Bug#372508: Preparing multiarch

Loïc Minier lool at dooz.org
Fri Aug 18 12:19:41 UTC 2006 

        Hi,

On Wed, Jun 14, 2006, Goswin von Brederlow wrote:
> But I think I have pango converted now. Just have to check package
> contents tomorrow so nothing gets lost and then the big task of making
> upgrades go smoothly. I think that will need some conffile moving or
> deleting for the generated files.

 We're in the process of replacing this part of the pango packaging
 because of the necessity to have a working pango between unpack and
 postinst.  THe Ubuntu packages moved the file to /var, I plan moving
 the file too, and this is a chance to include arch information.  If you
 have any suggestion / patch, please see:
 <https://launchpad.net/distros/ubuntu/+source/update-manager/+bug/41297>

 I also attach today's IRC log on the subject.

 The Ubuntu package also has some patches which seem closer to bi arch /
 ia32-libs stuff which I'm attaching for you.

 I would appreciate if you could process all the information and propose
 something, otherwise I might pick a layout incompatible with multiarch,
 and we would have to rediscuss this again in the future.


 You mentionned that Gtk differs from Pango in that it has an ABI
 version, I think pango has one too, so you should be able to share code
 between these.

   Bye,
-- 
Loïc Minier <lool at dooz.org>
-------------- next part --------------
14:21 -!- efaistos [~efaistos at m139.net81-64-30.noos.fr] has joined #gnome-debian
14:22 -!- efaistos [~efaistos at m139.net81-64-30.noos.fr] has left #gnome-debian []
14:28 -!- Q-FUNK [~q-funk at dsl-hkigw7-fe10f900-10.dhcp.inet.fi] has quit [Связь Ñ€Ð?сет бй пÐ?вÐ?]
14:37 -!- MrCooper [~daenzer at 195.226.9.186] has joined #gnome-debian
14:54 -!- decko [~decko at skywalker.ffclrp.usp.br] has joined #gnome-debian
15:06 -!- Q-FUNK [~q-funk at dsl-hkigw7-fe10f900-10.dhcp.inet.fi] has joined #gnome-debian
15:13 -!- seb128 [~seb128 at ANancy-151-1-18-128.w83-194.abo.wanadoo.fr] has quit [Ex-Chat]
15:21 -!- mvo_ [~egon at p54A66DC0.dip.t-dialin.net] has joined #gnome-debian
15:31 -!- mvo [~egon at p54A65EA7.dip.t-dialin.net] has quit [Ping timeout: 600 seconds]
15:34 -!- Q-FUNK [~q-funk at dsl-hkigw7-fe10f900-10.dhcp.inet.fi] has quit [Ping timeout: 600 seconds]
15:37 -!- stratus [~stratus at cronopio.rits.org.br] has joined #gnome-debian
15:39 -!- MrCooper [~daenzer at 195.226.9.186] has quit [Ping timeout: 600 seconds]
15:42 -!- ross [~ross at althur.gotadsl.co.uk] has joined #gnome-debian
15:42 -!- telemaco [~telemaco at 84.77.31.94] has joined #gnome-debian
15:51 -!- sadleder [~sadleder at p50812A0D.dip0.t-ipconnect.de] has joined #gnome-debian
15:51 -!- sadleder [~sadleder at p50812A0D.dip0.t-ipconnect.de] has left #gnome-debian []
15:57 -!- seb128 [~seb128 at ANancy-151-1-18-128.w83-194.abo.wanadoo.fr] has joined #gnome-debian
15:57 -!- mode/#gnome-debian [+o seb128] by JHM
16:05 <@lool> seb128: FYI, I've clarified the symbol issue wiht mcclasen and he said this is normal and due to the API changes in the fiel system backend; probably only affects gtk and libgnomeui as well as third party modules
16:15 -!- atrus [atrus at S01060000b48e3b44.cg.shawcable.net] has quit [Ping timeout: 600 seconds]
16:18 <@seb128> lool: ok, good, what I though but better to check :)
16:19 <@seb128> is there any issue with @debian.org emails atm? I got a surprisingly low number of emails today
16:25 -!- atrus [atrus at S01060000b48e3b44.cg.shawcable.net] has joined #gnome-debian
16:25 < adn> lool: is kov usually here?
16:26 <@seb128> adn: sometimes, why?
16:27 < adn> seb128: to thank him for update-notifier :)
16:27 <@seb128> ah, k
16:27 <@seb128> you can thank mvo_ who code it if you want :p
16:27 <@lool> adn: he hangs less on the IRC channel, but you can mail him ;)
16:28 <@lool> adn: the work of kov was to take the software from Ubuntu, and patch it for Debian
16:28 <@lool> adn: for example change the name of the distro in strings, or adapt the CD detection code to recognize Debian CDs (not only Ubuntu ones)
16:29 <@lool> adn: thanks for supporting his efforts BTW
16:33 -!- ross_ [~ross at althur.gotadsl.co.uk] has joined #gnome-debian
16:35 -!- ross [~ross at althur.gotadsl.co.uk] has quit [Ping timeout: 600 seconds]
16:41 < CIA-1> lool desktop/unstable * r7032 gnomemeeting/debian/ (changelog control control.in patches/recvers.dpatch): 
16:41 < CIA-1> * Fix FTBFS with bumping pwlib b-d to new version along with update of
16:41 < CIA-1>  recvers.dpatch (Closes: #377357)
16:41 < CIA-1> * Swap Maintainer and Uploader with Jose
16:42 < adn> lool: did you see my query?
16:42 < adn> mvo_: great work!
16:42 -!- ross_ [~ross at althur.gotadsl.co.uk] has quit [Ex-Chat]
16:42 < adn> seb128: heh
16:43 < mvo_> adn: woah, thanks :)
16:43 <@seb128> oh, mvo_ is around
16:47 < adn> mvo_: you're anonymously quoted in next DWN
16:47 < adn> I did not know your name
16:51 < mvo_> adn: nice! what is my quote :) ?
16:54 <@lool> adn: yes I did back then, i thought it wasn't necessary to explicitely acknowledge it, but I will in the future
17:09 -!- sadleder [~sadleder at p50812A0D.dip0.t-ipconnect.de] has joined #gnome-debian
17:09 -!- sadleder [~sadleder at p50812A0D.dip0.t-ipconnect.de] has left #gnome-debian []
17:11 < CIA-1> lool desktop/unstable * r7033 gnomemeeting/debian/ (changelog control control.in): 
17:11 < CIA-1> * Build-depend on libebook1.2-dev (>= 1.1.4) instead of
17:11 < CIA-1>  evolution-data-server-dev which now striclty depend on what it requires,
17:11 < CIA-1>  thanks Andreas Jochens. (Closes: #383132)
17:28 -!- MrCooper [~daenzer at 84-74-141-85.dclient.hispeed.ch] has joined #gnome-debian
17:37 -!- seb128 [~seb128 at ANancy-151-1-18-128.w83-194.abo.wanadoo.fr] has quit [Ex-Chat]
18:04 < adn> lool: great
18:04 < adn> mvo_: you'll see next week :)
18:05 < mvo_> adn: can't you /msg it to me? I'm *so* curious ;)
18:05 < adn> heh
18:12 -!- telemaco [~telemaco at 84.77.31.94] has quit [Download Gaim: http://gaim.sourceforge.net/]
18:49 < CIA-1> lool packages/unstable * r7034 gtranslator/debian/ (changelog control control.in): * Bump up Standards-Version to 3.7.2.
18:53 -!- gpfreitas [~guilherme at 201.37.234.246] has joined #gnome-debian
18:56 -!- herzi [~herzi at kiwi.mediascape.de] has quit [Remote closed the connection]
18:57 -!- mvo_ [~egon at p54A66DC0.dip.t-dialin.net] has quit [Ex-Chat]
18:58 -!- herzi [~herzi at kiwi.mediascape.de] has joined #gnome-debian
18:59 < CIA-1> lool packages/unstable * r7035 gtranslator/debian/ (changelog compat control control.in): * Bump up Debhelper compatibility level to 5.
19:04 < CIA-1> lool packages/unstable * r7036 gtranslator/debian/ (changelog patches/02_hebrew-translation-list.patch): 
19:04 < CIA-1> * New patch, 02_hebrew-translation-list, to set the Hebrew translation list
19:04 < CIA-1>  as suggested by Lior Kaplan. (Closes: #382629)
19:04 < CIA-1> lool packages/unstable * r7037 balsa/debian/ (changelog control control.in): 
19:04 < CIA-1> * Rebuild against libgmime-2.0-2-dev which supersedes libgmime2.1-dev.
19:04 < CIA-1>  (Closes: #383157)
19:05 <@jordim> non-stop lool
19:09 <@jordim> 19:07 < aj>      gnome | 1:2.14.2.1 |       testing | all
19:09 <@jordim> 19:07 < aj> pretty
19:09 <@jordim> good
19:10 <@lool> thanks for forwarding the good news, I was overly depressed by the stupidity of my work on gtk lately :)
19:15 < spoof> good work :)
19:16 <@lool> well, it's really ready since 3 weeks
19:17 < spoof> gtk 2.10?
19:17 <@lool> gnome 2.14
19:17 <@lool> for etch
19:17 <@lool> it didn't migrate because of deskbar-applet because of xulrunner
19:18 < spoof> gnome 2.16 will not be in etch? :/
19:18 <@lool> depends
19:26 -!- gpfreitas [~guilherme at 201.37.234.246] has quit [Ex-Chat]
19:38 -!- xjuan [~xjuan at host166.200-117-224.telecom.net.ar] has joined #gnome-debian
19:42 -!- telemaco [~telemaco at 84.77.31.94] has joined #gnome-debian
19:50 < CIA-1> lool desktop/unstable * r7038 pango1.0/debian/ (changelog patches/10_allow-empty-gpos-table.patch): 
19:50 < CIA-1> * New patch, 10_allow-empty-gpos-table, to allow an empty GPOS table, should
19:50 < CIA-1>  prevent warnings reported in Debian #377944, thanks John Wright; included
19:50 < CIA-1>  in upstream CVS and in >= 1.13.4 releases.
19:55 < CIA-1> lool desktop/unstable * r7039 pango1.0/debian/ (changelog po/ru.po): 
19:55 < CIA-1> * Add Russian Debconf templates translation, thanks Yuri Kozlov.
19:55 < CIA-1>  (Closes: #380131)
19:56 < CIA-1> lool desktop/unstable * r7040 pango1.0/debian/ (changelog update-pango-modules.in): 
19:56 < CIA-1> * Sort modules in update-pango-modules(.in), as requested by Daniel Glassey.
19:56 < CIA-1>  (Closes: #341463)
20:04 -!- telemaco [~telemaco at 84.77.31.94] has quit [Download Gaim: http://gaim.sourceforge.net/]
20:09 < CIA-1> lool packages/unstable * r7041 gtranslator/debian/changelog: * Rebuild against unstable's glib, ups.
20:10 < spoof> burn it! :)
20:13 < CIA-1> lool desktop/experimental * r7042 pango1.0/debian/ (6 files in 3 dirs): (log message trimmed)
20:13 < CIA-1> * Merge 1.12.3-2.
20:13 < CIA-1> * debian/watch:
20:13 < CIA-1>  - Update watch file
20:13 < CIA-1> * debian/copyright:
20:13 < CIA-1>  - Update source location
20:13 < CIA-1> * New patch, 10_allow-empty-gpos-table, to allow an empty GPOS table, should
20:17 -!- decko [~decko at skywalker.ffclrp.usp.br] has quit [Fui embora]
20:19 -!- josesanch [~jose at 227.Red-213-96-192.staticIP.rima-tde.net] has joined #gnome-debian
20:19 -!- mode/#gnome-debian [+o josesanch] by JHM
20:32 < CIA-1> lool desktop/experimental * r7043 pango1.0/debian/ (19 files in 4 dirs): 
20:32 < CIA-1> * New upstream releases, with API additions.
20:32 < CIA-1>  - Target at experimental.
20:32 < CIA-1>  - Drop patch 10_allow-empty-gpos-table, merged upstream.
20:32 < CIA-1>  - shlibs bumped (follow upstream version).
20:32 < CIA-1>  - Run debconf-updatepo.
20:35 < CIA-1> lool desktop/experimental * r7044 pango1.0/debian/ (changelog control control.in): * Bump up Standards-Version to 3.7.2.
20:36 < CIA-1> lool desktop/experimental * r7045 gtk+2.0/debian/ (4 files in 2 dirs): 
20:36 < CIA-1> - New patch, 011_gdk-directfb-cvs-changes, backport of CVS only build
20:36 < CIA-1>  fixes to permit compilation against directfb 0.9.25.1.
20:40 -!- gpfreitas [~guilherme at 201.37.234.246] has joined #gnome-debian
20:42 -!- gpfreitas [~guilherme at 201.37.234.246] has quit [Ex-Chat]
20:52 -!- zorglub [~zorglub at wahe.diwi.org] has quit [leaving]
20:53 -!- zorglub [~zorglub at wahe.diwi.org] has joined #gnome-debian
21:05 -!- josesanch [~jose at 227.Red-213-96-192.staticIP.rima-tde.net] has quit [Ex-Chat]
21:46 -!- damog [damog at merkel.debian.org] has joined #gnome-debian
21:54 -!- kristo [~kristo at ip-127-234.sn1.eutelia.it] has quit [Ping timeout: 600 seconds]
22:38 -!- slomo_ [~slomo at dslb-084-061-144-093.pools.arcor-ip.net] has quit [Ping timeout: 600 seconds]
22:41 -!- cwryu [~cwryu at 211.219.73.44] has quit [전 �만 갑니�.]
22:46 -!- slomo_ [~slomo at dslb-084-061-137-202.pools.arcor-ip.net] has joined #gnome-debian
22:57 -!- Q-FUNK [~q-funk at dsl-hkigw7-fe10f900-10.dhcp.inet.fi] has joined #gnome-debian
23:25 -!- mvo [~egon at p54A66DC0.dip.t-dialin.net] has joined #gnome-debian
23:32 -!- Q-FUNK [~q-funk at dsl-hkigw7-fe10f900-10.dhcp.inet.fi] has quit [Remote closed the connection]
23:50 -!- kristo [~kristo at ip-127-234.sn1.eutelia.it] has joined #gnome-debian
00:04 -!- herzi [~herzi at kiwi.mediascape.de] has quit [Leaving.]
00:11 -!- stratus [~stratus at cronopio.rits.org.br] has quit [Ex-Chat]
00:18 -!- Q-FUNK [~q-funk at dsl-hkigw7-fe10f900-10.dhcp.inet.fi] has joined #gnome-debian
00:19 < Q-FUNK> hm
00:19 < Q-FUNK> what could possibly cause a .desktop fie to not who in the menu, other than the hiden option?
00:19 < Q-FUNK> argh.. not show
00:24 < cortana> does it show when you run "gnome-menu-spec-test -in" ?
00:26 < Q-FUNK> nope
00:27 < cortana> then i guess the answer is, 'something else' :)
00:28 < Q-FUNK> indeed
00:28 < Q-FUNK> I wonder what
00:33 < Q-FUNK> it's wifi-radar
00:33 < Q-FUNK> currently only in ubuntu, it seems
00:39 < Q-FUNK> cortana: any way to force the panel to efresh it menu?
00:39 < cortana> kill it :)
00:41 -!- Q-FUNK [~q-funk at dsl-hkigw7-fe10f900-10.dhcp.inet.fi] has quit [Remote closed the connection]
00:47 -!- Q-FUNK [~q-funk at dsl-hkigw7-fe10f900-10.dhcp.inet.fi] has joined #gnome-debian
00:48 < Q-FUNK> found
00:48 < Q-FUNK> needed an exact path to the binsry, since it's not in the non-root users' path
00:49 -!- Q-FUNK [~q-funk at dsl-hkigw7-fe10f900-10.dhcp.inet.fi] has quit [Связь Ñ€Ð?сет бй пÐ?вÐ?]
00:55 -!- LizardKing [~LizardKin at h22.2.141.67.ip.alltel.net] has joined #gnome-debian
00:58 -!- seb128 [~seb128 at ANancy-151-1-48-92.w83-196.abo.wanadoo.fr] has joined #gnome-debian
00:58 -!- mode/#gnome-debian [+o seb128] by JHM
01:02 -!- LizardKing [~LizardKin at h22.2.141.67.ip.alltel.net] has quit [Leaving]
01:22 -!- xjuan is now known as xjuan|afk
01:26 <@jordim> seb128?
01:26 <@seb128> jordim: pong
01:26 <@jordim> seb128: so I have this pbook, with its clever 1 button touchpad
01:26 <@jordim> the button acts as X's button 1
01:27 <@jordim> the other day I plugged in a USB mouse for the first time I guess
01:27 <@jordim> and I found out that my config was for right-handeds, as I only have 1 button
01:27 <@jordim> so I changed it to left handed
01:27 <@jordim> and the touchpad started doing button 2
01:29 <@seb128> nice ;)
01:29 <@jordim> seb128: GTK bug?!!
01:30 <@seb128> gnome-settings-daemon one probably
01:33 <@jordim> good night folks
01:33 <@seb128> 'night jordim
02:08 -!- mvo [~egon at p54A66DC0.dip.t-dialin.net] has quit [Ex-Chat]
02:10 -!- seb128 [~seb128 at ANancy-151-1-48-92.w83-196.abo.wanadoo.fr] has quit [Ex-Chat]
02:19 -!- xjuan|afk [~xjuan at host166.200-117-224.telecom.net.ar] has quit [Leaving]
02:23 -!- damog [damog at merkel.debian.org] has quit [Leaving]
02:54 -!- kov [~kov at 201.67.64.102] has joined #gnome-debian
03:11 -!- MrCooper [~daenzer at 84-74-141-85.dclient.hispeed.ch] has quit [Ping timeout: 600 seconds]
03:39 <@lool> hey kov!
05:11 < kov> hey lool 
05:11 < kov> how's it?
05:21 < kov> oh well, bed time
05:21 < kov> see you
05:21 -!- kov [~kov at 201.67.64.102] has quit [Fui embora]
06:21 -!- cortana [~sam at 82-32-57-57.cable.ubr08.azte.blueyonder.co.uk] has quit [Ping timeout: 600 seconds]
06:42 -!- azeem [~mbanck at host109.natpool.mwn.de] has quit [Read error: 78 (Connection timed out)]
06:58 -!- cortana [~sam at 82-32-57-57.cable.ubr08.azte.blueyonder.co.uk] has joined #gnome-debian
08:13 < spoof> morning
08:16 -!- heikkih_ [~heikkih at 121.80-202-218.nextgentel.com] has joined #gnome-debian
08:19 -!- heikkih [~heikkih at 121.80-202-218.nextgentel.com] has quit [Ping timeout: 600 seconds]
08:40 -!- cwryu [~cwryu at 211.219.73.44] has joined #gnome-debian
09:16 -!- mvo [~egon at p54A66DC0.dip.t-dialin.net] has joined #gnome-debian
09:21 -!- MrCooper [~daenzer at 84-74-131-40.dclient.hispeed.ch] has joined #gnome-debian
10:04 -!- efaistos [~efaistos at m59.net81-66-147.noos.fr] has joined #gnome-debian
10:15 -!- cwryu [~cwryu at 211.219.73.44] has quit [전 �만 갑니�.]
10:25 -!- seb128 [~seb128 at ANancy-151-1-48-92.w83-196.abo.wanadoo.fr] has joined #gnome-debian
10:25 -!- mode/#gnome-debian [+o seb128] by JHM
11:12 -!- Q-FUNK [~q-funk at dsl-hkigw7-fe10f900-10.dhcp.inet.fi] has joined #gnome-debian
11:21 -!- seb128 [~seb128 at ANancy-151-1-48-92.w83-196.abo.wanadoo.fr] has quit [Ex-Chat]
11:22 <@lool> pff
11:22 <@lool> almost grabbed seb
11:22 -!- Q-FUNK [~q-funk at dsl-hkigw7-fe10f900-10.dhcp.inet.fi] has quit [Read error: 104 (Connection reset by peer)]
11:24 <@sjoerd> heh
11:24 <@lool> sjoerd: big hal update; many ubuntu patches too
11:24 <@sjoerd> yup
11:24 <@lool> sjoerd: did you find some ubuntu developper to work with?
11:25 <@sjoerd> i've been always working with martin (pitti) for gvm/hal
11:25 <@lool> hah, he's great at sharin
11:25 <@lool> g
11:25 <@sjoerd> but i mostly synced from their bzr repo 
11:25 <@lool> he explained me how good it's going in cupsys too
11:26 <@lool> sjoerd: you might want to offer him to share the SVN repo; perhaps it's a bit late if he's using bzr already
11:26 <@lool> sjoerd: I know he sometimes proposes kmuto to merge patches in the Debian SVN when they would be useful for Debian
11:26 <@sjoerd> i already did (years?) ago
11:26 <@sjoerd> but he didn't want it at that time
11:27 <@lool> hmm, nm then :-(
11:27 <@sjoerd> it's not a big problem as long as we both sync regularely
11:27 <@sjoerd> i know he always nicely syncs with the debian package, it's mostly missing time on my side
11:29 -!- seb128 [~seb128 at ANancy-151-1-48-92.w83-196.abo.wanadoo.fr] has joined #gnome-debian
11:29 -!- mode/#gnome-debian [+o seb128] by JHM
11:29 <@lool> ah seb128, do you think we could spend some time merging pango1.0 today?
11:30 <@lool> seb128: I see you moved the list of pango modules to another location for example and would like to sort that out before uploading
11:31 <@seb128> lool: ah, right, I planned to mail debian-gtk-gnome about that change and ETOOBUSY
11:32 <@lool> seb128: I intend to push forward the transition of the core libs (glib, gtk, pango spring to mind) as soon as they seem to be in a better shape than unstable's
11:33 <@lool> seb128: but I would like to benefit from the work ubuntu did in experiencing all transitions
11:33 -!- nanda [~nanda at 84-75-31-99.dclient.hispeed.ch] has quit [Ex-Chat]
11:33 <@seb128> lool: have you read http://mail.gnome.org/archives/gtk-devel-list/2006-August/msg00068.html ?
11:33 -!- Q-FUNK [~q-funk at dsl-hkigw7-fe10f900-10.dhcp.inet.fi] has joined #gnome-debian
11:33 <@lool> seb128: BTW, the reason I mail you in english is to be able to forward mail exchanges, as you did for directfb for me
11:33 <@seb128> ah, k
11:34 <@seb128> I'll reply in english next time then ;)
11:35 <@lool> I don't read gtk-devel-list, so I'm reading the thread now
11:40 <@lool> seb128: it's interesting, but I think they will hold the stable policy and keep the compatibility
11:40 <@seb128> they will
11:41 <@seb128> just pointing that 2.10 is not perfect yet
11:41 <@lool> sure
11:41 <@seb128> not sure if we should hurry it to etch
11:41 <@lool> I think we should, the transition is late enough already
11:41 <@seb128> ok, fine with me
11:42 <@seb128> that's just a choice to do between latest crack and stability
11:42 <@seb128> uptodate is fine with me ;)
11:44 <@lool> seb128: you're not saying that edgy is going to ship with a very broken Gtk, so let's say that we will have a Gtk better than edgy's with the time that we have between the two releases
11:44 <@lool> ;)
11:44 <@seb128> lool: I'm saying edgy is supposed to be "edgy"
11:44 <@seb128> but I don't expect a Debian stable to be edgy
11:45 <@seb128> that's a matter of what you expect from the distro
11:45 <@lool> so in Ubuntu stable releases, Gtk leaks memory and crashes?
11:45 <@seb128> dapper is the stable release
11:45 <@seb128> and will still be
11:45 <@seb128> edgy will be much less polished but with new crack, etc
11:45 <@lool> it's the LTS one, but isn't edgy stable?
11:45 <@seb128> we advice people who want a stable Ubuntu do keep using dapper
11:46 <@seb128> depend of what you can stable
11:46 <@seb128> it'll be stable in sense of "no new version of changes to it"
11:46 <@seb128> but we don't expect to have the polish level of dapper no
11:47 <@seb128> which doesn't mean we will not try to make it good
11:47 <@lool> ok, so since you're going to ship a good Gtk in edgy, I think we can ship an even better one in etch as we have more time, and upstream has more time
11:48 <@seb128> we expect corporate user to keep using dapper probably
11:48 <@seb128> ok
11:48 <@lool> and since you're already shipping Gtk 2.10 since many weeks, and that everyone is running edgy, I'm sure bugs will be sorter out :-P
11:48 <@seb128> what I was trying to say is that I expect a Debian stable being polish as dapper is, not as edgy will be
11:49 <@seb128> I'm sure GTK will have some leaks and issues to edgy but that's the price to pay to run latest crack ;)
11:49 <@lool> if we want to polish Gtk 2.10, we'd better start now, and we do want Gtk 2.10 for sure, even without GNOME 2.16
11:49 <@seb128> if you think that's fine for etch too fine with me
11:49 <@seb128> ok
11:50 <@lool> so, here's a short status report for Gtk 2.10: I could build the dfb and static flavors with many patches, dfb needs a cairo with different configure flags which the libcairo maintainer is not ready to do right now
11:50 <@lool> it's also missing the conflicts, which I'll build shortly, and some minor lintian / doc fixes
11:51 <@lool> does it need anything else?
11:51 <@seb128> dfb from 2.10 has different requirements of the backport Josselin did to 2.8?
11:51 <@seb128> cf cairo
11:51 <@lool> yes, because libcairo is required with PDF/PS support in Gtk 2.10 and I can't disable this
11:51 <@seb128> ah, k
11:51 <@lool> it's a bit stupid that it's not configurable BTW
11:52 <@seb128> nothing else required afaik, no
11:52 <@lool> ok, then I'll rebuild the conflicts list with our Contents.gz and write a wiki page for the transition, and try to prepare the release team to the transition
11:53 <@lool> (another thing to do is the 2.10.2 merge, but I'll continue on the 2.10.1 until it's complete)
11:53 <@lool> on the pango front, you moved the modules in /var/lib, which seems cleaner
11:53 <@lool> shouldn't we do this in gtk too?
11:54 <@seb128> yeah, might be a good idea to make the same sort of change for GTK
11:54 <@seb128> the move has been made by mvo because the previous way was breaking pango modules between the time the package content is installed and the time update-pango-modules is called
11:55 <@lool> is this because of .dpkg-tmp files as well?
11:55 <@lool> 41297
11:56 <@seb128> no, it's because pango look to the pango.modules and the path to that file was not matching the new location
11:56 <@seb128> like you have 
11:56 <@seb128> * pango module to 1.4 with config file pointing to it
11:56 <@seb128> * new version un packed, modules to 1.5 now but file still pointing to 1.4
11:57 <@seb128> * config file updated later when the postinst run pango-querymodules
11:58 <@lool> yeah, so you're moving the "low-level" list of modules to a location specific to the ABI?
11:58 <@lool> no, it's not specific, that's bad
11:58 <@lool> don't you think the same thing will happen with the next ABI change?
11:58 <@seb128> no, mvo did change the package to ship a config file with the package
11:58 <@seb128> so the config is unpackaged at the same time as the actual content
11:59 <@seb128> you don't have to wait on the postinst to get the correct location
11:59 <@lool> but what about modules which have added themselves to the list?
12:00 <@seb128> they are updated when update-pango-modules is runned later
12:00 < mvo> lool: those will only be updated later when pango-querymodules is run. but at least we will have a working pango between unpack and configure
12:00 < mvo> this broke e.g. debconf in gtk mode
12:01 <@seb128> hey mvo ;)
12:02 < mvo> hello seb128
12:03 <@lool> I'm still trying to understand why pango is broken between unpack and postinst; I understand it's broken between unpack and postinst for initial installs, but it's not clear to me what happens on upgrades
12:04 <@seb128> libpango1.0-0 modules are moved on the disk from /usr/lib/pango/1.4.0 to /usr/lib/pango/1.5.0
12:04 -!- Q-FUNK [~q-funk at dsl-hkigw7-fe10f900-10.dhcp.inet.fi] has quit [Ping timeout: 600 seconds]
12:04 <@seb128> and the config file is not updated yet, still pointing to /usr/lib/pango/1.4.0
12:05 <@seb128> so when debconf try to open a gtk dialog it shockes on pango not finding its modules where the config file is pointing
12:06 <@seb128> "Failed to load Pango module for id: 'BasicScriptEngineFc'Pango-WARNING **: pango_shape called with bad font, expect ugly output at /usr/share/perl5/Debconf/Element/Gnome/Note.pm line 35, <GEN1> line 5."
12:06 <@lool> ah the full path is in there
12:06 <@seb128> yep
12:06 <@lool> (I imaginated the list to be a list of module names expanded in a standard location)
12:07 <@lool> ok, so if the list is a regular file shipped in the package, the switch of the pango library happens at the same time as the switch of the module list, sounds good
12:08 <@seb128> right
12:10 <@lool> is it ok to keep it in /etc in the udeb?  sounds inconsistent to me
12:11 -!- Q-FUNK [~q-funk at dsl-hkigw7-fe10f900-10.dhcp.inet.fi] has joined #gnome-debian
12:11 <@seb128> I guess it could be changed for the udeb though
12:12 <@seb128> though the udeb is not likely to have a dist-upgrade issue
12:12 <@lool> in general, this file was never truly a config file since it was overwritten by pango on upgrades, so I'm okay with blowing away local user configs
12:12 <@seb128> but consistency is good
12:14 <@lool> does it break anything with respect to modules in other packages?
12:15 <@lool> +#if defined(__linux__) && defined (__i386__)
12:15 <@lool> uh
12:16 <@lool> is it required to have per arch dirs below /var/lib?
12:17 <@lool> I don't see anything about this in the FHS, and I don't expect /var to be shared; is this for bi-arch?
12:17 <@seb128> that's a hack from doko
12:17 <@seb128> I don't think Debian does that
12:17 <@seb128> we ship an ia32-libs package
12:17 <@seb128> on amd64
12:18 <@seb128> for openoffice, since it doesn't (or didn't) run on amd64
12:18 <@lool>  ia32-libs |       1.15 |      unstable | source, amd64, ia64
12:18 <@lool> we seem to do too
12:19 <@seb128> if you have the Ubuntu pango src package the hack is debian/patches/ia32-hack.patch
12:19 <@seb128> we have a similar patch for GTK
12:19 <@seb128> not sure how why that's not required for Debian if they use ia32-libs too
12:20 <@lool> actually, this looks very similar to the requests we had for multi-arch in the Debian BTS
12:21 <@lool> ok, finally there's the question of user configuration
12:21 <@lool> the list of modules never truly was configurable because it was overwritten, but upstream would like to see it configurable IMO
12:21 <@lool> and users too
12:21 <@lool> (for any module type)
12:22 <@seb128> why does it need to be configurable?
12:22 <@lool> to filter out some modules for example
12:22 <@seb128> the update-pango-modules should list everything no?
12:22 <@seb128> if you don't want some modules don't install them?
12:22 <@lool> imagine you don't want to list a module which is required by dependencies
12:23 <@lool> for example gtk's bundled pixbuf loader
12:23 <@seb128> right, but that looks rather a feature request to me
12:23 <@lool> yes
12:23 <@lool> I wonder if it's required though
12:23 <@seb128> we don't have a lot of complains about that afaik
12:23 <@seb128> not sure of the best way to do that
12:23 <@seb128> having an ignore list?
12:23 <@lool> we had that guy who needs alphanum ordering for his module to be loaded last
12:24 <@lool> presumably it's named zzz_module-name.so
12:24 <@lool> yes, I am pondering for some time to have a filter on our various modules; I would like to avoid implementing it at the same time as Gtk 2.10, but I ponder whether it's a required feature
12:26 <@seb128> that doesn't look like something many people would have use for
12:26 <@seb128> but still nice to have
12:34 <@lool> mvo: out of curiosity, what made one choose /var/lib over /usr/lib?
12:44 <@lool> actually, I think it's wrong to ship a file in /var in a package
12:44 <@lool> ARGH FUCK
12:44 <@lool> gnopernicus: /var/scrollkeeper/ca/scrollkeeper_extended_cl.xml
12:44 <@lool> *AGAIN*
12:44 <@lool> cortana: see, I told you
12:45 -!- sadleder [~sadleder at p50812A0D.dip0.t-ipconnect.de] has joined #gnome-debian
12:45 -!- sadleder [~sadleder at p50812A0D.dip0.t-ipconnect.de] has left #gnome-debian []
12:45 -!- sadleder [~sadleder at p50812A0D.dip0.t-ipconnect.de] has joined #gnome-debian
12:45 -!- sadleder [~sadleder at p50812A0D.dip0.t-ipconnect.de] has left #gnome-debian []
12:47 <@seb128> lool: we should get a lintian warning for that
12:47 < mvo> lool: I can't remember the rational anymore, it may just have been a mistake :)
12:49 <@lool> mvo: the module list is really dynamic, so it makes sense to have some tool generate it, and keep it in /var, but shipping a pre-computed one in a .deb seems wrong because of MD5 sums for example
12:49 -!- herzi [~herzi at kiwi.mediascape.de] has joined #gnome-debian
12:49 <@lool> it also means that if your generated file had been updated and you upgrade pango, it gets overwritten cruelly
12:50 <@lool> this wouldn't happen would be e.g. change the package name of pango
12:52 <@seb128> lool: that's why http://librarian.launchpad.net/2456323/pango1.0_1.12.2-0ubuntu2.debdiff2 has that
12:52 <@seb128> "-	dh_md5sums -a -Nlibpango$(apiver)-udeb
12:52 <@seb128> +	dh_md5sums -a -Nlibpango$(apiver)-udeb -Xvar/lib/pango/pango.modules"
12:52 <@seb128> I've just noticed that the edgy package has not that change
12:52 <@lool> seb128: yes, I've read that already
12:52 <@seb128> it probably got dropped while merging
12:53 <@lool> hmm in the udeb, that seems wrong
12:53 <@lool> but I've read about dh_md5sum in the launchpad bug
12:53 <@seb128> good comment
12:53 <@lool> it's what makes me think it's wrong to have the file in /var/lib
12:54 <@lool> ah no, it's not in the udeb
12:54 <@lool> -N is to exclude the package I suppose, my mistake
12:55 <@seb128> right
12:56 < mvo> lool: I agree that the current solution to ship the pre-generated version and overwrite it later is not good. but the old solution to have a broken pango (and in effect broken everything that depends on gtk) from unpack until configure (that may be a long time) is not really appealing either. I'm open for good ideas :)
12:56 < mvo> IIRC I described various possible solutions in the bugreport?
12:56 <@lool> I think I'll come back to my original idea: have one file per ABI
13:01 <@lool> there are two problems, the fact that the file is shipped in the .deb, and the fact that we want programs linked to pango to be launchable the most of the time; perhaps the solution is a two level list: if $some-dynamically-update-file exists, use it, otherwise use the pre-generated list of modules of the pango package itself
13:08 < mvo> the two level list sounds good to me
13:14 <@lool> mvo: actually, that's quite close to your /var/lib/pango/pango.modules.d idea, except your idea is more generic, but the complexity is quite close
13:37 -!- azeem [~mbanck at host109.natpool.mwn.de] has joined #gnome-debian
13:44 -!- decko [~decko at skywalker.ffclrp.usp.br] has joined #gnome-debian
14:06 -!- Anarka [anarka at c-b867e255.29-28-64736c14.cust.bredbandsbolaget.se] has joined #gnome-debian
14:13 -!- SpiritC [anarka at c-7a66e255.29-28-64736c14.cust.bredbandsbolaget.se] has quit [Ping timeout: 600 seconds]
-------------- next part --------------
=== modified file 'a/pango-1.12.3/pango/modules.c'
--- pango-1.12.3/pango/modules.c~	2006-08-04 11:24:59.777313056 +0200
+++ pango-1.12.3/pango/modules.c	2006-08-04 11:41:18.190571632 +0200
@@ -24,6 +24,8 @@
 #include <string.h>
 #include <limits.h>
 #include <errno.h>
+#include <unistd.h>
+#include <sys/utsname.h>
 
 #include <gmodule.h>
 #include <glib/gstdio.h>
@@ -489,9 +491,28 @@
   dlloaded_modules = g_hash_table_new (g_str_hash, g_str_equal);
   
   if (!file_str)
-    file_str = g_build_filename (pango_get_sysconf_subdirectory (),
+#if defined(__linux__) && defined (__i386__)
+    {
+      struct utsname uts;
+      char *sysconfdir = g_getenv("PANGO_SYSCONFDIR");
+      if (!sysconfdir || !access(sysconfdir, R_OK|X_OK))
+	{
+	  uname(&uts);
+	  if (!strcmp("x86_64", uts.machine)
+	      && !access(SYSCONFDIR "/pango32", R_OK|X_OK))
+	    sysconfdir = SYSCONFDIR "/pango32";
+	  else
+	    sysconfdir = "/var/lib/pango/";
+	}
+      file_str = g_build_filename (sysconfdir,
+				   "pango.modules",
+				   NULL);
+    }
+#else
+    file_str = g_build_filename ("/var/lib/pango/",
 				 "pango.modules",
 				 NULL);
+#endif
 
   files = pango_split_file_list (file_str);

More information about the Pkg-gnome-maintainers mailing list