Tag Archives: Linux

Correcting obsolete dependency requirements in DEB packages

I use the excellent XMind mind-mapping tool. Moving from Linux Mint 13 (Ubuntu 12.04 LTS) to Xubuntu 14.04 LTS revealed a problem with the way XMind was packaged for Ubuntu. It had a hard-coded dependency on specific versions of the Java runtime – either of openjdk-6-jre, oracle-java6-jre, sun-java6-jre or sun-java5-jre. That wouldn’t do. The 14.04 LTS uses openjdk-7.

Of  course, the DEB package is built to support the widest range of Java runtimes and on some systems, they can go back up to Java 5 – when Sun Microsystems was still alive. But that should not be a reason to force up-to-date systems to pull in older runtimes. We can let XMind know about the issue and wait for them to update the dependencies and rebuild the package. Or we can do it ourselves.

DEB files are compressed archives with some control information. The dpkg-deb utility makes it very easy to modify DEB files.

$ mkdir -p extract/DEBIAN
$ mkdir build
# dpkg-deb -x package.deb extract/
# dpkg-deb -e package.deb extract/DEBIAN
[...do something, e.g. edit the control file...]
# dpkg-deb -b extract build

The control file was edited as follows:

-Depends: libgtk2.0-0 (>= 2.8.0), openjdk-6-jre | oracle-java6-jre | sun-java6-jre | sun-java5-jre
+Depends: libgtk2.0-0 (>= 2.8.0), default-jre | openjdk-6-jre | oracle-java6-jre | sun-java6-jre | sun-java5-jre

Now, the package will first check for the default Java package on the system and then move on to its list of specific runtimes.

Winner! In The Green Corner – Cinnamon!!

This is my new desktop. I am confident of keeping it till the next the LTS two years away.


Long time GNOME users were a worried lot because of GNOME 3 with its GNOME Shell. Ubuntu was not helping matters with its Unity desktop either. Thankfully, Linux Mint, led by Clem, stepped up to the challenge. They provided not one, but two alternatives to those fleeing from GNOME Shell and Unity – MATE, for those wanting to stick with GNOME 2.3x and Cinnamon, for those wanting to try GNOME 3 with a classical desktop layout.

I was willing to move to GNOME 3, but abandoning existing desktop metaphors and work-flows sounded too much of a hassle. Sue me! I was also suspicious of my three-year old laptop’s ability to run these modern desktops.

I decided to use MATE at first.  However I ran into a few issues.

  • There was noticeable flickering in the installer slide-show. This did not inspire confidence in the compositor or whatever component it is that takes the output from all desktop windows and renders it on your desktop.
  • The installer crashed once. It has been a long time since a Linux installer crashed on me!
  • MATE developers renamed various apps to maintain compatibility with GNOME 3. I found my myself double-guessing and hunting for apps. Though I agree – this wouldn’t be an issue after a few days of usage.
  • Some functionality was missing. Some applets and utilities were missing. Expected and forgiven, given that it is a new project with less developer resources than GNOME.
  • I missed the Scale and Expo Compiz plug-ins. Don’t ask me which does what. One exposes all the windows on a particular workspace and the other exposes all the workspaces.

In general, the environment felt neither solid nor familiar. If I was going to poke around and keep crossing my fingers, I ‘d rather use Cinnamon! At least I would be happy using a new, shiny piece of software. And boy, did it work out! I am very happy with Cinnamon. None of the four issues I mentioned above appeared in Cinnamon.

It’s not all rainbows and unicorns though. Cinnamon has some issues to clear out before it can dethrone GNOME 2.3x. Underneath Cinnamon is not GNOME 2.3x but GNOME 3.x. No matter how much Cinnamon is tuned and tweaked, it will never be GNOME 2.3x. Users will find that this is for the better.

Cinnamon is so good now, that after a week of usage, I actually have no complaints. All the issues I have are subjective.

  • Cinnamon is not feature heavy. Its developers should resist the temptation and user pressure to create a resource hogging Compiz clone. We don’t need rotating cubes, wobbly windows or fire trails after the mouse pointer.
  • Cinnamon’s implementation of Scale and Expo absolutely ROCK!
  • I actually prefer the Cinnamon Menu to the earlier Mint Menu.
  • I counted three tool-tip styles – Cinnamon, GNOME 3 and GTK+ theme. There are that many systems working together (whew!) to give you this awesome desktop. Come on guys, Linux Mint is now at the stage where you should have a OCD about such things. Get all of ’em styles linked and synced!
  • Some notification area icons (e.g. for Transmission and GNote) take up almost double the usual space. This doesn’t happen with others like Pidgin or Firetray for Thunderbird. So, is there a problem here or not?
  • The panels disappear when the gksudo prompt appears. It is disconcerting – the first time it appeared, I though Cinnamon had crashed.
  • I could not take screen-shots using the Print Screen button when a menu was in use in Pidgin. I don’t recollect this being a problem in GNOME 2.3x.
  • The reason I wanted to take a screen-shot was because the menu was disappearing below the Cinnamon panels – rendering the top-most and lower-most items difficult to reach.
  • It would be nice to have a placeholder applet on the panel to mark hot-spot corners, for users using the classic GNOME layout.
  • Applets need a ‘gravity’ sensor. For instance, check the cinnamon settings applet in the screen-shot. The icon (^) is fine if placed on the lower panel. It should have rotated 180 degrees when placed on the upper panel.
  • Cinnamon’s most exciting feature – applets and extensions – could be its breakout feature. It already has some sort of quality control and user feedback to up-vote good extensions and applets. There is also a PPA for Cinnamon extensions, applets and themes – users can easily and regularly upgrade Cinnamon components.
  • I tried the Coverflow AltTab extension and a timer applet, but had to remove them when Cinnamon started crashing just as I was getting down to work for the first time on the new setup. There are no issues running Cinnamon as it came out of the box.  May be these new applets and extensions need some more testing. May be Cinnamon needs better crash handling – something like Firefox’s Crash Reporter perhaps?

Despite these issues, Linux Mint with Cinnamon is a solid bet and I can’t wait to see how it matures in the future.

On the OS front, I have been more forgiving towards the default apps that come with the base install. About the only change I did was substituting Totem and GNOME MPlayer with SMPlayer. Everything else is as it came out of the box.


Enhanced by Zemanta

New host!

If you are a discerning reader of my posts, you would have noticed a very “disturbing” change on my blog. I have put a link in my footer that I didn’t need to, and which almost smacks of affiliate marketing. Worry not, my gentle reader. It is just that I am that excited about the move to Leanservers.

One reason is that I was becoming wary of the unlimited resources that A2 hosting was promising to everybody. Who are they fooling?

Most importantly, I wanted to move to a host that provided Nginx (say it with me – Engine X; N – gin – X). There is a small trickle of knowledge that has burst above ground recently – unless you are a big IT setup running enterprise applications on big iron IBM servers, Nginx serves your needs quite well.

Apache‘s HTTPD is a great server in its own right. I cut my teeth on it when learning web development. But over the years, Apache has become for Linux what IIS has become for Windows. Web servers like Nginx and Lighttpd are providing a welcome challenge to this unintended lock-in. The tag-along benefits are also welcome – better performance on virtual machines, easier configuration, cleaner config files and faster releases.

However, Nginx does not offer per directory configuration like Apache. Quite a few apps need Apache’s .htaccess file to reach 100% functionality. Some are implemented, inadvertently, to run only on Apache. I suppose this happens because the developer was trying to offer an ability to simply extract and run – think of WordPress’s famous five-minute install. Nginx’s increasing popularity is now changing the way web developers write their apps.

That being said, there are very few hosting providers for Nginx. This is, in part, due to its configuration limitations and maybe because most Nginx users are good enough at their job to put it on a VPS. I did not want the hassle of a VPS for my personal blogs – that would be repeating earlier mistakes.

Leanservers came to the rescue! No promise of unlimited resources (just like a proper VPS provider) and reasonable rates. I was a sceptic initially, but the guys at Leanservers (a huge shout out to Mr Unai Rodriguez) patiently answered my queries – yes, I can use Nginx the way I wanted to on a shared host. There were a few hiccups, but I have moved both blogs to Leanservers.

I am getting good scores on Y Slow and Pagespeed. The site is responsive. I unleashed Xenu’s Link Sleuth on Walk Alone and it performed at least 50% better than when A2 was hosting it. Of course, this is the honeymoon and things can go haywire later on. But I am hopeful. And did I say, excited?


Enhanced by Zemanta

battle report

What can I say? I’ve been allowed to continue as HoTT – TT is Tech Team, go figure out the rest. I suppose I must bow down and say “Thanks. Thanks.. O Ye wizards from the comp/it/etrx streams for letting a student from Electrical Engineering become HoTT”. In the bargain, nearly half the team is made up of IT wizards because that stream sacrificed a HoTT slot. Phew! No wonder we stay tuned during elections.

The comp/it/etrx (cie for short) people believe the TT belongs to them. Why? Because they use Flash. And maybe some programming and scripting. They use lpt interface programing. They handle lots of PCs and peripherals. And that’s all.

To tell you what I feel – THIS SUCKS. Period. In these times, to use those points as a criteria for exclusive ownership over the TT is anachronistic. These people have an over-inflated opinion of themselves. They forget that they got equal competitors churned out by the thousands from various B.Sc. IT courses, NIIT, Aptech and lots of similar institutes. Frankly, a formal degree in cie has lost its relevance in these times. Rank amatuers can beat these comp/etrx guys at their jobs – such are these fields. As for IT, please show me where these people work. They are simply taking away programming and system admin jobs. I, an amatuer who looks on Linux and related stuff as a hobby, could study 4 books on say, Linux Administration, Apache, MySQL and Perl – and beat a BE Comps/IT student hands-down at server design and admin anytime. (ok don’t sweat – I don’t have the time or inclination). These people can’t appreciate the fact that someone outside their direct discipline can do things as good as them or (again, don’t sweat) better.

Visit Joe’s blog on his site. He has a post on the irrelevance of Comp Science Engineering. Read it.

Enhanced by Zemanta

point of conflict

How does one persuade a guy to use Linux? When he is using w32 since he’s using a PC. The only thing he has used or is interested in are w32 apps, frequently MS products. By nature of the OS being proprietary, quite a few apps are proprietary. I leave the rest to your interpretation. This guy treats his PC as appliance – like a toaster, or maybe a glorified DVD player + calculator + TV + game console. His immediate concern is getting the toast – or his job on the PC done. And the only way he can/wants to do it is the way he was taught or like he learnt it. He recently migrated to Win XP. And now, is coming to terms with multi-tiered user levels, advanced config options which were never seen in Win 98/Me and much more.

I feel sad. I was like him once. Not having a choice. Not being aware of a choice being present. Not knowing that things could be different, and better, than they were now.

He is sorting out issues with his newly installed w32 build 5.x. He scoffs when I tell him that wizards are/should be the last option. He believes that having less wizards indicates a high-minded attitude among the developers of a system – that it is intended to drive away newbies. And then he says that LUGs are symptomatic of some shortcomings in Linux – that “linuxers are handicapped”. He refuses to believe that something which survived through so much (so much that I can’t fit it all here – visit Eric’s and Joe’s sites for more info and links) won’t last into the future. Though he is gracious enough to admit that the same fate may fall on w32. I believe that Linux will survive. Perhaps by another name. Maybe BSD will become dominant. Solaris. Maybe even Mac OS or the proper, authentic Unix itself will become open and dominant. Am I being over-confident? Will the principles of OSI and FSF fall away?

I don’t think so. I’ll do my best to keep them up – my bit.

You might want to read more of this GNU/Linux v/s Windows slugfest at Sandesh’s blog.

Enhanced by Zemanta