05-10-07

Revision 2 as of 2008-08-06 16:24:20

Clear message

Okt 05 10:00:09 dholbach good morning everybody and welcome to another MOTU Q&A session!

Okt 05 10:00:26 dholbach I think we'll start again with a round of introductions

Okt 05 10:01:10 dholbach I'm Daniel Holbach, been MOTU for quite a long time already, am trying to make becoming a MOTU easier and to make Ubuntu Universe rock even harder

Okt 05 10:01:56 norsetto I'm Cesare Tirabassi, fresh MOTU

Okt 05 10:02:15 dholbach ... and great contributor Smile :-)

Okt 05 10:02:21 lool I'm Loïc Minier, would like to become MOTU

Okt 05 10:02:30 norsetto and living proof that anybody can become a MOTU Smile :-)

Okt 05 10:02:36 dholbach norsetto: hahaha Smile :)

Okt 05 10:03:01 dholbach lool: great to have you around

Okt 05 10:03:12 dholbach who else do we have?

Okt 05 10:03:14 * \sh is Stephan Hermann, former MOTU, contributor and lurker for the first time Wink ;)

Okt 05 10:03:35 slangasek I'm Steve Langasek, MOTU wannabe

Okt 05 10:03:52 dholbach heya slangasek, hey \sh

Okt 05 10:03:57 dholbach anybody else? Smile :)

Okt 05 10:04:40 dholbach ok... seems we have only people around who are quite familiar with packaging already

Okt 05 10:04:53 dholbach do you have any questions yourself?

Okt 05 10:05:07 dholbach maybe about processes or how the MOTU team deals with things?

Okt 05 10:05:20 lool Yes; I'd like to know how build failures on buildd should be handled and why I receive such notices

Okt 05 10:06:08 lool For example I received two build failures for synaptic; it seems one arch was set to Needs-Build again (hppa IIRC)

Okt 05 10:06:23 dholbach good question

Okt 05 10:07:15 dholbach Soyuz (the buildd part of Launchpad) sometimes sends out build error mails when build-deps can not be properly installed, in a lot of cases this is not your fault at all

Okt 05 10:07:58 dholbach that's something the Soyuz folks should fix, it's a bit annoying to have to check the buildlog before just deleting the error mail

Okt 05 10:08:05 lool Am I supposed to do something with the notice?

Okt 05 10:08:18 slangasek roll it up and hit lamont over the head with it Wink ;)

Okt 05 10:08:21 dholbach If you figured out that it's not your fault, no Smile :)

Okt 05 10:08:28 lool Hehe

Okt 05 10:08:38 lool I wonder whether other bug reporters get similar spam and why?

Okt 05 10:09:00 norsetto lool: we all get ......

Okt 05 10:09:02 lool (I didn't prepare the synaptic upload, I only reported a bug)

Okt 05 10:09:22 dholbach lool: oh?

Okt 05 10:09:36 lool Ah wait I did

Okt 05 10:09:49 dholbach lool: maybe that's part of the ChangelogClosesUbuntuBug thing?

Okt 05 10:09:55 lool mvo didn't replace my name when he added his changes

Okt 05 10:09:57 dholbach oh good *phew* Smile :-)

Okt 05 10:10:02 dholbach right

Okt 05 10:10:10 lool Ok, I'm relieved as well

Okt 05 10:10:36 dholbach in any case it's a good idea to talk to the folks on #launchpad if anything like that creeps up and is not explicable

Okt 05 10:11:28 dholbach any other questions?

Okt 05 10:11:59 dholbach do we have somebody here who's new to packaging and would like to benefit from the packaging experience we have gathered in #ubuntu-classroom right now? Smile :)

Okt 05 10:12:13 dholbach a package review? build problem? something?

Okt 05 10:12:45 lool I have another question about attaching debdiffs to bug reports; is it ok to use "UNRELEASED" in the target dist or should I use gutsy?

Okt 05 10:13:16 dholbach lool: I personally would use 'gutsy' and also (LP: #123456)

Okt 05 10:13:26 dholbach that way it's just a matter of applying it, trying and uploading it

Okt 05 10:13:45 dholbach I personally find UNRELEASED only useful for VCS commits

Okt 05 10:14:00 dholbach lool: is that common for sponsored uploads in Debian?

Okt 05 10:14:06 lool What do you mean by "and also (LP: #123456)"? You mean instead of "; LP: #123456." which I used?

Okt 05 10:14:19 dholbach oh, no - that's not what I referred to

Okt 05 10:14:28 dholbach "LP: #123456." should be fine too

Okt 05 10:14:40 slangasek dholbach: not IME; I agree with you that UNRELEASED is only useful when you're sharing around something in an intermediate state

Okt 05 10:14:45 lool dholbach: It's just that I saw a sponsor requesting an updated debdiff with s/UNRELEASED/gutsy and wondered whether this was a strict requirement

Okt 05 10:14:55 dholbach I just meant that if you make use of ChangelogClosesUbuntuBug and specify the upload target it's less work for the sponsor

Okt 05 10:15:13 lool Ok; naurally I close bugs in changelogs

Okt 05 10:15:15 dholbach lool: it's not strict afaik

Okt 05 10:15:37 norsetto I would say its good practice though

Okt 05 10:15:49 dholbach that's great - it's mainly new contributors who forget to use it, but you all know how to use it Smile :-)

Okt 05 10:16:23 dholbach hey adam_b - how are you doing?

Okt 05 10:16:30 dholbach adam_b: here for the MOTU Q&A session?

Okt 05 10:16:34 lool Ok; I have a question about bug states when reporting: I think I now have the bug triaging capability, but I'm new to it and I'm sometimes in doubt with bug states: should I set a bug to "confirmed" after I report it?

Okt 05 10:16:47 adam_b dholbach: I'm doing ok, yep thought I might listen in

Okt 05 10:16:53 norsetto lool: no, we should never confirm our own bugs

Okt 05 10:16:54 lool (In order to avoid bug triage)

Okt 05 10:17:02 dholbach adam_b: rock on

Okt 05 10:17:39 dholbach adam_b: if you have any questions regarding packaging, becoming a MOTU or something else - just ask Smile :)

Okt 05 10:18:06 dholbach lool: in some cases it's OK to do that - I think pitti's requestsync script for example sets to 'confirmed' automatically

Okt 05 10:18:45 dholbach but the archive admins will recognize your name on the bug report anyway and probably know that you know what you're doing

Okt 05 10:19:43 dholbach any other questions?

Okt 05 10:19:45 lool Hmm bash tells me requestsync is in devscripts, but I have devscripts and no requestsync script

Okt 05 10:20:08 dholbach lool: thanks for that - I'll tell mvo to update command-not-found-data

Okt 05 10:20:13 dholbach lool: it moved to ubuntu-dev-tools

Okt 05 10:20:22 lool Ok, thanks

Okt 05 10:20:23 adam_b dholbach: there's a package from debian 'mapnik' that is half in ubuntu, just wondered why that might be? there is .gz's but no .debs

Okt 05 10:20:59 dholbach adam_b: ok, let's take a look at http://launchpad.net/ubuntu/+source/mapnik together

Okt 05 10:21:27 dholbach it seems the source for mapnik 0.4.0-2 arrived in Ubuntu alright

Okt 05 10:21:39 adam_b cool

Okt 05 10:21:44 dholbach if you look at https://launchpad.net/ubuntu/gutsy/+source/mapnik/0.4.0-2

Okt 05 10:21:53 dholbach you will find that it failed to build on all architectures

Okt 05 10:21:59 dholbach let's take a look at the build logs

Okt 05 10:22:14 adam_b failed to build

Okt 05 10:22:33 dholbach python scons/scons.py PROJ_INCLUDES=/usr/include PGSQL_INCLUDES=/usr/include/postgresql PROJ_LIBS=/usr/lib DESTDIR=/build/buildd/mapnik-0.4.0/debian/tmp PREFIX=/usr BIDI=yes PYTHON=/usr/bin/python2.5 --clean

Okt 05 10:22:33 dholbach scons: Reading SConscript files ...

Okt 05 10:22:33 dholbach Checking for main() in C library m... no

Okt 05 10:22:33 dholbach Could not find header or shared library for m, exiting!

Okt 05 10:22:33 dholbach make: *** [clean] Error 1

Okt 05 10:22:57 dholbach a scons build issue, it seems

Okt 05 10:23:13 adam_b its cpp app and build thinks its c?

Okt 05 10:23:27 adam_b oh no next line

Okt 05 10:24:15 dholbach it has problems in Debian as well, it seems

Okt 05 10:25:56 lool dholbach_: Missed some lines?

Okt 05 10:26:03 adam_b when I build it in a gutsy chroot I have to build with sudo python scons/scons GDAL_INCLUDES=/usr/include/ install

Okt 05 10:26:20 dholbach_ adam_b: you shouldn't have to use sudo - try fakeroot

Okt 05 10:26:31 adam_b ok

Okt 05 10:26:50 adam_b even for the install part?

Okt 05 10:27:27 slangasek only for the install part

Okt 05 10:27:29 dholbach slangasek: I saw this http://bjorn.haxx.se/debian/testing.pl?package=mapnik

Okt 05 10:27:50 slangasek for the non-install part, you should deliberately avoid using fakeroot or sudo, to make sure the package builds cleanly as a normal user

Okt 05 10:28:18 slangasek dholbach: well, "waiting for boost" isn't a bug in mapnik

Okt 05 10:28:42 lool Here it cleans fine; perhaps it's a missing build-dep

Okt 05 10:28:58 slangasek (hmm, I immediately want to retract that comment after saying it, clearly depending on boost /is/ a bug Wink ;)

Okt 05 10:32:26 adam_b so in ububut build its looking for 'm' which is math.h (in SConstruct)

Okt 05 10:34:53 dholbach I'm afraid I can't offer an answer to this one

Okt 05 10:35:18 adam_b fakeroot gives: scons: *** [/usr/local/lib/libmapnik.so] /usr/local/lib/libmapnik.so: Permission denied

Okt 05 10:36:14 dholbach upstream have received a similar report: http://trac.mapnik.org/ticket/14 - but it's not of much help

Okt 05 10:36:31 slangasek adam_b: then you're missing the DESTDIR=$(CURDIR)/debian/tmp from the original

Okt 05 10:40:07 dholbach I think it might make sense to raise awareness of this failing to build on ubuntu-devel-discuss@lists.u.c

Okt 05 10:40:17 dholbach not sure we'll find an answer right here, right now

Okt 05 10:40:40 dholbach do we have any other questions?

Okt 05 10:42:09 dholbach adam_b: did you start to actively dive into MOTU things already?

Okt 05 10:44:45 adam_b dholbach: nope, just contemplating writing an app that will use mapnik

Okt 05 10:45:03 dholbach adam_b: ahh nice

Okt 05 10:45:06 dholbach hey huats

Okt 05 10:45:15 huats hey

Okt 05 10:45:25 huats dholbach: what a surprise to see you here

Okt 05 10:45:28 dholbach welcome huats, another soon-to-be-MOTU Wink ;-)

Okt 05 10:45:44 huats dholbach: Wink ;-)

Okt 05 10:45:45 dholbach huats: do you have any questions? anything we could take a look at together?

Okt 05 10:45:55 huats dholbach: yes I have

Okt 05 10:46:07 dholbach rock on

Okt 05 10:46:09 dholbach Smile :-)

Okt 05 10:46:09 huats I am searching it Smile :-)

Okt 05 10:46:18 dholbach I knew we could count on huats Smile :-)

Okt 05 10:46:38 huats alwasy count on me where there are questions to ask

Okt 05 10:46:42 dholbach hehe

Okt 05 10:46:45 * dholbach hugs huats

Okt 05 10:46:59 * huats hugs dholbach, his Master

Okt 05 10:47:17 huats it is related to bug #137513

Okt 05 10:47:30 dholbach http://launchpad.net/bugs/137513

Okt 05 10:47:47 lool (Just for the record: I don't get the failure on mapnik in pbuilder)

Okt 05 10:48:18 huats the last comment from pkern bothers me a bit...

Okt 05 10:48:22 dholbach lool: me neither

Okt 05 10:48:49 dholbach lool: we had weird scons problems in sbuild (at least I think so) before

Okt 05 10:49:41 dholbach huats: it's best to ask asac about that

Okt 05 10:50:06 huats dholbach: ok

Okt 05 10:50:09 dholbach huats: I don't know about a policy decision on iceweasel in package names

Okt 05 10:50:20 huats it is the only one apparently...

Okt 05 10:50:39 dholbach right

Okt 05 10:50:49 dholbach in debian there are lots of iceweasel-* binary packages

Okt 05 10:50:56 dholbach in ubuntu it would be just this one

Okt 05 10:51:07 * huats deserves a price : /me has asked a question about something that daniel doesn't know !!!

Okt 05 10:51:13 dholbach so best to check with asac and ask if it should be renamed to mozilla-torbutton or firefox-torbutton

Okt 05 10:51:22 huats dholbach: ok

Okt 05 10:51:27 huats dholbach: I will

Okt 05 10:51:35 dholbach super thanks

Okt 05 10:52:16 dholbach anything else? anything more we can review? Smile :)

Okt 05 10:52:35 huats I have a second question to fire : can you explain me a bit the process to transmit bug fixes to debian... I know norsetto tried to many times, but I am sure it is someting interesting for many of us

Okt 05 10:53:20 dholbach soren wrote submittodebian (in ubuntu-dev-tools) maybe that'd help with it?

Okt 05 10:53:31 huats oh

Okt 05 10:53:33 huats great

Okt 05 10:53:43 huats I'll have a look

Okt 05 10:54:26 norsetto dholbach: I think huats has some problems about what should be sent to Debian, in particular re. debdiff

Okt 05 10:54:50 huats norsetto reads my mind

Okt 05 10:55:05 dholbach right... I think the diff the debian maintainer could apply to the current source in sid would be most useful

Okt 05 10:55:19 dholbach but we have two experienced Debian folks here, ... Smile :-)

Okt 05 10:55:43 norsetto three ....

Okt 05 10:57:13 dholbach slangasek, lool: huats just asked " can you explain me a bit the process to transmit bug fixes to debian... I know norsetto tried to many times, but I am sure it is someting interesting for many of us" - do you have an answer for that?

Okt 05 10:58:21 huats dholbach: if you want we can have a look to a precise bug entry to see what is sendable to debian...

Okt 05 10:58:23 dholbach submittodebian is particularly useful for merges it seems

Okt 05 10:58:26 lool The best case would be to confirm the bug is still present in Debian by actually running a Debian and checking the bug and patch, then using reportbug to report the bug with the patch; that's probably a long process for simple bugs though

Okt 05 10:58:58 dholbach submittodebian has the advantage that you can edit the diff

Okt 05 10:59:04 dholbach and remove stuff that is ubuntu specific

Okt 05 10:59:06 lool So I guess one simpler case is simply checking whether the packages were patched in Debian or not

Okt 05 10:59:15 huats ok

Okt 05 10:59:23 slangasek dholbach: I was refraining from comment because my views on how Ubuntu should send bugfixes back to Debian seem to be at odds with the prevailing opinion within Ubuntu Smile :)

Okt 05 11:00:00 slangasek but as a Debian maintainer, if someone in Ubuntu has made a change that they think fixes a bug, I would like to hear about it directly through the Debian BTS

Okt 05 11:00:01 dholbach slangasek: right, but given the question from somebody who wants to submit a patch to the debian bts? Smile :)

Okt 05 11:01:35 slangasek btw, the actual submittodebian script seems to be missing from current ubuntu-dev-tools, only the manpage is there

Okt 05 11:01:42 dholbach oh?

Okt 05 11:02:08 huats dholbach: so I'll use the debian BTS ...

Okt 05 11:02:09 dholbach slangasek: I'll fix that in a sec - there's another change to it pending

Okt 05 11:02:17 dholbach (need to talk to Søren about that)

Okt 05 11:03:31 dholbach ok, asked him

Okt 05 11:03:46 dholbach if there are no more questions, this is the end of this MOTU Q&A session

Okt 05 11:04:12 dholbach we'll have one next week - I'll just ask for a proposal of a different time before Smile :-)

Okt 05 11:04:24 dholbach thanks everybody for your questions and answers

Okt 05 11:04:27 dholbach have a nice day

Okt 05 11:04:28 huats dholbach: thanks to you

Okt 05 11:04:30 lool Thanks!

Okt 05 11:05:21 slangasek cheers

Go back to MOTU/Q&A/Logs.