Michal Čihař - Archive for Jan. 1, 2009

This evening (CEST), the first code from this year GSoC has been merged to phpMyAdmin trunk. And I think it is really interesting piece of code.

Derek Schaefer has spent his time to improve import abilities of phpMyAdmin. You can now import arbitrary spreadsheet documents and you should be able to get reasonably looking table with all of the data.

And as a side effect, we got better export to various Excel formats (including XLSX).

Thanks Derek, your contributions were really great and it was a pleasure to work with you (especially when compared to responsiveness of other students).

Gammu 1.25.91 is out! The changelog is short, but all changes are quite interesting - thanks to Iain Hibbert we now have support for Bluetooth on other BSD variants and I spent time travelling around Spain in trains by writing documentation and translating Gammu to Czech language.

Full list of changes:

• Documented most commands in man page.
• Completed Czech translation of Gammu.
• Avoid crash on some Nokia smses (bug #959).
• Added support for Bluetooth on more BSD systems, thanks to Iain Hibbert.
• Added new command setsmsc.
• Added option to disable checking for battery and signal in SMSD.

You can download from usual place: http://cihar.com/gammu/,

Debian users will find packages in experimental soon, SUSE, Fedora and Mandriva users can check experimental repository in OpenSuse Build Service.

Recently, SourceForge.net has updated their release system. The change does not seem to be that bad (besides the fact that people got used to the old system and need to learn something new), but the change completely broke their RSS exports for file releases.

But hey, you have freedom of choice! You can either use old RSS feeds, which are not being updated or new RSS feeds, which duplicate same data for all entries. Sorry, these links require login.

At the end I've found a way - you can download whole project RSS, where you ignore most content, but get just the file names and then download separately RSS for each file. And now you have complete data!

Yesterday I arrived back from Spain and DebConf, so I should be finally back on track to process mails, bug reports and other similar things. If you're waiting for my reply more than month, you should be soon handled ;-).

I really enjoyed DebConf and meeting all the people there. And as well I did enjoy rest of travel through Spain. Except the day when they stole my backpack in Barcelona. But I can hardly do anything about that, it just made my vacation quite expensive because loosing my DSLR camera with lenses.

It has been really nice week in here, but it's time to leave. Last night, wake up in the morning and leave by train (in my case to explore Barcelona).

Thanks to all people who organised DebConf 9, from my POV everything went really smoothly and I enjoyed time here. I finally met people I meet daily on the mailing lists and now I can finally imagine people behind those email addresses.

Now it's time to pack my things and finally have some sleep, otherwise I will not be able to wake up tomorrow and I will miss my hardly booked train.

While being on DebConf 9 I managed to do some work on Gammu, the famous tool for getting access to data in cell phones. The greatest feature of this release is reducing memory consumption by 800 MiB. Other code changes are much smaller, but there is also another important change - I now use Git for development, you can git clone git://gitorious.org/gammu/mainline.git.

Full list of changes:

• Fixed getting firmware revision on AT phones (bug #951).
• Added support for home/work mobile number (bug #952).
• Fixed handling of error response for security status in AT.
• Fixed segfault on systems with very low memory limits.
• Development now goes on in Git repository, please check README.
• 6510 driver now dynamically allocates memory for file caches, what reduces Gammu memory usage by 800 MiB.
• Added testcases for AT+CNMI replies.
• Fixed range parsing in AT engine.
• Added reconnect support in PostgreSQL driver in SMSD.
• Gammu debug level no longer can be controlled separately in SMSD.
• Fix loop when Nokia phone does not allow to access files (bug #954).

You can download from usual place: http://cihar.com/gammu/,

Debian users will find packages in experimental soon, SUSE, Fedora and Mandriva users can check experimental repository in OpenSuse Build Service.

After losing some data from svn, because I just did not have the latest backup and I only had data in local copy, I decided to finally migrate to Git.

For now the repositories are hosted on gitorious:

• Gammu
• Wammu
• Ukolovnik

The svn server will stay there for some time, but I want to get rid of it in future.

PS: Translations from Pootle are now directly commited to separate Git repository under user data which he used to register in Pootle. I'll merge this repository time to time to mainlines.

It looks like some things will be changing in Debian in next months. The most known thing which is already widespread know is switching to time based freeze cycles, what should make us synchronised with Ubuntu LTS release cycles. This gives advantage of sharing work done on security support, because both distribution will be quite similar code base.

The other thing which we saw today is new website proposal. From what I could see in the audience, it has been accepted quite positively, so maybe Debian will finally have nice looking website in (near?) future. The discussion showed up that also somebody else is working on that, so we will have even two competing proposals. I just hope it will not end up in some huge flamewar :-).

Well I really love these automatic test and "rewards". Today, Wammu has just received the "Windows 7 Compatible" award on ForteDownloads.com. Wow.

But as far as I know, it has some problems on Windows Vista with permissions and I doubt it will be better on Windows 7. Is anybody actually using these systems? I don't.

This is just a small rant about issue which did hit me today. One of MyISAM tables on my database server has been crashed. Well you can expect that kind of thing after a hardware failure. I did start REPAIR TABLE and nothing happened for quite a long time.

Finally I managed to look into the logs and found this jewel:

[ERROR] /usr/sbin/mysqld: Disk is full writing '/tmp/STwW9hIc' (Errcode: 28). Waiting for someone to free space... Retry in 60 secs

Well MySQL, nobody else is not using this vserver and the /tmp is a ram filesystem. So if you filled it up, nobody is going to make space for you there and you will wait for ages.

Maybe continuous waiting is not the right idea...