My last update about running activities is dated back to Paris Marathon (April 10th). Time for a new update, then. After a very busy end March-early April, with that crazy succession of 55km Ecotrail and 42km Paris marathon with only two weeks between both, I needed a kind of "break". Moreover, I had another goal set with the 35m trail race named "Trail des Cerfs", on May 15th, in my neighbourhood. So, end of April was a little bit quieter. Still, one week after Paris marathon, I completed a very hard trail run in Seine Valley: only 18km, but about 600-700m positive slope. That was indeed a good moutain-style run (yes, in Seine valley....) and a good way to check that I recovered quite well. Later on, I slowed down a little with even a Sunday (24th) where I *didn't run at all*..:-).. Just a few short uns during the week and a short Saturday run. Well, that was indeed Easter Sunday and Monday is a holiday in France, so I ran an half-marathon on 25th...:-)...on a loop where I already have some references, the fastest being 1h46. This time it was 1h52'...but, well, that was two weeks after a marathon... Still, a small pain at the back of my ankle decided me to stop running long distances until May 15th. So, I only did short trainings up to no more than 12km, even during my stay in Germany for SambaXP. As a consequence, I was very well for the 35km (600m positive slope) Trail des Cerfs, which I completed in 3h31', 168th out of 328 runners. More interestingly, I completed it in negative split with a much faster second half, ending up at about 12km/h. For seomthing that (given the slope and the nature of the paths) is comparable to a marathon, that's very promising. The rest of May was spent quietly with a standard training and progressively resuming longer distances. I originally planned to do my spring solo marathon on May 28th in Versailles Castle park, but I stopped at km28 and completed the distance two days after with a 14km run..:-)... Stopping was wiser as I was feeling that the last loop would have been a great pain. That was indeed the first time I give up on a marathon. I prefer this to happen for a non official race, after all! June will not see any "big goal". I might be running one race somewhere in my neighbourhood but the point will mostly be accumulating kilometers before summer. Summer (which includes DebConf that makes training a little bit more hectic) will not see any race, probably (unless I find one in banja Luka!) but more some specific preparation for Berlin marathon. Here, the "safe" target for me is, first, breaking my record. I'm probably not ready for 3h30 and even 3h35 might be difficult. It seems that I'm now reaching a barrier where progress becomes harder. Before Berlin, I'll run the Bois d'Arcy half-marathon, two weeks before. Last year, I managed to set a very good mark, there, with 1h37. Not sure that i'll try to break this only two weeks before Berlin. After Berlin? We'll see, but my last big goal of the year will be the 67km "Le Puy-Firminy" night race, in November (no web site, this is indeed a "small" local race). Yet another adventure beyond marathon distances, in the unknown...

This week saw again some progress for samba packages. May is indeed my samba month..:-). So, I uploaded samba 3.5.8~dfsg-3 and -4. For whatever reason, -3 was rejected three times and, as I already tagged it in our SVN then started adding changes for an upcoming -4, I finally uploaded that -4. It fixed about 10 more bugs and the bug count is now 62. Upstream also release 3.6.0-rc1, which I uploaded in experimental and which fixed an RC bug. I also began working on Pootle 2.1.6 as an attempt to resync with upstream. I integrated patches by Alaa Abd El Fattah to have it use dbconfig-common to setup a database. However, this revealed that more invasive work is needed to satisfy dependencies...that may depend on the database backend chosen by users. As I'm quite ignorant with these things, I might need some help, here. I uploaded a new release of xgridfit to fix an important bug and avoid an NMU. xgridfit is team-maintained in pkg-fonts. I'm not very active in this team ATM. I shoul dresume work for it as I proposed a talk about it at Debconf11..:-) I continued hard work on French DDTP. Martijn made changes to stats pages and it makes interesting challenges to complete. You know I like challenges.. Smith reviews continue to go on: I launched a call for translations for slapos.tool.format and started the review of ntfs-3g. Finally, on Saturday....I upgraded churro (aka i18n.debian.net) to squeeze. Really. All alone..:-)...on a server that lies in Spain. "Meme pas peur" (not even scared) as we say in French. I'm really highly confident in squeeze upgrades now....and I was right. After only 5 hours work (mostly looking at packages downloads and configuration), nearly everything was back to work. As of now, we got three glitches, with only one bignot yet fixed:

• the wiki that uses MoinMoin is severely broken. So, as of now, http://i18n.debian.net yields an error (but all other i18n.d.n URLs are working)
• HTTPS was broken for ddtp.debian.net, which prevented people to login to work on the DDTP. I partly fixed that today (Monday) and Felipe will have a closer look at VirtualHosts in Apache to fix the remaining glitches
• cron jobs for i18n robots were no longer working. After some headache and googling, it turned out one file in /etc/cron.d was group-writable, which cron translates to "WRONG INODE INFO".

Quite active week, finally. This week should be quieter as I'm on holidays with Elizabeth in our country house in Normandy, in Seine valley (this house belonged to Elizabeth parents and she inherited it). I'm in deep love with this house and there are tons of things to do there...which leaves few room for hacking. We'll see what will be next weekly summary!

This week, I continued triaging bugs in the samba source package. The result is interesting Many bugs are tagged "moreinfo", most of them being old bugs, many likely to be user error....or something fixed for quite a while. So, I expect the bug count to decrease again. Moreover, I did a call for translations 10 days ago, so about 10 bugs are indeed transaltion updates which will be closed in next upload. Samba upstream also released their 3.6.0rc1 version this week, and 53 hours after the annoucement, it landed in Debian experimental. I pushed tasksel 3.02 in git. Indeed, I upload it quite a while ago and forgot to push the release and tags. I'm not entirely used to git, as it seems..:-) On l10n front, I launched a Smith review for ntfs-3g and launched a call for translations matlab-support (and got thanks from maintainer). Finally, after last week push, the French team completed to translation of package descriptions for the top500 packages in popcon. As a consequence, I moved the target further and we nos have "Popcon600" on the stats page. French now reached rank 3 in the DDTP.

Old blog entries popped out from my laptop after a restore where I couldn't restore file dates. Sorry..:-)

This week, I continued triaging bugs in the samba source package. The result is interesting Many bugs are tagged "moreinfo", most of them being old bugs, many likely to be user error....or something fixed for quite a while. So, I expect the bug count to decrease again. Moreover, I did a call for translations 10 days ago, so about 10 bugs are indeed transaltion updates which will be closed in next upload. Samba upstream also released their 3.6.0rc1 version this week, and 53 hours after the annoucement, it landed in Debian experimental. I pushed tasksel 3.02 in git. Indeed, I upload it quite a while ago and forgot to push the release and tags. I'm not entirely used to git, as it seems..:-) On l10n front, I launched a Smith review for ntfs-3g and launched a call for translations matlab-support (and got thanks from maintainer). Finally, after last week push, the French team completed to translation of package descriptions for the top500 packages in popcon. As a consequence, I moved the target further and we nos have "Popcon600" on the stats page. French now reached rank 3 in the DDTP.

Old blog entries popped out from my laptop after a restore where I couldn't restore file dates. Sorry..:-)

After my rehearsal from yesterday, I finally booked my flights to Edinburgh for Debconf 7. I'll fly from CDG terminal F on June 15th 16:15 (AF5052) and will fly back to CDG on June 24th 09:05 (AF5051). The return flight is somewhat more expensive than later flights (about 30EUR more) but much more convenient for me. See you in EDI, ladies and gentlemen...

New pictures again (I leave tomorrow for holidays, so I hurry up). Again among my favourite ones, pictures of people just socializing. Amaya once wrote that Debconf is more about socializing than hacking and I couldn't agree more.

... I just learned how to use Mercurial after Daniel converted his darcs repository for aptitude to Mercurial. Being such a VCS wizard when you're told to be the DD with no programming skills is kinda scary...:-)

There are some bloggers who I'm really happy to see back in the blogosphere.

Geez, I got a Nokia M770 for X-Mas. /me does the geek dance.

That was a damn busy week. It was mostly centered about attending SambaXP, the annual Samba user and developers conference, in G ttingen, Germany. The only free software conference I attend with expenses paid by my employer, Onera. This year was the 10th edition and, as last year, to the "who was there for the nth edition" game, I won by staying alone as they asked who attended all editions of the conference. :-) That was a great week, with time spent with people as interesting as Andrew Tridgell, Jeremy Allison, John Terpstra, Volker Lendecke, Kai Blin, to name a few. A good opportunity, again, to get input from our packaging work for that big piece of software, as well as getting visibility about the future of Samba. I also had a great, even if short talk, with the kind Karolin Seeger, the release manager of Samba for 3 years now. We talked about....children, as she's now a mother since last year (with a non negligible impact on her professionnal life, as often in Germany). Great meeting, too, with Brad Kuhn, from the Software Freedom Conservacy, who had a keynote about GPL licence enforcment activities. It becomes more and more sure that Samba3 and Samba4 will reconverge together after the Samba Team releases Samba 3.6. It brings plans for our packaging work: I think we'll stick with having samba 3.6 in wheezy while the brand new shiny Samba4 probably stays separate in some way. Our users (and /me first) clearly need stability in the file and print services first. Of course, I did some packaging work there: samba 3.6.0pre3 was uploaded to experimental, about 10 days after its official announcement. I also worked on the samba *binary* package bugs, triaging them as usual. We now have 51 bugs opened against the samba binary package: 18 unclassified, 11 moreinfo (several likely to be closed as unreproducible or user error), 1 wontfix, 8 with a pending patch and 13 forwarded upstream. I'm also thinking about a possible way to ask about SMB2 support in samba: it won't be activated by default in 3.6 (mostly because us, distros, requested for that and, by "us", I mean Debian, RHEL, SuSe and their derivatives, so quite a large consensus). Still, it would be good to put some light on SMB2 support and a debconf question about it could be a solution (not shown by default and defaulting to no SMB2). I also worked quite extensively on packages maintained by No l K the, Ralf Treinen and me, aka "the pkg-running team". I did setup a git repo for my new "garmin-ant-downloader" package, that allows downloading track files from Garmin Forerunner 405 GPS watches (guess what is the brand and model of mine!). My first packaging git repository! Thanks to Ralf for his advice and help in this. I triaged bugs in the other two packages we maintain: pytrainer (more bugs forwarded upstream) and garmin-forerunner-tools (which was later uploaded by Ralf). I also setup a team mailign list so, now, we're a real team...:-) Few activity on the l10n front: a few Smith reviews are in progress and I completed 1 or 2 French translations, and reviewed some others. Regular activity, then. The only specific stuff is that I'm now pushing harder for the French DDTP effort, doing many reviews and translations there. We try to reach 100% in the "popcon500" packages. Later, we'll try to head at reaching the hieghts reached by the Italian and German teams, who are, on this l10n activity, way ahead from us. Finally, during the SambaXP conference, and as usual (except last year because of too heavy work duties), I visited my German friends, living "close to" G ttingen, accomppanied by Luk Claes and his friend and colleague Ivo, who were also attending SambaXP. Great barbecue at Andreas and Kathrin Tille's place, facing the Wernigerode castle at sunset. And the best sp tzle ever at Meike and Alex Reichle's place in Hildesheim, with a french touch on the salad's dressing as well as great Chilean wine brought by Meike's coworker Wolfram. Always a great time to see these good friends even if that means driving a few hours (and being flashed....twice!...by german speed cameras on the way to and back Andreas place!) To complete the week, I ran a 34km/800m+ trail today in the Rambouillet forest, completing it in 3h31. I'll probably blog separately about running updates as it is now quite some time that I didn't. Guess what? I'll be sleeping well tonight...

Yes, finally, I decided to upload my Debconf 10 pictures. Not to my own gallery, as usual, because it is apparently broken (viewing pictures works but upload plugins aren't). So, they are on Debconf gallery site. So, please enjoy a few more gazillion of weird Debconf pictures. Sorry to the Debconf 10 official report editors: I should have done this before so that you have pictures of Debconf runners.

As usual at the beginning of a month, I gave spam reviewers a GO to review spam on debian-boot as well as french-speaking mailing lists. I triaged bugs in the pytrainer package and forwarded some more with patches (thanks to Matt Kraai and Jukub Wilk). Some have already been fixed upstream. Apparently, pytrainer's upstream is very reactive and opened to such bug reports. I fixed several forgotten commits in D-I packages after the previous week's mass upload. Thanks to Coling Watson for prodding me about this. I launched a debconf templates review for a recently appeared package: slapos.tool.format. I indeed even filed a bug report for debconf abuse for it..:-) However, I dropped the templates review for apparmor as the package maintainer launched a translation update round on his own. I completed the review and translation update round for bastet. I sponsored fonts-hosny-amiri and added Ahmed El-Mahmoudy to DMUA for it. This was, Ahmed should be able to upload the package by himself in a near future. He's doing clean work with that package, so adding him to DMUA was not a big deal.

As usual at the beginning of a month, I gave spam reviewers a GO to review spam on debian-boot as well as french-speaking mailing lists. I triaged bugs in the pytrainer package and forwarded some more with patches (thanks to Matt Kraai and Jukub Wilk). Some have already been fixed upstream. Apparently, pytrainer's upstream is very reactive and opened to such bug reports. I fixed several forgotten commits in D-I packages after the previous week's mass upload. Thanks to Coling Watson for prodding me about this. I launched a debconf templates review for a recently appeared package: slapos.tool.format. I indeed even filed a bug report for debconf abuse for it..:-) However, I dropped the templates review for apparmor as the package maintainer launched a translation update round on his own. I completed the review and translation update round for bastet. I sponsored fonts-hosny-amiri and added Ahmed El-Mahmoudy to DMUA for it. This was, Ahmed should be able to upload the package by himself in a near future. He's doing clean work with that package, so adding him to DMUA was not a big deal.

...when waking up, learning, live, the death of Bin Laden, and then spend one hour wondering if I should be happy or sad that we end up being happy of this. Confused, still confused.

Will I succeed maintaining a weekly log of what I'm doing in Debian? We'll see. At least, let's try. D-I work: After spending most of the end of week 16 uploading D-I components, I completed this during this week. I hunted down the remaining arch:all, arch:i386 and arch:amd64 packages that need an upload because of l10n changes. The main point was bringing published D-I work inline with the work of translators. That was mostly meant to publicly have an image with Uyghur activated (which I blogged about separately). Finally, I ended up doing some changes to the fonts-ukij-uyghur package in order to use the right font to display the language in D-I, as well as improving the readability by increasing the font size just as we're already doing for Arabic and Persian. Samba packaging: I got release managers approval to upload a new release to stable-proposed-updates. It fixes several important bugs in samba 3.5.6, often affecting Win7 clients. I'm always happy to see that we can do that even when issues are ot security or "release critical" issues. I also integrated the newly-published samba 3.6.0-pre3 in the package maintenance SVN. However, we're still having problems with symbols disappearing from libwbclient, that prevent me to upload 3.6.0pre3 in experimental. Steve Langasek's help here will be welcome. Pytrainer and "gant" packaging: I setup a mailing list for the pkg-running maintenance team so that we can move the few packages some of us maintain (pytrainer and garmin-forerunner-tools as of now, and soon "garmin-ant-downloader", the helper tool for 405 watches). I also began looking at pytrainer bugs and forwarding some of them upstream. I also began packaging "gant" (had to rename it as we already have a gant package, totally unrelated). I just got a notice from Klaus Ethgen about improvements he did to the software. We'll try to figure out what to do as this software doesn't really have a clear upstream and published versions. l10n work: a few packages introduced debconf templates and I launched reviews of English for them (while mumbling about uncoordinated work, of course..;-)). I also reviewed about one hundred package description translations in French DDTP (something I recently began to commit to do even though it means onlinen work for me, which I'm not used to). No l10n NMUs this week but a few maintainer uploads after I (gently) kicked their ass..:-)

I really don't know how I ended up spending this entire evening listening to Bhutanese music. But I did well. I particularly enjoyed Rang Sem Kar Ga Mikar and Gayser Jang Lu Jonpi Gang, which I both found on YouTube. The former reminded be the pleasure and (apparent) simplicity of Bhutanese dances, which I had the privilege to share with guests in Thimphu a few years ago. The latter gives a great idea of the peacefulness of that place of the world (I recommend scenery views some of which probably taken at Dochula Pass). And that language is so beautiful even when one doesn't understand a word. Maybe some readers of this blog will become listeners of Bhutanese music for some time after reading this. I just can wish this. And, doh, I really can't wait until I go back there. I will, one day or another. "Ku Zu Zang Po La" to those who made me discover that place.

Last week, I announced the activation of Uyghur language in Debian Installer. For once, I even went further: I rebuilt all D-I packages that include localization (at least for i386/amd64 architectures) and uploaded them. As a consequence, D-I daily builds now allow previewing how an install done in Uyghur can be like. For this, please go to the Debian Installer development page, download the netboot ISO image (not netinst) and launch D-I with "suite=unstable" passed as argument when booting the ISO image. You should end up with something like this, this or this. There are a few glitches, still: for instance, as the font we chose has no Bold variant, the screens "titles" do not use an Uyghur font but an Arabic or Persian one. As a consequence, glyphs ar enot properly joined together. I'm investigating this issue with Uyghur translators and we're trying to choose an appropriate font among the gazillion that have been created by UKIJ. Please notice that Uyghur looks like Arabic but, according to those who can read Arabic, is really different from Arabic (or Persian). I hope that this great work by Gheyret Tohti, Abduqadir Abliz and other people working on this will give ideas to other so-called "minorities", here or there in the world, who want to be able to install Debian in their language. There is still a lot of work to do, such as having end-user software localization for Uyghur and, then, a uyghur-desktop task, but this is at least the first language that gets added for wheezy. In case you would be wondering, the current count of activated languages in D-I is 71.

This week, I finally switched to use pytrainer to track my running data. Up to now, I was still using Garmin proprietary software on Elizabeth's Windows copmuter for this. Mostly because my GPW watch is a Forerunner 405, that uses a bloody crappy stupid USB stick with proprietary communication protocol (ANT+). So, the use of proprietary crappy software was kinda mandatory. There is however some code floating around to handle this device in Linux, named "Gant". This is far from being very polished, but I finally managed to make it work more or less reliably to get the GPS traces from my watch to my Debian laptop. So, during the week , I brought all my old running data back from "Garmin Training Center" (the Windows software provided by Garmin) back to my laptop and imported it in pytrainer. And I imported new data from my daily runs with help of that "gant" software. So, this is now time to really jump into maintenance of pytrainer and help my friend No l (that will be easy as pytrainer is kinda well maintained), then package "gant" for Debian, as part of the work in the "pkg-running" maintenance team (Ralf Treinen, No le K the and /me). New challenge for me: learn how to create a git public repo and maintain a package in it (I already participate to maintenance work with git for several things, but I never initiated any). Other plans: some bug triage on pytrainer (if some is needed, I indeed don't know yet) and eventually deal with things that have to be reported upstream. Looking at Launchpad might also be a good idea as I suspect that many are using pytrainer with Ubuntu.