Search Results: "philipp"

4 September 2006

Branden Robinson: On Leader Reports

Philipp Kern: to put it simply, you're right. I appreciate your frankness, and yours is exactly the kind of feedback I need to do a better job. The primary reason for transitioning DPL reportage from monthly reports to blog entries is because I have found that most things in Debian that require my attention do not unfold neatly on a month-to-month basis. The ongoing attention to stable security updates, for example, began in June, carried forward through DebConf, and brings me to Oldenburg this weekend. Activity tends to be "bursty" instead of steady, which makes sense given that most of us don't have salaried jobs we can dedicate to Debian (I don't, certainly — my day job just has some overlap with my avocation). The emphasis I placed on visibility, demystification of Debian processes, and reform of the ones that don't work well remain my focus. To that end, it seems logical to me that it would only compound a failure to issue monthly reports by withholding information until I do issue one again. I'm trying to keep people informed, and if the information I've been sharing via my blog entries about DPL matters is of little value, then I need to know that — please tell me. planet.debian.org, by virtue of its .debian.org address, is an official project resource, which is why I think it's a suitable forum to use for communication with Debian's developers and users. Similarly, I am acting in my elected position when I post to any other Debian mailing list, even though subscription to other lists by developers is not required, and I act in that position when I speak with the press. I expect to be held to account for my statements in those forums, and I suspect you'd agree that I couldn't avoid responsibility for them by claiming that I didn't make them to debian-devel-announce. Whether I have good reasons for falling behind on the monthly reports or not, and whether or not I still think now they are the best way to maintain visibility is beside the point. I fully expect, if I run for election against next year, for my rivals to make much of the failure. My primary concern, however, remains serving the membership of our organization, and the users of our distribution, as best I can. To that end, I have a question for you and others: should I re-post blog entries about leadership matters to debian-devel-announce, or is that needlessly duplicative? Along the same lines, is it desirable that I stop (or reduce) blogging in favor of emails to debian-devel-announce? Let me know what you think; post to the Planet, and/or mail me at leader@debian.org.

30 August 2006

Philipp Kern: Re: SCR 3310 on Linux

So finally I got my card reader running on Linux. It was not that hard, but the main problem was caused by the proprietary Linux driver I installed. Thanks to Ludovic Rousseau's Generic CCID driver the reader is able to run just fine with only Free Software. It works even with libchipcard2, but a tiny fix to udev was necessary to fix a permission problem (you might need to adapt the vendor and product id to your local card reader, as shown by lsusb):
/etc/udev/rules.d/smart_card.rules:
BUS=="usb", SYSFS idVendor =="04e6", SYSFS idProduct =="5116", GROUP="chipcard", MODE="0664"
With this tiny fix chipcardd2 correctly opens the card as a non-privileged user. My thanks also go to Thomas Viehmann, who provides an HBCI branch of Gnucash on Alioth. Apart from the application crashing a bit too often, the HBCI part works fine with my card and PIN. The only problem left is the sudden disconnect of the reader when it's not used for a few minutes. Only a restart of chipcardd2 and gnucash resolves the problem temporarily. All my homebanking activities are now properly signed, and the retrieval of transactions works just fine. To credit my bank a bit, I am a customer of 1822direkt, a subsidiary of the Frankfurter Sparkasse, since a few weeks. They provided me with a free checking account, together with a free(!) HBCI card. Transactions are mailed just-in-time (checked for hourly during working hours) PGP-encrypted. (Yes, they are running PGP5i on OS/2. I needed to create a RSA key using PHP5i/Unix without a passphrase. The public key output of PGP5i had to be sent to the bank, and after an import of the resulting public and private key into my GPG instance I was able to read the statements correctly. The mail interface did not cope with GPG-generated keys.)

28 August 2006

Philipp Kern: Lazyweb: SCR 3310 on Linux?

Dear Lazyweb, is there any way I could get a SCM SCR 3310 running on Linux with libchipcard2? pcscd seems to start up correctly, but I lack applications to check if chipcard accesses come through properly. But I don't get any successful connection to the chipcard reader with neither the free CCID driver nor the proprietary SCM one.

20 August 2006

Philipp Kern: Gobby 0.4.0rc4 released

We just released Gobby 0.4.0rc4 (together with its library dependencies net6 1.3.0rc3 and obby 0.4.0rc4) with the hope that it is one of the last release candidates before the final 0.4 release. Our application tests all went very well, without any further desyncronisation. Please test this release, as we need to weed out any further bugs. From our point of view those packages should be usable for productive working, but your mileage may vary, of course. I will push this branch into Debian unstable as soon as the final release is available, to enter Etch in time. Packages are already in Debian experimental (which, of course, might require that unstable is also available on your system). For i386 only, though. I lack build machines for other architectures, and although experimental is autobuilt nowadays, builds would need to be requeued manually due to the bumped library dependencies. Backports to Sarge are not yet available, do not hesitate to request them if needed, it will be easier to backport them than it is for the 0.3 branch. Sobby is missing, too, as it needs some updates to be made in the course of this week. If you use Ubuntu Dapper, just follow these instructions. Changes from the last set of release candidates:
For more changes between the 0.3.x and 0.4.x branch please read the NEWS file.

18 August 2006

Philipp Kern: The future of (G)obby

Armin provides a nice outlook into the future of (G)obby, a collaborative editor and the corresponding library to embed this feature into other applications. A C rewrite of the obby library is currently in progress, but it's still quite some time away. We will release the 0.4.0 branch soon, though. There will be another set of release candidates for all components in the next few days, which hopefully pave the way to a final release of that series. It was delayed for a long time due to a desyncronisation issue, which seems to be fixed since 0.4.0rc3. The 0.4.0 branch of Gobby will have the following advantages over the 0.3.0 one (which is in Debian unstable and Ubuntu dapper):

16 August 2006

Philipp Kern: Drag'n'drop for Ruby-GTK2

Uh, yeah. So I "lost" a whole day due to two "signals" missing in the ruby-gtk2 wrapper. Both GtkTreeDragSource's "row-draggable" and GtkTreeDragDest's "row-drop-possible" are implemented as virtual functions, instead of plain signals. This means that a new gobject class needs to be created to properly use those callbacks. Well, Armin Burgmeier and I had a nice long Gobby session (completely over IPv6 by the way, welcome to the bleeding-edge of technology, where no port forwards are required) about this problem. Him being the main developer of Gobby, he knows much about Gtk, so finally we came up with this patch, which adds both functions as signals to the Ruby-Gtk API. It's sad that my program will require a patch to the GUI wrappers in order to run correctly, but this is, as far as I see, the only way to get proper drag'n'drop in Ruby.

11 August 2006

Philipp Kern: Broken reportbug

People are going crazy when reportbug is broken, not able to report new bugs. Well, now I know how we could release etch in time... Here is the fix:
--- reportbug_ui_text.py.broken 2006-08-11 14:43:01.000000000 +0200
+++ reportbug_ui_text.py        2006-08-11 14:43:28.000000000 +0200
@@ -223,6 +223,7 @@
             break
         list.append(entry)
     ewrite('\n')
+    return list
 def get_password(prompt=None):
     return getpass.getpass(prompt)
Pipe it into cd /usr/share/reportbug; patch. Well, I hate my blog, I does not display heredocs correctly, otherwise that would have been a copy'n'paste receipt. Though I don't know if the patch fixes all the bugs introduced.

5 August 2006

Zak B. Elep: Ok, I need to post now

Ah, well. I’m such an errant blogger, being focused in doing actual work. But here are some updates:

Alexis Sukrieh: Backup Manager 0.7.4-1 uploaded to sid

I finnaly took the time to package 0.7.4 for Debian, thanks to Philipp Kern, it has just been uploaded. This package closes about ten bugs and provide the last available “development” release of Backup Manager. I hope this package will please those who will “apt-get” into it. I now plan to fix as much upstream bugs as I can, and moreover, work on the inclusion of rdiff-backup.

24 July 2006

Philipp Kern: XFS on root on powerpc

Thank you very much, Aur lien G R ME. So yaboot on XFS is finally fixed. (Which means that root on XFS is finally possible again on powerpc.) Ah and this is also fixed in Edgy's yaboot 1.3.13a-1ubuntu1.

Ross Burton: Sound Juicer "I Need A Little Love To Ease The Pain" 2.14.5

Sound Juicer "I Need A Little Love To Ease The Pain" 2.14.5 is out. Tarballs are available on burtonini.com, or from the GNOME FTP servers. This releases rolls in the bug fixes from 2.15.4:

Ross Burton: Sound Juicer "Two Undernourished Egos Four Rotating Hips" 2.15.4

Sound Juicer "Two Undernourished Egos Four Rotating Hips" 2.15.4 is out. Tarballs are available on burtonini.com, or from the GNOME FTP servers. Not a lot going on, sorry.

19 July 2006

Philipp Kern: Melting

Tomorrow we will encounter 38 C over here. I will be melting. Now. Really.

3 July 2006

Philipp Kern: Graduation

So after a week of post-graduation stress I am back at the computer working for Debian and for my Summer of Code participation. Two weeks ago I had to prepare myself for two oral examinations, which took place on Monday and Tuesday last week. On Tuesday I received my transcript, and all people celebrated this hilariously. The parties continued until Friday when we earned the final certificate. I completed my Abitur with a grade of 1.2 (with 1.0 being the best grade). I think that I am satisfied with this result. Next in line is the move to Karlsruhe to enroll for the CS department at the University of Karlsruhe.

28 June 2006

Philipp Kern: Blog folds

MJ Ray, couldn't you just export the complete entries via your RSS feed? (This also applies to other users of folding.)

21 June 2006

Philipp Kern: Debian NEW queue

Package count in NEW: 222
(NEW queue summary)
Well, that is a huge queue again. Could the ftp-masters please fast-track packages which fix RC bugs? (Like being stuck in NEW for a new SONAME because of an incompatible ABI change which caused dependent packages to FTBFS, like it's the case with gmp.) P.S.: And no, this is indeed intended as a general call, not to fast-track gmp, which is in NEW since only 2 days.

17 June 2006

Philipp Kern: Some more progress with reportbug

It is a daily fight with the oddities of PyGTK. It seems that there are severe issues in conjunction with the way I use the threading in Python. The patch is almost complete in terms of functionality. (Not everything is committed yet, because I will review all core changes again, to keep them both compact and sensible.) But there are issues with the GUI, like it not reacting after a program switch in conjunction with the GTK accessibility module. Scrolling the log window to the bottom does not work neither, even if I call the corresponding function scroll_to_mark in an idle handler (and thus in the main thread). Still those are minor, mainly cosmetic issues. It's just that the deadline is approaching. (Considering that I face the final examinations for my High School degree in a bit more than a week, I need to hurry up. I get my topics on Monday, so there is one week of preparation time for both a presentation and all content of one class.)

14 June 2006

Philipp Kern: Blog Comments

The captcha protector provided by Serendipity once worked correctly. But due to changes in one of the latest versions of the spam blocking plugin the captcha images went unreadable. Thus legitimate users were turned down by it. The protection was automatically activated seven days after the post was published, but I deactivated it now, after two complaints sent via mail. The comments could, however, still be held for moderation. So, thank you for your input! If I know of a problem I try to solve it as soon as possible.

9 June 2006

Philipp Kern: First AM report finished

Quite some time has passed since I accepted to process Alexis Sukrieh's NM application (back in August 2005) but today I fired his AM report out to the Front Desk, where it now waits for a review if any things are missing. I hope that I did not leave out anything crucial, due to him being my very first applicant to process as an application manager. Cheers.

7 June 2006

Philipp Kern: Embedding Lua into qmail

So I spent most of yesterday's time on embedding a Ruby interpreter into qmail, and most of today's time on embedding a Lua interpreter into it instead. The latter worked well enough, the former did not and the resulting scripting environment was just weird. Not even the File object did work, I was not able to put some strings into a debugging file. Why I do that? Well, I needed greylisting, which seems to require a bit more policy than expected. There was a patch for qmail-smtpd which called a shell script to check if a certain file is present, representing the greylisting triplet. I first changed the script to sanitise its input variables and it looked to work well. But no, it broke sender verify. Another machine of mine, running exim, was not able to do a proper callout to this server, out of dubious reasons I have not yet fully explored. I wanted more control over the policies when to accept mail (considering that qmail does not implement anything like this by itself), so embedding looked like a good choice. I still hope that Lua provides me enough functions through its libraries, just like Ruby would. And why I use qmail... The commercial web hosting panel I got for free with the server, called Plesk, uses it, so I will stick to it until I find a drop-in replacement. I cannot be bothered to either write a free replacement for the panel which takes some work off me, or to take the challenge and replace qmail on the live server, having no other server equipped with the panel to try it on. Especially considering that I do this hosting for free in my spare time...

Next.

Previous.