, your first Python program should have been:

The Greek kryptos graphia, which translates to ‘hidden writing’ [see Cryptography Old And New, p50, LXF189] followed by a new line using the correct script.

Python 3

The Unicode revolution Traditionally, text was encoded in ASCII, in which each character is encoded as a 7-bit codepoint, which gives you 128 characters to play with. Some of this number are invisible teletype codes (ASCII originated in the 1960s) and once we've counted the familiar alphanumeric characters there isn't really much room left. Since we like things to be bytes, several 256-character extensions of the ASCII encoding emerged. The most notorious of these is ISO-8859-1, sometimes called Latin-1. This widely-used character set (and the related Windows-1252) contains almost all the accents required for the Latin-scripted languages, as well as the

characters used in the romanisation of other languages. As a result, it’s fairly common in the western hemisphere, but doesn't really solve the problem elsewhere. The correct solution would be a standard encoding (or maybe a couple of them) which accounts for as many as possible of the set of characters anyone on Earth might conceivably wish to type. Obviously this will require many more than 256 characters, so we'll have to do away with one character encoding to one byte (hence the divergence of codepoints and byte encodings), but it's for a greater good. Fortunately all the wrangling, tabulating and

other rigmarole has been done, and we have an answer: Unicode. This accounts for over 100,000 characters, bidirectional display order, ligature forms, and more. Currently there are two encodings in use: UTF-8 which uses 1 byte for 'common' characters (making it entirely backwards compatible with ASCII) and up to four bytes for the more cosmopolitan ones, and UTF16 which uses two bytes for some characters and four bytes for others. Unicode has been widely adopted, both as a storage encoding standard and for internally processing tests. The main raison d’etre of Python 3, is that it's predecessor did not do the latter.

print ('Hello world') Which is perfectly compatible with Python 2 and 3. If you were a fan of using a comma at the end of your print statements (to suppress the newline character), then sad news: This no longer works – instead we use the end parameter, which by default is a new line. For example print ('All on', end=" ") print ('one line') does just that.

Print in Python 3 A significant proportion of Python programs could be made compatible to 3 just by changing the print syntax, but there are many other, far less trivial, things that could go wrong. To understand them, we must first be au fait with what really changed in Python 3. Most of the world doesn't speak English, in fact, most of the world doesn't even use a Latin character set, even those regions that do tend to use different sets of accents to decorate the characters. As a result, besides the ol' ASCII standard, numerous diverse and incompatible other character encodings have emerged. Each grapheme (an abstraction of a character) is assigned a codepoint, and each codepoint is assigned a byte encoding, sometimes identically. In the past, if you wanted to share a document with foreign characters in it then plain ASCII wouldn't help – you could use one of the alternative encodings, if you knew the people you were sharing it with could do the same, but in general you needed to recourse to a word processor and with a particular font, which just moves the problem elsewhere. Thankfully, we now have a widely adopted standard: Unicode (see The Unicode revolution box, above) that covers all the bases, and is backwards compatible with ASCII and (as far as codepoints are concerned) its Latin-1 extension. We can even have Unicode in our domain names, although internally these are all still encoded as ASCII, via a system called Punycode. Python 2 is far from devoid of Unicode support, but its handling of it is done fairly superficially (Unicode strings are sneakily re-encoded behind the scenes) and some third-party modules still won't play nice with it. Strings in Python 2 can be of type str (which handles ASCII fine, but will behave unpredictably for codepoints above 127) or they can be of type unicode. Strings of type str are stored as bytes and, when printed to a terminal, are converted to whichever

The PyStones benchmark will likely be slower in Python 3, but the same won’t be true for all code. Don’t be a Py3k refusenik without first trying your code.

encoding your system's locale specified (through the LANG and LC_* environment variables in Linux). For any modern distro, this will probably be UTF-8, but it's definitely not something you should take for granted. The unicode type should be used for textual intercourse – finding the length of, slicing or reversing a string. For example, the Unicode codepoint for the lowercase Greek letter pi is 03c0 in hex notation. So we can define a unicode string from the Python console like so, provided our terminal can handle Unicode output and is using a suitable font: >>> pi = u'\u03c0' >>> print(pi) ϖ >>> type(pi) >>> len(pi) 1 However, if we were to try this on a terminal without Unicode support, things will go wrong. You can simulate such a scenario by starting Python with: $ LC_ALL=C python Now when you try and print the lowercase character pi you will run into a UnicodeEncodeError. Essentially, Python is trying and failing to coerce this to an ASCII character (the

Quick tip Arch Linux is one of few distributions to use Python 3 by default, but it can live happily in tandem with its predecessor (available in the python2 package).

Never miss another issue Subscribe to the #1 source for Linux on page 34. www.tuxradar.com


Python 3

only type supported by the primitive C locale). Python 2 also tries to perform this coercion (regardless of current locale settings) when printing to a file or a pipe, so don't use the unicode type for these operations, instead use str. The str type in Python 2 is really just a list of bytes corresponding to how the string is encoded on the machine. This is what you should use if you're writing your strings to disk or sending them over a network or to a pipe. Python 2 will try and convert strings of type unicode to ASCII (it's default encoding) in these situations, which could result in tears. So we can also get a funky pi character by using its UTF-8 byte representation directly. There are rules for converting Unicode codepoints to UTF-8 (or UTF-16) bytes, but it will suffice to simply accept that the Pi character encodes to the two bytes CF 80 in UTF-8. We can escape these with an \x notation in order to make Python understand bytes: >>> strpi = '\xCF\x80' >>> type(strpi) >>> len(strpi) 2 So ϖ apparently now has two letters. The point is: if your Python 2 code is doing stuff with Unicode characters, you'll

have to have all kinds of wrappers and checks in place to take account of the localisation of whatever machine may run it. You'll also have to handle your own conversions between str and unicode types, and use the codecs module to change encodings as required. Also if you have Unicode strings in your code, you'll need to add the appropriate declaration at the top of your code: # -*- coding: utf-8 -*The main driving force for a new Python version was the need to rewrite from the ground up how the language dealt with strings and their representations in order to simplify this process. Some argue that it fails miserably here (see for example, Armin Ronacher's rant on his blog http://bit.ly/ UnicodeInPython3), but it really depends on your purposes. Python 3 does away with the old unicode type entirely, since everything about Python 3 is now Unicode-orientated. The str type now stores Unicode codepoints, the language's default encoding is UTF-8 (so no need for the -*- coding decorator above) and the new bytes object stores byte arrays, like the old str type. The new str type will need to be converted to bytes if it's to be used for any kind of file I/O, but this is trivial via the str.encode() method. If you're reading Unicode text files you'll need to open them in binary 'rb' mode and convert the bytes to a string using the converse bytes.decode() method (see the picture for details, see bottom, p84). Python 3, however, brings about many other changes besides all this Unicode malarkey, some of these are just the removal of legacy compatibility code (Python 3, unlike 2.7, doesn't need to be compatible with 2.0), some of them provide new features and some of them force programmers to do things differently. But wait, there's more: For example, functions can now be passed keyword-only arguments even if they use the *args syntax to take variable length argument lists and catching exceptions via a variable now requires the as keyword. Also you can no longer use the ugly <> comparison operator to test for inequality, instead use the much more stylish != which is also available in Python 2.

Automating conversion with 2to3

Aleph, beth, gimel… The 2to3 program show us how to refactor our Python 2 code for Python 3. Shalom alekhem.

For reasonably small projects, and those that aren't going to encounter any foreign characters, it’s likely that your Python 2 code can be automatically made Python 3 compatible using the 2to3 tool. This helpful chap runs a predefined set of fixers on your Python 2 code, with the goal of emitting bona fide Python 3 code. Using it can be as simple as

The division bell One change in Python 3 which has the potential to flummox is that the behaviour of the division operator / is different. For example, here's an excerpt from a Python 2 session >>> 3/2 1 >>> 3./2 1.5 >>> 3.//2 1.0 which shows the first command operating on

two ints, and returning an int, thus in this case the operator stands for integer division. In the second example, the numerator is a float and the operator acts as floating point division, returning what you'd intuitively expect half of three to be. The third line uses the explicit floor division operator // which will return the rounded-down result as a float or an int depending on the arguments. The latter operator was actually backported to 2.7, so it's behaviour is the same in Python 3, but the

behaviour of the classic division operator has changed: if both numerator and denominator are ints, then an int is returned if one divides the other, otherwise a float is returned with the correct decimal remainder. If at least one of the arguments is a float (or complex), then the behaviour is unchanged. This means the / operator is now as close to mathematical division proper as we can hope, and issues involving unexpected ints will no longer cause harm.

If you missed last issue Head over to http://bit.ly/MFMissues now! 86 LXF195 March 2015

www.linuxformat.com

Python 3

The popular matplotlib module has been Python 3 compatible since v1.2, for all your graphing and plotting requirements.

$ 2to3 mypy2code.py which will output a diff of any changes against the original file. You can also use the -w option to overwrite your file, don't worry – it will be backed up first. Some fixers will only run if you specify them explicitly, using the -f option. An example is buffer which will replace all buffer types with memoryviews. These two aren't entirely compatible, so some tweaking might be necessary in order to successfully migrate. Using 2to3 can save you a whole heap o' effort, since searching and replacing print statements manually is a drudgerous task. The pictured (over on the left there, page 86) example shows the changes to a simple three-line program – the unichr() function is now chr(), since Unicode is now implicit, and the print line is reworked, even if it uses % to format placeholders.

Parlez-vous Python Trois? Another option is to create 'bilingual' code that is compatible with both Python 2 and Python 3. While you should really only target one specific Python version, this middle ground is very useful when you're porting and testing your code: You might still need to rework a few things in Python 2 while still enjoying the new features and learning the new syntax. Many Python 3 features and syntaxes have already been backported to 2.7, and many more can be enabled using the __future__ module. For example, you can get the new-fangled print syntax by using the following: >>> from __future__ import print_function This is used with the from in this way, and __future__ doesn't behave like a standard module, instead it acts as a

compiler directive which in this case provides the modern print syntax. We can likewise import division to get the new style division, or unicode_literals to make strings Unicode by default. There is another module, confusingly called future which isn't part of the standard library, but can help ease transition issues. When Python 3.0 was released, it was generally regarded as being about 10% slower than it's predecessor. This was not surprising, since speed was not really a priority for new version and many special-case optimisations were removed in order to clean up the code base. Now that we're up to 3.4 (3.5 is scheduled for final release in September 2015), it would be nice if performance had been improved. Unfortunately, this hasn’t been the case, which you can verify for yourself using the PyStone benchmark. We tested it [see p85] on an aging machine in the office, which has now come back from the dead twice and hence possesses supernatural powers far beyond it's dusty 2.3GHz processor would suggest. But don't be misled; PyStone tests various Python internals of which your code may or may not make extensive use. It’s important to test your code in both versions to get a more accurate picture. You can always use Cython [see Coding Academy, p84, LXF193] to speed up code that is amenable to C translation (loops, math, array access), or the bottleneck module. Guido van Rossum, the author of Python, says Python 2.7 will be supported until 2020, but that's no excuse to postpone learning the new version now. Python 3 adoption is increasing, so you won't be alone. Half the work is retraining your fingers to add parentheses. LXF

www.tuxradar.com


Julia

Julia: Dynamic programming Mihalis Tsoukalos explains the necessary things that you need to know to start programming in the fast and dynamic Julia language.

Our expert Mihalis Tsoukalos is a Unix admin, programmer, mathematician and DBA. You can reach him at www. mtsoukalos.eu and @mactsouk.

The Julia shell is where you do most of your experimentation. You can see that integer division results in a floating point result.

J

ulia (http://julialang.org) is a fast dynamic language for technical computing designed and developed by Jeff Bezanson, Stefan Karpinski, Viral Shah and Alan Edelman. It’s also a functional language to address the requirements of high-performance numerical and scientific computing while also being capable of doing general purpose programming. Julia uses an LLVM-based just-in-time (JIT) compiler to run its programs. The compiler combined with the design of the language enables it to approach and often match the performance of C code. You can also call C functions directly from Julia without the need for wrappers, special APIs or other nasty tricks. Julia's design and implementation is driven by the following three principles: first, It has to be fast; second, It has to be a dynamic programming language; and third, it also has to be expressive. The core of Julia is written in C but the rest of the language is written in Julia itself, which means you can see how it works behind the scenes and easily make changes. Even in 'untyped' programming languages, there's some kind of type used by the compiler. Julia thinks that if the compiler uses types, then why not enable the programmer to also use types if they so choose? In Julia, it's up to the programmer to mention the type of a variable. However, there


www.linuxformat.com

are situations where you need to talk about types, so the type system of Julia is good and sophisticated. As you can probably tell, Julia is attempting to replace R, Matlab, Octave and the NumPy module of Python in arithmetic computing by counting on faster performance. The reactions from users of its competition is good so far and we'll show you why.

Installing Julia On Ubuntu, you can install Julia with sudo apt-get install julia. After installation, you can check the install with $ julia -v julia version 0.2.1 and the Julia version of "Hello World!" is: println("Hello World!") You can save the program and execute it from the Linux shell as follows: $ julia hw.jl Hello World! The Julia shell is the best place to explore the language and try to understand its features (pictured above are some simple interactions with the Julia shell). You can exit the Julia shell by pressing Control+d or by typing quit(). You can also run Julia code from the shell without saving it first. This is very handy for small programs: $ julia -e 'for x in ARGS; println(x); end' Hello World Hello World

Julia The ~/.juliarc.jl file used by Julia is similar to the .bashrc file. It executes Julia code whenever the Julia shell is executed. $ cat ~/.juliarc.jl println("Using the .juliarc.jl file!") Julia is a case-sensitive programming language. Variable names must begin with a letter (A-Z or a-z), underscore, or a subset of Unicode code points greater than 00A0. As a result 88a is an invalid variable name whereas _99 is valid. Julia uses many types of integer variables, both signed (Int8, Int16, Int32, Int64 and Int128) and unsigned (Uint8, Uint16, Uint32, Uint64 and Uint128) as well as bool variables (Bool) and single chars (Char). Additionally, it supports three types of floating point variables: Float16, Float32 and Float64. The number at the end of a type denotes the number of bits required for that particular type. Julia also defines the Int and Uint types, which are aliases for the signed and unsigned native integer types of the system that Julia runs. So, on a 64-bit system Int is an alias for Int64.

Getting to know her Julia also supports Fractions with the help of the // symbol. So 4//5 is a fraction and isn't converted to 0.8. If the numerator and the denominator have common factors, Julia will automatically simplify the fraction; otherwise, it will leave it as it is. You can do calculations with fractions using the standard operators. Julia also supports complex numbers. The global constant im is the complex number i. So, 4 - 2im is a complex number. You can do complex number arithmetic using the standard operators. Now, let us see some Julia code. The code for calculating Fibonacci numbers is the following: $ cat fibonacci.jl fib(n) = n < 2 ? n : fib(n - 1) + fib(n - 2) The fib() function takes just one argument as its input. As you can see, you don’t have to declare its type. Should you wish to declare it, you can do it as follows: fib(n::Int8) = n < 2 ? n : fib(n - 1) + fib(n - 2) If you save your Julia code in a separate file, you can use: julia> include("fibonacci.jl") fib (generic function with 1 method) julia> fib(10) 55 > fib("test") ERROR: ìsless` has no method matching isless(::ASCIIString, ::Int64) in fib at /Users/mtsouk/docs/article/working/Julia.LXF/code/ fibonacci.jl:2 julia> fib(' ') 2178309

Profile module The Profile module provides tools that help developers to improve their code. The module takes measurements on running code and helps you to understand how much time is being spent on each executed line of code. You run the function you want to profile once in order to compile it and run it again by putting the @profile keyword in from of it. Last, you execute the Profile.print() command to get the output.Julia at a fast pace.

julia> fib(32) 2178309 julia> @elapsed fib(40) 0.883197884 Although you can't find the Fibonacci number for the "test" string, you can find the Fibonacci number for a single character; Julia automatically uses the ASCII code of the character as the input to the fib() function! The @elapsed macro evaluates an expression and returns the number of seconds it took to execute as a floating-point number, discarding the resulting value. It is very useful for benchmarking your code. Two other useful macros are called @linux and @unix. They can help you identify the operating system your program runs on and act accordingly: julia> @linux? println("Linux!") : println("Not Linux!") Not Linux! julia> @unix? println("UNIX!") : println("Not UNIX") UNIX! Not Linux! Despite its simplicity, Julia is also a programming language for the advanced programmer. The code_native() function enables you to look at the Assembly code the JIT compiler generates for the println() function [pictured, bottom of p90]. If you are comfortable with Assembly code, you can clearly understand that Julia has the potential of being almost as fast as C code, despite the fact that's primarily a dynamically typed language. This kind of optimised code can’t be generated for Python or Ruby easily, because the JIT compiler can’t know the type of arguments being passed. Calling a C function is done using the ccall() function. The following code calls the getpid() system call to get the process ID of the calling process: julia> pID = ccall((:getpid, "libc"), Int32, ()) 44763 julia> typeof(pID) Int32

Quick tip Julia supports TCP sockets. By just running listen(1234) you can create a socket that listens to port number 1234 on localhost. You can then connect to it in Julia by typing connect(1234). If you give client = connect(1234), you should then close the socket by typing close(client).

Why Julia was created? As the creators of Julia say: "We want a language that’s open source, with a liberal licence. We want the speed of C with the dynamism of Ruby. We want a language that’s homoiconic, with true macros like Lisp, but with obvious, familiar mathematical notation like Matlab. We want something as usable for general programming as Python, as easy for statistics as R, as natural for

string processing as Perl, as powerful for linear algebra as Matlab, as good at gluing programs together as the shell. Something that is dirt simple to learn, yet keeps the most serious hackers happy. We want it interactive and we want it compiled. (Did we mention it should be as fast as C?)" Not asking for much then. Julia also wants to make the development of numerical

code better, faster and more efficient than with other programming languages. The library, largely written in Julia itself, also integrates mature C and Fortran libraries for linear algebra, random number generation, signal processing, and string processing. In addition, the Julia community provides a large number of external packages through the built-in package manager.

If you missed last issue Call 0844 848 2852 or +44 1604 251045 www.tuxradar.com


Julia Most operators in Julia are just functions; therefore 1+2+3 is equivalent to writing +(1,2,3). You can also execute expressions, such as +(-(1,2),3). Anonymous functions are also supported in Julia and their main usage is for passing them as arguments to other functions. Functions in Julia can return multiple values or other functions. The following program reads an integer from the user and returns its factors: function factors(n) f = [one(n)] for (p,e) in factor(n) f = reduce(vcat, f, [f*p^j for j in 1:e]) end return length(f) == 1 ? [one(n), n] : sort!(f) end The result of the factors() function is an array that contains the desired numbers. You can use the input() function to get user input: function input(prompt::String="") print(prompt) chomp(readline()) end You can use the factors() function in combination with the input() function to automatically process user input. The trick can be done as follows: julia> factors(int(input("Give me a number: "))) So the readline() function enables you to get user input and the print() or the println() functions enable you to print output on screen. The difference between the two functions is that println() automatically prints a newline character. Julia has error handling capabilities. The following output shows how to handle division by zero errors when doing integer division: julia> try div(100, 0) catch x println(typeof(x)) end DivideError Julia treats arrays as first class citizens with powerful functionality. You can easily generate a new vector with 100 random integer elements and name it myVec using the myVec = rand(Int32, 100) command. The first element of

myVec can be accessed by using the myVec[1] notation; therefore the first element has an index number of 1, not 0. Similarly, its last element is accessed as myVec[100]; not myVec[99]. The last element of a list or array can also be accessed using the end symbol as its index. So far it may appear that arrays in Julia are the same as in any other programming language. Where Julia truly excels is in its multidimensional array capabilities: julia> anArrayA = [-1 -2 -3; 1 2 3; 10 20 30] # command 1 julia> anArrayB = [1 2 3; -1 -2 -3; -10 -20 -30] # command 2 julia> arrayAPlusB = anArrayA + anArrayB # command 3 julia> productAB = anArrayA * anArrayB # command 4 The first command creates an array named anArrayA with two dimensions. The second command creates a second array named anArrayB. The third command adds the two arrays and stores the result in a variable called arrayAPlusB. Similarly, the fourth command finds the product of the two arrays and saves it using a new variable. As you can see, you can do matrix calculations using the standard operators. The anArrayA .+ 2 command adds 2 to all elements of the array! The sum(anArrayA, 1) command calculates the sum over the first dimension of the array: (-1 + 1 + 10), (-2 + 2 + 20) and (-3 + 3 + 30). Julia also supports Sparse Matrices. A sparse matrix is a matrix most of whose values are equal to zero. To reduce storage costs, such matrices are represented by storing only the values and coordinates of their non-zero elements.

Strings in Julia Strings are defined with double quotes and can contain any Unicode character. Single characters are defined using single quotes. There exist the usual functions for converting a string to uppercase or lowercase, called uppercase() and lowercase() respectively. The length() function can be used for getting the length of a string. String concatenation is also supported with the help of the * operator, instead of the + operator. Alternatively, you can concatenate strings using the string() function. Searching a string for a char or another string can be done with the search() function. Similarly, the replace() function, which takes three arguments, replaces a string with another. Search() also supports regular expressions. Three other handy functions are repeat(), matchall() and eachmatch() and their use is illustrated in the following examples: julia> thisx3 = repeat("This ", 3) "This This This " julia> "This" ^3 # The same as repeat("This",3) "ThisThisThis"

External programs

Seeing the Assembly code of your Julia code is easy. Here’s what Assembly makes of the println() function, for instance.

Julia also enables you to run external programs and get their output. You can execute the ls command by typing run(`ls`). You can assign its output to a new variable by typing lsOutput = readall(`ls`). If you want to run the variable command in your shell, you can do it by putting an $ in front of the variable name, for example: run(`$command`). Please be warned that this is not a safe way to write a program!


www.linuxformat.com

Julia julia> r = matchall(r"[\w]{4,}", thisx3) 3-element Array{SubString{UTF8String},1}: "This" "This" "This" julia> r = eachmatch(r"[\w]{4,}", thisx3) RegexMatchIterator(r"[\w]{4,}","This This This ",false) julia> i = 1 # have to declare it first 1 julia> for(word in r) println("Match ", i) i += 1 # i++ is not supported end Match 1 Match 2 Match 3 The matchall() function returns a vector with RegexMatches for each match whereas the eachmatch() function returns an iterator over all matches.

Implementing Bubble Sort Julia’s Bubble Sort algorithm, as you can see below, has a slightly different format than the C for loop: function bubbleSort(inputVec) elements = length(inputVec)-1 for vecIndex in elements:-1:1 for pass in 1:vecIndex # Swap them if needed if inputVec[pass] > inputVec[pass + 1] tmp = inputVec[pass] inputVec[pass] = inputVec[pass + 1] inputVec[pass + 1] = tmp end end end return inputVec end In order to get the vecIndex numbers in reverse order, it is necessary to use the elements:-1:1 format. Comments in Julia begin with the hash character. You can now use the bubbleSort() function to sort an array that is called myVec by executing the following command print(bubbleSort(myVec)) You can also use the Pkg.status() command to show you information about the installed packages. (The image, right,

Here’s how to install the Date Julia package to construct a date.

shows how to install and use a Julia package.) You can add a package with the Pkg.add() command and remove it with the Pkg.rm() command. Last, the Pkg.update() command enables you to update every installed package to its latest and greatest version. Julia also allows you to look at the source code and full development history of all installed packages, and you can make changes and fixes to them. So, for instance, you can find the source code of the Dates package inside the ~/.julia/v0.3/Dates/ directory. Should you wish to generate your own package, then you can begin by using the command Pkg.generate(). Julia allows you to work with both text and binary files. This section will show you how to work with text files. The presented program reads a text file, line by line and converts each line in uppercase, and saves the output into a new text file. The code looks like this: input = ARGS[1] output = ARGS[2] OUTPUT = open(output, "w") open(input,"r") do f for line in eachline(f) print(OUTPUT, uppercase(line)) end end close(OUTPUT) When you are using do to read a file, the file closes automatically when Julia reaches its end. Otherwise, you should close the file using the close() command. Julia has many more capabilities than the ones presented in this article, and as a modern programming language it has many handy features mainly for scientific computing. LXF

Next issu MySQL e: MariaDB

Text plotting in Julia TextPlots is an extremely simple plotting library which generates terminal plots using Braille characters. This enables you to plot any continuous real-valued function or any relatively small collection of data points. The image (right) shows the output of the following commands: julia> Pkg.add("TextPlots") julia> using TextPlots julia> plot(cos) julia> plot([x-> cos(x) * sin(x), x-> x/5], -6:6) As you can understand, TextPlots allows you to plot multiple function simultaneously and define the range of

the values of x. The following code changes the output of the plot() command and produces different kinds of plots: julia> plot(cos, border=false) julia> plot(cos, border=false, title=false, labels=false) julia> plot([1, 2, 5, 8], [10, 11, 12, 3]) The last command draws pairs of (x,y) points on screen using plot(). You can find more information about how to use the TextPlots Julia package at https://github.com/sunetos/ TextPlots.jl. Other useful packages for plotting that you may like to try are Winston, PyPlot and Gadfly.

www.tuxradar.com

The TextPlots package enables you to plot on your terminal using Braille characters and is very handy when you don’t have a graphical interface.


Got a question about open source? Whatever your level, email it to [email protected] for a solution.

This month we answer questions on: 1 Fedora 21 package manager 2 Backing up onto DVD 3 The Case of the Disappearing Printers

1

4 Copying Linux Format DVDs to

USB sticks 5 Dealing with Misbehaving hotkeys + UEFI booting

Apper not appy

I recently downloaded and installed from Fedora-Live-KDE-i686-21-5.iso. The install went smoothly and I have successfully applied all available updates without any problems. I was impressed by how quickly and smoothly this proceeded. However, the system does have a problem in that it cannot install extra packages from the standard Fedora repo. Clicking on System Settings and then on Software Management takes me to the Lists and Groups section. When I click on any of the icons within the

Fedora 21’s KDE package manager falls over if you try to browse groups. For now, use a different package management frontend, such as yumex.

Groups section, the system throws up a dialog stating "A problem that we were not expecting has occurred. Please report this bug with the error description." Clicking on this dialog's Details button reveals SearchGroups not supported by backend.

Enter our competition

Win!

if ($letter == winner) get $books Get your questions answered and exploit our generosity. Linux Format is proud to produce the biggest and best magazine Get into Linux today! about Linux and free software that we can. A rough word count of LXF193 showed it had 55,242 words. That’s a few thousand more than Animal Farm and Kafka’s The Metamorphosis combined, but with way more Linux, coding and free software (but hopefully less bugs). That’s more than most of our competitors, and as for the best, well… that’s a subjective claim, but it’s one we’re happy to stand by.


As we like giving nice things to our wonderful readers, the Star Question each month will win a copy of Martin O’Hanlon’s Adventures in Minecraft, and it’s a great intro to fun Minecraft API projects for you or a youngster. For a chance to win, email a question to lxf.answers@ futurenet.com, or post it to www.linuxformat. com/forums to seek help from our lively community of readers. See page 94 for our star question.

www.linuxformat.com

I'm at a loss to understand this error. Have I downloaded the correct ISO file? The ISO file name does mention i686, am I correct in thinking that PI is i586 and PII is i686? Is there some way that SearchGroups can be enabled in the backend? stuarte9, From the forums You are right in thinking that i686 is 32-bit, PII or later. 64-bit ISOs are generally marked with amd64 or x86_64. This is a known issue where Apper, the KDE package manager for Fedora, tries to call a function that‘s not implemented on the package management backend. This is one of those situations where it’s always someone else's fault, depending on who you ask. Is it the backend's fault for not implementing the function or the frontend's fault for not gracefully falling back to an alternative. Whoever it is they will eventually need to fix it, but until then the Groups functionality of Apper is currently broken. That means you need to use an alternative package management front end for the time being. This can be the command line with yum or a different graphical interface. For those that installed from LXFDVD194, which contains both Gnome and KDE frontends, there’s already the choice of gnome-software. As you installed from the KDE-only ISO, you may choose to install gnome-software from the command like with yum install gnome-software then use gnome-software to browse available packages. As this is a Gnome program, installing it on a KDE system will bring in a large

Answers Terminals and superusers We often give a solution as commands to type in a terminal. While it is usually possible to do the same with a distro’s graphical tools, the differences between these mean that such solutions are very specific. The terminal commands are more flexible and, most importantly, can be used with all distributions. System configuration commands often have to be run as the superuser, often called root. There are two main ways of doing this, depending on your distro. Many, especially Ubuntu and its derivatives, prefix the command with sudo, which asks for the user password and sets up root privileges for the duration of the command only. Other distros use su, which requires the root password and gives full root access until you type logout. If your distro uses su, run this once and then run any given commands without the preceding sudo.

number of dependencies. Alternatively, you could use YUM Extender (yumex), the package manager used by LXDE, which has a smaller dependency footprint. You will see that browsing groups in gnome-software does not give the same problems as with Apper. This issue only relates to the software groups features, updates or searching and installing packages by name works fine in Apper.

2

Scripts and hotkeys

I am using Xubuntu 14.10. I have installed xclip and written a one-line shell script to read the system date: date "+%A %d %B %Y" | xclip -selection clipboard I have also written a one-line shell script that outputs the date: xclip -o -selection CLIPBOARD Both of these work from the command line.

As a refinement, I made a keyboard shortcut to each of these scripts. The script which reads the date works when invoked by the keyboard shortcut. The script which should output the date appears to do nothing. I have the feeling that I am missing something obvious but I can't think what it is. elenmar, From the forums Diagnosing these problems can be tricky because you cannot see any messages from the command when running it from a keyboard shortcut but running it from a terminal changes its execution environment. When you open a shell session in a terminal, it executes your profile, usually ~/.bash_profile or ~/.bashrc, which doesn't happen when running a script from a keyboard shortcut (or cron task for that matter). So how do you get to see the output from your script when it is not connected to a terminal window? By adding these lines near the top of the script: exec >>~/myscript.log 2>&1 set -x The exec command can be used to change redirections for a shell process, this line redirects both standard output and standard error to a file. The first redirection sends standard output to the file, then 2>&1 sends standard error to stdout. Now you can inspect the output from the program after calling it from a hotkey. You can view the output in real time by running this in a terminal tail -f ~/myscript.log and then pressing the key. The other line set -x tells the shell to print each command before it runs it, adding this lets you see exactly what the script is doing and which commands are

creating any messages. If you suspect the problem is caused by the different execution environments, you can make sure your script loads the profile before running anything with one of: source ~/.bash_profile source ~/.bashrc Alternatively, you can have a single script that simply outputs the date in the format you need, but I assume you have a reason for storing that date and using it later. Be aware that any other program could change or clear the clipboard's contents, so it may be safer to use a temporary file rather than the clipboard.

3

Backing up DVDs

I want to rip, store and back up some of my DVD collection onto my NAS so that I can watch them via my Raspberry Pi and XBMC. Ideally, I'd like to store the files in MKV or MP4 format. I looks like I will need to remove any copy protection, save the copy as an ISO and then convert the resulting file to either MKV or MP4. So what are your favourite programs to do this? Noob_Computa_Ninja, from the forums First of all, you have to decide how much of each DVD (please note it may be illegal in your territory to copy DVDs or circumvent DRM protection, in the UK it’s now legal for personal use) you want to keep. Do you want the whole DVD, complete with menus and features, or just the main title? If you want the whole DVD, and have libdvdcss installed on the player, you can simply copy the entire DVD to an ISO file with this command. cp /dev/dvd name.iso This gives you the truest copy, but at the

A quick reference to...

Finding files inux has two main tools for finding files; locate and find. The former works by keeping a database of files on your system. It is very fast, but limited to files that were present when the database was last updated. Most distros install a cron script to keep the file database up to date. Locate is also limited to searching on file names, eg: locate somefile To add -i for a case-insensitive search. The alternative is find, which searches the filesystem directly. It’s slower, and can only search areas the users have permission to read, but provides completely up to date information. It also enables you to specify which directories to search or exclude and search on properties other than the file

L

name, such as the owner or creation date. find -name '*somefile*' find /usr -iname '*someotherfile*' find /usr -maxdepth 2 -iname '*whatfile*' The first line starts the search in the current directory, and descends into all subdirectories. The second starts in /usr instead, and performs case-insensitive searching. The third does the same, but descends a maximum of two directories. With more options, find is a far more flexible option, but locate is great for quick searches. You can pipe locate's output through grep to search specific directories locate -i myfile | grep /home/ Note also that locate searches for a substring, whereas find looks for an exact match, hence the use of * as a wildcard. Have you ever wondered where a program keeps its config files? Try using

www.tuxradar.com

Kfind adds a friendly face to the powerful find command. touch /tmp/now

then run the program, edit its preferences and quit. Then do find ~ -newer /tmp/now to find all files that have changed, which will include your program's configuration and, probably, little else.


Answers expense of storage space, each ISO image can be up to 9.4GB in size. If you want a single movie file containing the main feature of each DVD, there are a number of programs you can use, both command line and GUI, once again depending on your needs. If you want a straight copy of the title with no transcoding, you can use MPlayer's dump options. mplayer -dumpstream -dumpfile movie.mpg dvd://1 This will dump the first track of the DVD to a file instead of playing it. The only processing done is to remove any encryption. If you want to transcode to a different format, a good example of a GUI program for this task is HandBrake (which comes with a command line tool too). HandBrake should be in most distros' package repositories, so install it in the usual way and start it up. While HandBrake doesn't appear to have an option to read directly from a DVD, by pressing the Source button and then selecting /dev/dvd (or /dev/sr0, depending on your system) it will read the DVD in that drive and enable you to choose the title to transcode, then you can set encoding parameters. This is an important consideration when using a Raspberry Pi for playback. You need to choose codecs and compression that work well with your Pi. Fortunately HandBrake has a couple of features that help with this. You can encode a small part of a video for testing, enabling you to try various options without having to transcode the whole DVD every time, and you can save these settings in a preset for easy recall next time. You may as well start with the Normal

Star Question +

Winner!

4

HandBrake makes transcoding video easy, and lets you define your own custom presets.

LXFDVD to USB

Why can I not make a bootable USB from the LXF192 DVD, please? I have tried on two different 64-bit machines. I’ve booted from the LXF192 disc, loaded the Ubuntu 14.10 Mate version and I’ve used disc creator. After 20-30 minutes it gave me an error and the process stopped. I tried again on a 64-bit Compaq HP laptop, erased the USB stick and again after 20-30 minutes the same thing happened again. Why? I never seem to be able to use LXF discs to make bootable USB from the attached DVD that comes with my subscription. Am I the only one who has this kind of trouble?

el10, From the forums Disc creators like the one you tried, and generic ones, such as Unetbootin, make certain assumptions that are not always true. In this case, it assumes that you have booted from an Ubuntu disc and that is what it expects to find in your optical drive. In fact, there's a completely different disc in there and you have booted from an ISO image file on that disc. Enough of why that approach doesn't work, the good news is that not only is what you want possible, it is far simpler than messing around with various disc creators. Both the LXFDVDs themselves and the various distros ISO images on them are what are known as hybrid images. This means they can be used on either an

This month’s winner is Peter Woolley. Get in touch with us to claim your glittering prize!

UEFI conundrum

I read Neil Bothwick’s article on configuring Grub,[p78, LXF193] however, it didn't answer the problem that I’m experiencing. I bought a Lenovo Flex 10, having been happy with Lenovo machines before. Unlike previous models this hasn’t been easy to get Linux working. It has a 32-bit UEFI firmware and a 64-bit processor. This makes installation much harder, though not impossible. I’ve got as far as booting into Ubuntu 13.04 using a pendrive I prepared using rufus in Windows. From this, I have run Gparted and set up my drive as follows: Disk partition table set as GPT: Partition Filesystem Size Flag Mountpoint /dev/sda1 FAT32 511MB boot /dev/sda2 EXT4 10GB none / /dev/sda3 Swap 1GB none However, I am unable to boot using the HD. It isn't recognised by the firmware and I


preset, and leave the video codec as H.264 but change the audio codec to AAC. Try encoding a short clip and see how it plays back. Then you can adjust the settings if the playback is not smooth enough for you. Once you have the settings you want, click the Save icon below the Presets list. If you want encode a number of DVDs, and have sufficient space, you can rip each to an ISO file as above, then set up encoding in HandBrake and add each one to your ripping queue. Then you can select Queue > Start Queue to batch process all the DVDs.

get a kernel panic. The strings I use from the Grub console are: linux (hd1,gpt2)/boot/vmlinuz-3.8.0-19-generic root root=dev/sda/2 initrd (hd1,gpt2)/boot/initrd img-3.8.0-19generic I was wondering if I can avoid all this by using an EFI stub loader? Right now I just have an expensive paperweight! Peter Woolley Getting started with UEFI can be a little daunting, especially if you have a 32-bit UEFI. There are a couple of significant points to bear in mind. The first is that it’s not enough to make your EFI System Partition FAT format, it must also be marked as such in the partition table, by setting the partition type to EF00. There also seems to be some confusion over your disk identification. As a laptop, I would expect it to have a single disk, /dev/sda in Linux terms, as described in your partition table, but this is hd0 in Grub

www.linuxformat.com

terms. Grub 2 has taken the rather confusing step of counting disks from zero but partitions from one, so sda2 is hd0,gpt2. The other important point is the chicken and egg situation of installing Grub for a UEFI system. In order for Grub to install to the right place, it needs access to UEFI information from your EFI firmware, which it gets from /sys/firmware/efi/efivars. If this directory is empty, or doesn’t exist, Grub will install as if for a BIOS boot, but this information is only available if you’re already booted using the EFI firmware, so check the files are there before trying to install Grub. Another factor is your mixing of 32-bit UEFI with 64-bit CPU and OS. The Linux kernel supports this, but only from version 3.15 onwards. Ubuntu 13.04 comes with version 3.8.8 of the kernel, which doesn’t support what you want to do. Having said all that, you may well have already read the follow-up to that article [p74, LXF194], which explains Gummiboot for EFI booting.

Answers optical disk or a USB flash drive. To copy the entire DVD to a flash drive, run the following: sudo dd if=/dev/dvd of=/dev/sdX bs=4k sync Note you must get the mount point right as giving the wrong device will trash any data on it. To be sure, run: dmesg -w and then plug in the flash drive. You will see something like this usb-storage 1-2:1.0: USB Mass Storage device detected scsi host7: usb-storage 1-2:1.0 scsi 7:0:0:0: Direct-Access SanDisk Cruzer Edge 1.20 PQ: 0 ANSI: 5 sd 7:0:0:0: [sdb] 31266816 512-byte logical blocks: (16.0 GB/14.9 GiB) so in this case you would want /dev/sdb. Use the device name without a trailing number, you are writing to the whole device. This gives you a copy of the whole DVD, so it is a fairly slow process. The alternative is to copy just the ISO image you want. In your example of wanting the Ubuntu image, you would use the same command but set it to the path to the ISO image, instead of the whole drive: sudo dd if=/media/LXFDVD192/ubuntu-14.10desktop-remix-amd64.iso of=/dev/sdX bs=4k sync The sync command makes sure that all data is written to the flash drive before the command prompt returns, to avoid unplugging it too soon. In either case, the flash drive should not be mounted while writing to it.

5

Invisible printer

Since upgrading from Ubuntu 14.04

to 14.10 I have had no end of trouble in getting the computer to see my Officejet 4620 printer, there was no problem with 14.04. Strangely enough, provided that I switch the printer on for a minute or two prior to starting the computer, there seems to be no problem. However, if I start the computer first, followed by the printer then I cannot print at all, and I have to restart the computer to solve the problem. Not only that, but when I tried to connect my camera to the computer for the first time, the computer couldn’t 'see' it either. The camera has a straightforward USB connection. Next, I tried it on another computer running a Ubuntu 14.10 OS that has nothing on it except a downloaded 14.10 fresh install. That computer could not see the camera! Then I connected the camera to a laptop running Ubuntu 12.04 and was able to transfer my photos to it – no problem! Bryan Mitchell This looks like a problem with hotplug detection on your computer. The first thing to do is to check whether the kernel even sees the appearance of the printer. Open a terminal and run: dmesg --follow Then connect the printer and watch the output. Is udev detecting the printer? Switching it on before booting means the printer is already there when udev and CUPS start up. If you load the CUPS web interface at http://localhost:631 – is the printer visible? If it is visible but disabled, you can enable it from the web interface or with this command: sudo cupsenable printername

Help us to help you We receive several questions each month that we are unable to answer, because they give insufficient detail about the problem. In order to give the best answers to your questions, we need to know as much as possible. If you get an error message, please tell us the exact message and precisely what you did to invoke it. If you have a hardware problem, let us know about the hardware. If Linux is already running, you can use the Hardinfo program (http://hardinfo.berlios.de) that gives a full report on your hardware and system as an HTML file you can send us. Alternatively, the output from lshw is just as useful (http://ezix.org/project/wiki/HardwareLiSter). One or both of these should be in your distro’s repositories. If you are unwilling, or unable, to install these, run the following commands in a root terminal and attach the system.txt file to your email. This will still be a great help in diagnosing your problem.

uname -a >system.txt lspci >>system.txt lspci -vv >>system.txt

If CUPS doesn't see the printer at all, you could try running sudo /etc/init.d/cups restart after powering up the printer, to force CUPS to detect it. It's a bit of a sledgehammer to crack a nut, but a far smaller hammer than rebooting the whole computer. The camera problem appears to be separate and a bug has already been raised for it on Launchpad (http://bit.ly/LXF195bugs) where the blkid command mis-identifies some filesystems. This may have been fixed in the repositories by the time you read this answer, if not there’s a Deb file you can download from http://packages.ubuntu.com/vivid/libblkid1 (this is referenced in that bug report's comments) that solves the issue. LXF

Frequently asked questions…

Mobile broadband What is this mobile broadband I keep hearing about? Is is some sort of ADSL over a mobile phone connection? Sort of. It uses the 3G mobile phone network, but the technology is not ADSL, it uses HSPA (High Speed Packet Access) which is designed for mobile use. This is the technology used by smartphones that require a constant (or at least frequent) connection to the Internet. Does this offer a broadband level of speed? No, it's not broadband in the original sense of the word (neither are some of the slower landline connections for that matter) but it is fast enough for important tasks,

such as reading email and watching YouTube videos. The companies bandy around various speeds, but these are all heavily dependent on signal strength and other related factors. What do I need to use it? You need a broadband modem and an account with a mobile telecoms company. Most of these include the modem, but you can use your own. The telcos all use basically the same modem, a USB dongle, which takes a SIM card to authenticate the connection via your account. What does it cost? As with most technology, the price varies according to the supplier,

the length of contract and the data allowance. Expect to pay around £15 in the UK, which will give you around 3GB of data transfers per month. How long do I need to commit to generally? Contracts vary in length from one month to 18 months, or you can get pay as you go services. The longer term contracts come with a free modem. How well does this technology work with Linux? Very well nowadays. As all the companies seem to provide modems from a single manufacturer, Huawei, and this has drivers in the Linux kernel, it should ‘just work’. You need to set it up as a dialup connection, using

www.tuxradar.com

something like KPPP or GnomePPP, with the modem device as /dev/ttyUSB0. Because it uses the SIM card for authentication, you can put whatever you like for username and password (although the software does expect you to provide something for these). Does it matter which distro I use? The standard PPP setup should work with any distro. Those that use Network Manager have an advantage as this now supports 3G modems. With such a distro, plugging in the modem for the first time can pop up a requester for you to select your provider, then it sets it all up for you. After that, you can put the modem on and offline through the Network Manager menu.


On the disc Distros, apps, games, books, miscellany and more…

The best of the internet, crammed into a phantom-zone like 4GB DVD.

Distros fter installing each of the distros from this month's DVD, the differences between the installers was clear. Ubuntu has a clear and simple graphical installer. Fedora, the oldest and most experienced in this field, has the rather pretty but at times confusing Anaconda installer while ArchBang opens a terminal window, albeit a very attractive one, and throws questions at you in plain text. At first ArchBang's approach seems rather geeky and unfriendly to those preferring the GUI, but it has some significant advantages. All the stages of the installation process are presented in the main menu, you can jump to where you want at the press of a key – try doing that with a GUI. You can immediately see how far you are through the usual Q&A ritual of an installation and it doesn't do anything until you’ve answered all the questions and told it to proceed. That’s a little slower but a lot more comforting. In our desire to make everything as user-friendly as possible with a nice pointy-clicky GUI have we actually gone the other way? Have we made it look more friendly while making it less functional? Friendly is good, but dumbing down is not.

A

Important

NOTICE! Defective discs

In the unlikely event of your Linux Format coverdisc being in any way defective, please visit our support site at www.linuxformat.com/dvdsupport for further assistance. If you would prefer to talk to a member of our reader support team, email us at discsupport@futurenet. com or telephone +44 (0) 1225 822743.


An intermediate user’s distro

64-bit

Fedora & Ubuntu On last month's LXFDVD, we included a remix of Fedora 21, with several extra desktops. But as we’ve gone in-depth with the latest releases of

A newbie friendly distro

Fedora (there are three distinct versions now: Cloud, Server and Workstation) in one our features this issue (see p54), it made sense to have it again so you can try out the new features that Jonni describes. Note, however, that this time it only has its standard Gnome 3 desktop (or 3.14 to be precise). As is so often the case, there was a flurry of updates in the weeks following the release, and so many more people using it resulted in more bugs being discovered and fixed. So this month's Fedora offering is also a respin, this time with all the updates from the six weeks after the initial release. The Ubuntu 14.10 on the LXFDVD is also updated.

32-bit

PCLinuxOS 2014.1 In the dim and distant past, before Ubuntu became the world’s most popular distro, Mandrake was top of the heap. The merger with Connectiva and then the fork to Mageia are fairly well known, but there was an earlier fork of Mandrake called PCLinuxOS. This started life as an extra software repository for Mandrake – probably the main source of unofficial packages – and it gradually became a distro in its own right. It was, and still is, an easy to use distro ideally suited to new Linux users. It was also the distro that introduced the idea of a live CD that could be used to both demo a distro (or use a distro) and then install the distro to a system directly once a user was satisfied with the experience. Note, however, that the current PCLinuxOS live boot doesn’t play nicely

www.linuxformat.com

with USB devices, either flash disks or external DVD drives. You will need to boot from an internal DVD drive, using either the LXFDVD or burning the PCLinuxOS ISO image to a DVD, in order to use it.

New to Linux? Start here What is Linux? How do I install it? Is there an equivalent of MS Office? What’s this command line all about? Are you reading this on a tablet? How do I install software? Open Index.html on the disc to find out A lightweight distro

32-bit

ArchBang 2015.01 ArchBang is another lightweight distribution that takes its inspiration from the excellent #!CrunchBang Linux. As the other part of the name implies, this is one of the increasing number of distros using Arch Linux as a base.

If you are impressed by the live CD experience and want to install ArchBang, beware. The installer isn’t a friendly mouse-driven affair. It’s a text-based system that asks you a series of questions, so you need a reasonable level of experience and knowledge about your system. However, it doesn’t touch your disk until you have answered all the questions, so it’s perfectly safe to try. When we tried a test installation, we chose Grub as the bootloader and the installed system hung without booting. Trying again with the Syslinux choice resulted in a trouble-free setup. Lightweight desktops are often touted as ideal for older hardware, but they fly on a modern computer, leaving more of your system’s resources to do the actual work.

A NAS distro

Download your DVD from www.linuxformat.com

And more! Free eBook LibreOffice: The Complete Guide Get everything you need to master the powerful productivity suite, LibreOffice for the unbeatable price of, well, nought at all. Bargain! This month’s LXFDVD includes an eBook from our TechBook series that would normally set you back £5.99, and in it you’ll find a 100 pages of guides and tips, all written in plain English by our office experts. The guide goes through all the applications in the suite. For example, you’ll learn how to make a stylish newsletter in Writer, use master documents and make mail merges. In Calc & Math, we’ll show you how to manage your money, split unwieldy worksheets and make complex spreadsheets more manageable with pivot tables. In Impress, you’ll learn to create knockout presentations and in Draw we’ll dive into flowcharts and logos. LibreOffice even includes a database app and we’ll help you build a DB and how to use forms and queries.

64-bit

OpenMediaVault Our final distro this month is specialised, but an excellent choice for that perfectly good computer that has been gathering dust since you bought yourself a new one for Christmas. OpenMediaVault 1.9 turns a computer into a NAS (Networked

Attached Storage) box. This isn’t a live distro, you need to install it to your hard drive, so copy the ISO image from the DVD to a CD or USB stick and boot with that. Once installed, it will give you a URL that you can use to admin and use the NAS.

System tools Essentials Checkinstall Install tarballs with your package manager. GNU Core Utils The basic utilities that should exist on every operating system. Hardinfo A system benchmarking tool. Kernel Source code for the latest stable kernel release. Memtest86+ Check for faulty memory. Plop A simple boot manager to start operating systems. RaWrite Create boot floppy disks and other images in Windows. SBM An OS-independent boot manager with an easy-to-use interface. WvDial Connect with a dial-up modem.

www.tuxradar.com


Get into Linux today! Future Publishing, Quay House, The Ambury, Bath, BA1 1UA Tel 01225 442244 Email [email protected] 20,238 January – December 2013 A member of the Audit Bureau of Circulations.

LXF 196

will be on sa le Thursday 1 9 March 201 5

EDITORIAL Editor Neil Mohr [email protected] Technical editor Jonni Bidwell [email protected] Operations editor Chris Thornett [email protected] Art editor Efrain Hernandez-Mendoza [email protected] Editorial contributors Neil Bothwick, Chris Brown, David Eitelbach, Kent Elchuk, Matt Hanson, Russ Pitt, Les Pounder, Mayank Sharma, Richard Smedley, Alexander Tolstoy, Mihalis Tsoukalos, Steven Wu Illustrations Shane Collinge

ADVERTISING For ad enquiries please contact: Key Accounts - sales manager Richard Hemmings [email protected]

MARKETING Marketing manager Richard Stephens [email protected]

PRODUCTION AND DISTRIBUTION Production controller Marie Quilter Production manager Mark Constance Distributed by Seymour Distribution Ltd, 2 East Poultry Avenue, London EC1A 9PT Tel 020 7429 4000 Overseas distribution by Seymour International

LICENSING International director Regina Erak [email protected] Tel +44 (0)1225 442244 Fax +44 (0)1225 732275

Next month: The best 100 open source tools ever! Don’t miss out on the best FOSS tools, utilities and apps known to man, well, the LXF team at least.

Build a Tor box We (finally) explore how open source and a bit of electronics can create an anonymising box.

CIRCULATION Trade marketing manager Juliette Winyard Tel 07551 150 984

SUBSCRIPTIONS & BACK ISSUES UK reader order line & enquiries 0844 848 2852 Overseas reader order line & enquiries +44 (0)1604 251045 Online enquiries www.myfavouritemagazines.co.uk Email [email protected]

MANAGEMENT Content & Marketing director Nial Ferguson Head of Content & Marketing, Technology Nick Merritt Group editor-in-chief Paul Newman Group art director Steve Gotobed Editor-in-chief, Computing Brands Graham Barlow LINUX is a trademark of Linus Torvalds, GNU/Linux is abbreviated to Linux throughout for brevity. All other trademarks are the property of their respective owners. Where applicable code printed in this magazine is licensed under the GNU GPL v2 or later. See www.gnu.org/copyleft/gpl.html. Copyright © 2015 Future Publishing Ltd. No part of this publication may be reproduced without written permission from our publisher. We assume all letters sent – by email, fax or post – are for publication unless otherwise stated, and reserve the right to edit contributions. All contributions to Linux Format are submitted and accepted on the basis of non-exclusive worldwide licence to publish or license others to do so unless otherwise agreed in advance in writing. Linux Format recognises all copyrights in this issue. Where possible, we have acknowledged the copyright holder. Contact us if we haven’t credited your copyright and we will always correct any oversight. We cannot be held responsible for mistakes or misprints. All DVD demos and reader submissions are supplied to us on the assumption they can be incorporated into a future covermounted DVD, unless stated to the contrary. Disclaimer All tips in this magazine are used at your own risk. We accept no liability for any loss of data or damage to your computer, peripherals or software through the use of any tips or advice. Printed in the UK by William Gibbons on behalf of Future.

Stay anonymous online The ultimate showdown of anonymising distros, and we have no idea who’s writing it!

Build a home router Take an old system and create the ultimate home router run by Linux, mad-genius Jonni is your guide. Contents of future issues subject to change – hang on, I recognise that Tor box from last month…


www.linuxformat.com

Future is an award-winning international media group and leading digital business. We reach more than 49 million international consumers a month and create world-class content and advertising solutions for passionate consumers online, on tablet & smartphone and in print. Future plc is a public company quoted on the London Stock Exchange (symbol: FUTR). www.futureplc.com

Chief executive Zillah Byng-Maddick Non-executive chairman Peter Allen &KLHIÀQDQFLDORIÀFHURichard Haley Tel +44 (0)207 042 4000 (London) Tel +44 (0)1225 442 244 (Bath)

We are committed to only using magazine paper which is derived from well-managed, certified forestry and chlorinefree manufacture. Future Publishing and its paper suppliers have been independently certified in accordance with the rules of the FSC (Forest Stewardship Council).

9000

9012

Linux Format UK

Recommend Documents

This is a test.

Written By:

{{author.name}}