27 Aug 2005 shlomif   » (Master)

Python-IL Activity

The second meeting took place in July. We were five people there: Gabor Szabo, Beni Cherniavsky, Amit Aronovich, Chen Levy and I. We mostly talked about the lessons from the Israeli Perl activity, and about the possibility of having a conference for all the P-languages (Perl, Python, Ruby, etc.)

After the meeting, we were given a lecturing hall in Aduva. So the next meeting along with presentations was coordinated for 16 August. There were a few glitches along the way, but it was very successful: over 20 people came, and we had a lot of fun. The first presentation which was supposed to be about TestOOB (pronounced "tess-toob") actually covered testing in general (not even specifically to Python) most of the time, which was nice. The shorter presentations afterwards had more glitches because they involved computer demonstrations, and there were a few hardware problems.

The room was crowded and we had to insert a few more chairs. If fewer people show up to the next presentations than this one, we'll be OK, but otherwise, we may need to be looking for a larger place.


I did quite a lot of work on www.perl.org.il. I added more CPAN authors and modules to the Israeli Perl Projects page. I heavily refactored the templates to avoid duplicate code and markup. Recently, I also created a modified look and feel, which I will hope will enable a multi-level navigation menu that will eliminate the need for the Miscelleanous Content Page. Most of my patches were accepted there.

News from my other Blogs

A use.perl.org Journal Entry about my article about HTML::Widgets::NavMenu that was published in Perl.com. Another one about the new Acme::Gosub module which brings BASIC-like Retro-programming to Perl 5.

In my Hebrew Linmagazine blog: A few more anecdotes about the IGLU Upgrade and Summary of August Penguin 4 from my personal POV. The latter sparked an active discussion about Debian vs. the world. (and some other smaller discussions)

Computer Problem 1: Kopete and Jabber

I recently upgraded to Mandriva 2006 Beta2. I won't talk about the upgrade right here. I'll just mention that afterwards Kopete (the KDE Instant-Messaging client) had problems connecting to Jabber. When trying to connect using it, it said: "There was a connection error: Operation is not supported.". After compiling Kopete, making sure it will compile with "-g" and without "-O2", trying to debug with gdb, with many problems along the way, and not succeeding (it's a long and sad story), I ended up concluding that I should try to manually tweak the configuration file.

In the config file ($HOME/.kde/share/config/kopeterc) in the Account_JabberProtocol_* section, there were several keys starting with PluginData_JabberProtocol. I removed them (and most other keys) and then started Kopete again. Then Jabber did not work with a different error. After I went to the Jabber account configuration, and saved it, it worked, so it is possible that removing the extraneous keys was not necessary. It may be a backward-compatibility problem between KDE 3.3.x and KDE 3.4.x.

Problem solved, but futzing with gdb, etc. took several hours.

Computer Problem 2: cp -i -f

The -i flag causes UNIX commands like rm, cp or mv to be interactive and request confirmation before deleting or overriding files. -f is the opposite: force such deletions without requests. Many people alias these commands to rm -i, cp -i, etc. for safety.

When I was working on Eskimo I tried to copy a file using cp -f (with the alias cp="cp -i" in place.) That still asked for confirmation - apparently the -f flag was ignored. On my Mandriva system a similar invocation worked.

It turns out Mandriva applies a patch to GNU cp to allow the later "-f" flag to override the "-i". Debian did not apply a similar patch. Now the question is which policy is the philosophically correct one. Note that with the "rm -i -f" command, GNU rm happily applies the "-f" flag.

Bug in GNU grep

I came across a bug in GNU grep fixed it, and submitted a patch. It turns out the bug was already fixed in the CVS version, but without the test case that I proposed. So a similar test case was added. It's strange that a new version of grep was not released yet as this bug is a crash. In any case, delving into the code was fun.

Revision of the Beast

* Revision 666:

File svn-commit.tmp saved File svn-commit.tmp not changed so no update needed. Sending MANIFEST Transmitting file data . Committed revision 666. shlomi:~/progs/perl/www/Nav-Menu/trunk/module$

This was a commit to the MANIFEST of HTML::Widgets::NavMenu to add two test-related files that I had already added to the repository. (the MANIFEST lists all the files that are part of the Perl distribution). It was part of a larger amount of work on HTML::Widgets::NavMenu, SVN::RaWeb::Light and possibly other modules within the web-cpan repository. A short time and 34 revisions later I also hit on revision 700. Right now, I'm at revision 720.


For some reason, I believed that the Perl 5 IO::Scalar module is part of the core Perl distribution. However, it is part of the CPAN IO-stringy module. Maybe what misled me was the fact that it was installed on my Mandriva system, and the fact that it is such an elementary module. In any case, I had to add it as a dependency to SVN::RaWeb::Light and HTML::Widgets::NavMenu, and accidently reported it as uninstalled in PONIE (the ongoing Perl 5 implementation for PONIE), when I tried to see if HTML::Widgets::NavMenu passes all of its tests there. The good news is that it does after IO-stringy is installed.

GIMP Hacktivity

Looking for some GIMP Hacktivity to do, I looked into making sure the GIMP resources can be categorized (as was suggested by a mailing list post). Sven Neumann had suggested a pre-requisite and I wrote a patch as a result. More information can be found there.

Mozilla Bookmarks Patch

The Mozilla and Firefox browsers have an annoying behaviour with pasting bookmarks. I decided to write a patch to remedy it. This involved hacking the Javascript "chrome" code of the Mozilla internals. I neglected working on this for a long while, because Mozilla yelled at me from some reason I could not understand at an XPCOM statement. As it turned out, I was misled by the Javascript debugger, which caused a "Next" or "Step" statement into a certain XPCOM call, to run indefinetly. (guess I now have another bug to report). Having sorted this out, fixing the patch was relatively uneventful and the patch was ready.

I posted it at the SeaMonkey bug record, but then it turned out that SeaMonkey has a separate Bookmarks Editor, and my patch is Firefox specific. So I filed a separate entry for Firefox (after going over all the Firefox Bookmarks bugs and seeing nothing relevant there). Right now, my patch is waiting for a review, and has been for over a month. I hope it is reviewed soon, because this bug is very annoying. (and I have a fix, damn it!) This was my first Mozilla patch ever.

Latest blog entries     Older blog entries

New Advogato Features

New HTML Parser: The long-awaited libxml2 based HTML parser code is live. It needs further work but already handles most markup better than the original parser.

Keep up with the latest Advogato features by reading the Advogato status blog.

If you're a C programmer with some spare time, take a look at the mod_virgule project page and help us with one of the tasks on the ToDo list!