[00:03] k, testing locally on karmic, then I'll up;load to hardy ppa [00:03] lots of build targets are required [00:04] well, I left them all [00:04] you should leave most. yes [00:04] just replace the logic tha currently deals with installing the stuff [00:04] with install-xpi i guess [00:05] and your new logic for getting new orig [00:05] would help if I spelt the build targets right (not dh_shlibdeeps) :) [00:05] lol [00:05] hmm. [00:05] don't worry [00:05] are we doing anything about the maxVersion? [00:05] I'm, fixing it all [00:06] yeah, I have it set to 3.0.* [00:06] ah ... that happens in update-debian-files? [00:06] anyway. [00:06] you misspelled dh_shlibdeeps ? [00:06] that means you didnt start with the original one :( [00:06] right [00:06] we will see [00:06] hopefully the diff is minimal :;) [00:07] I know mine works..now I just need the old stuff [00:07] anyways [00:07] i know the old thing works [00:07] ;) [00:07] http://pastebin.com/f2e34c6b7 line 39 in the new rules file [00:07] micahg: create a diff -u [00:07] but that changed [00:07] and paste that [00:08] http://pastebin.com/f2eb71257 [00:09] -build-stamp: configure-stamp [00:09] +build-stamp: configure-stamp [00:09] - [00:09] (23) [00:10] row 71 [00:10] 75 [00:10] does it not like my spaceS? [00:10] 140 [00:10] not sure [00:10] make those diffs go away ;) [00:10] k [00:11] 149 152ff [00:11] they have extra spaces before :) [00:12] whats the difference of update-xpi and rebuild-debian-files ? [00:12] just the wget? [00:12] asac: yes [00:12] what uses rebuild-debian-files? [00:12] or is that external API? [00:12] external if you need to fix the control file after you download [00:13] make update-xpi: wget rebuild-debian-files [00:13] i guess [00:13] k [00:13] should I try pushing to hardy? [00:14] + cat $(CURDIR)/debian/locales.list >> $(CURDIR)/debian/packages.list [00:14] + sort -u $(CURDIR)/debian/packages.list > $(CURDIR)/debian/packages.list.new [00:14] + mv $(CURDIR)/debian/packages.list.new $(CURDIR)/debian/packages.list [00:14] that probably should be [00:14] oh [00:14] is the >> intentional? then nevermind [00:14] consider to make emtpy line in between [00:14] yes, so we keep all the previous ones [00:15] so we always have a stub package if we published a translation previously [00:15] * micahg should probably comment :) [00:15] the sort gets rid of duplicates? [00:15] yep [00:16] second and third line might be cat FILE | sort -u > $(CURDIR)/debian/packages.list [00:16] combinable [00:16] maybe do this ;) [00:16] even though I'm writing to the same file I'm reading? [00:17] cat $(CURDIR)/debian/packages.list $(CURDIR)/debian/locales.list | sort -u > $(CURDIR)/debian/packages.list [00:17] hmm [00:17] oh [00:17] my brain still had the > in it [00:17] but you can combine 1 and 2 like it ;) [00:17] i would hope [00:18] but not important enough [00:18] so just keep ;) [00:18] asac: before I had uniq FILE | sort > [00:19] micahg: ok. did you do a local test build? [00:19] do the languages install? [00:19] yes, works fine in karmic [00:19] ugh [00:19] jsut erased it [00:19] please test the following languages: de, fr, es, pt, en_GB, if more you can take them [00:19] oh and also zh [00:19] (if you do release tests of locales) [00:20] how do you list the files in a .deb again? [00:20] if you have the deb? [00:20] yeah [00:20] installed its dpkg -L packagename [00:20] dpkg -c filename [00:21] yep, appears to be working :) [00:21] install them . test them [00:21] and then upload to a ppa for lucid [00:21] and then try hardy [00:21] is it enough to just install the .debs locally to test or do I need the language-selector test again? [00:21] no [00:21] just install locally [00:21] and use [00:21] k [00:21] export LANG=de [00:21] firefox [00:22] export LANG=zh_TW [00:22] firefox [00:22] export LANG=zh_CN [00:22] firefox [00:22] s/firefox/thunderbird/ [00:22] export LANG=fr [00:22] etc. [00:22] oh, I was using the language addon in TB to switch [00:22] thats not how we do it [00:22] we just use LANG env ;) [00:22] k [00:22] you probably need to uninstall that [00:23] oh ... nowadays we seem to have issues with some extensions [00:23] when starting firefox twice [00:23] so start twice to test ;) [00:23] second time for greasemonkey it doesnt start anymore for instance [00:23] hardy worked now :) [00:23] same could be true for translations ... until someone has debugged that we should use extra care [00:24] good ;) [00:24] well done [00:24] thanks [00:24] maybe give the full roundtrip a try too ... e.g. does updating xpis sstill work etc. [00:25] * micahg will have to remove the broken .xpis again, but ok [00:26] broken xpis? [00:26] yeah, tehre were 3 languages that had whitespace [00:27] ah [00:27] yes [00:29] bug...package names need to be lowercase... [00:29] probably ;) [00:35] all is ok now :) [00:38] test build in ppa on lucid and karmic [00:38] asac: should I spin a TB3 build as well? [00:43] bdrung: thanks, all is working on hardy :) [00:55] micahg: yes [00:55] k [01:12] asac: is it worth denting about pushing stuff to my personal PPAs? [01:13] asac: also, have you reviewed my changes for TB3? [01:31] asac: TB3 built fine on lucid amd64 [01:31] i think all security stuff should be up now [01:32] minus xul 1.9.1 lucid [01:32] k [01:32] that's great then [01:32] 63 pending builds [01:32] BTW, debian apaprently has a fix for sparc [01:32] well [01:32] we should disable-jit there [01:32] until we have upstream fix [01:32] we have that fix in 1.9.1 i thought [01:33] at least karmic ? [01:33] for sparc? [01:33] yeah [01:33] so in karmic 1.9.1 builds for sparc [01:34] ah, yes [01:35] so, that needs to be ported fwd as well? [01:35] to 3.6? [01:36] yes. though i think there is an upstream fix pending [01:36] armin76 would know ;) [01:37] I get this now ... [01:37] (process:20696): Gtk-WARNING **: Locale not supported by C library. [01:37] Using the fallback 'C' locale. [01:37] when I try this: export LANG=de [01:37] thunderbird [01:38] is it in german? [01:38] hmm [01:38] might be you need for language packs [01:38] but tbird should be german still, isnt it? [01:38] no [01:38] try [01:38] de_DE [01:39] export LANG=de_DE [01:39] or [01:39] export LANG=de_DE.UTF-8 [01:39] niope [01:39] then there is something wrong ;) [01:39] the warning shouldnt be the problem [01:39] * micahg probably needs the langpacks [01:40] try it... but afaik, tbird doesnt care for that [01:40] * micahg has it [01:40] no go [01:40] rather check if all is right [01:40] when I use the locale switcher, it works [01:40] have you uninstalled that now? [01:40] yes [01:40] the langpacks show up in tb [01:40] go to config editor [01:41] search for matchOS [01:41] most likely that messed with that setting [01:41] ah, I might not have that... [01:41] it should be default aka true [01:41] you should have that [01:41] but probably set to false [01:41] false :) [01:41] from the extension shit [01:41] ;) [01:41] yeah [01:41] sorry for bad tone [01:42] says the default is false [01:42] thats a problem then [01:42] for firefox we ship it like: [01:42] grep matchOS /etc/firefox/pref/* [01:42] pref("intl.locale.matchOS", true); [01:42] so in the syspref file [01:42] * micahg is chcecking with a clean profile [01:42] most likely we dont even have a syspref patch? [01:42] yes, try everything clean [01:43] false [01:43] we dont even ship /etc/thunderbird ? [01:43] we need the syspref feature from firefox [01:43] and the link from APPDIR/defaults/sysprefs /etc/thunderbird/pref [01:43] yes,w e do [01:44] http://pastebin.com/f37dff5ad [01:44] micahg: thats not shipped by our tbird 3 package, or is it? [01:44] yes, that's from the tb3 package [01:44] if its shipped we need a patch and a link like above [01:45] APPDIR/defaults/sysprefs /etc/thunderbird/pref [01:45] like in firefox [01:45] b ls debian/patches/*sys* [01:45] debian/patches/add_syspref_dir.patch [01:46] thats what we need in tb3 i guess [01:46] for the syspref dir [01:48] should I pull hte patch from ff3.5? [01:48] or rather xul191 [02:04] I grabbed the xulrunner patch and it's bulding now [02:21] micahg: from 3.6 [02:21] ffox [02:21] but i think it just applies [02:21] from 1.9.1 [02:23] asac: I pulled the patch from xul191 since that's what TB3 is built on [02:23] right [02:23] all good i guess [02:23] should be almost done building [02:23] micahg: remember that the link needs to be created [02:23] are the rest of my patches ok? [02:23] no idea [02:23] ;) [02:23] * micahg forgot the link...I'll let it build first :) [02:25] i assume you picked the other patches from xul etc.? [02:25] then they should be fine [02:25] please commit the rest of the stuff (incl. the link and the script) i will review and upload tomorrow [02:26] you can make a release commit on your branch ... just in case that its perfect ;) [02:26] try a few cases for the profile migration and see if decide later makes it pop up again and so on [02:26] i have to drop to bed ;) [02:26] the locale packages are in your ppa now? [02:26] yes [02:27] the debhelper 5 ones? good [02:27] well, I have it for dh6 [02:27] since that's what hardy has [02:27] i am quite sure your script works with 5 too [02:27] probably :) [02:27] the previous rules didnt have anything newer [02:27] please use whatever was used in the previous package [02:27] that was 4 [02:27] then that [02:27] and I had lintian complaints [02:28] i think 5 is decent old [02:28] but whatever. i might change it on upload [02:28] k [02:28] it definitly needs to match what you have as lower versoin in control [02:28] that's what I set both at 6 [02:29] well, control is 6~ [02:29] do I need a .js file for tb [02:29] use hte one from the old pakcage [02:29] package [02:29] tbird 2 [02:29] cat /etc/thunderbird/pref/thunderbird.js | pastebinit [02:29] http://pastebin.com/f5532a62 [02:30] dont ask me why we have those ;) [02:30] matchOS is important [02:30] no idea about pref("extensions.update.enabled", true); [02:30] feels like its really old [02:33] ok, I got rid of the line and added it to the .install file [02:34] the line? [02:34] you mean dh_install? [02:34] debian/thunderbird.js etc/thunderbird/pref [02:34] k [02:34] now,do i need to link that to the TB dir in addition? [02:34] yes [02:34] like in ffox3.6 [02:35] k [02:38] got it [02:41] * micahg is changing compat back to 4 [02:43] asac: good night [02:43] I'll try to be on around 13:00 UTC [02:59] asac: locales work now :) [05:01] asac: orz... respinning language-pack updates now. [07:17] asac: uploading [08:41] morning [10:22] ArneGoetje: i wanted to update the searchplugins today ... now we need another run :( [10:22] gaa [10:23] asac: actually I need to leave now... how much time do the changes need? [10:23] no clue [10:23] i will talk to pitti [10:23] just give me instructions how to run that stuff [10:23] i dont like the idea doing it manually :/ [10:23] i have all the info now [10:24] asac: respin the language packs? [10:24] yes [10:24] we need to do that [10:25] asac: the latest tarball for lucid in unpacked in /srv/language-packs.ubuntu.com/langpack-o-matic/temp [10:26] asac: edit rosetta-lucid/timestamp.txt and bump the version number to .3 and +1 for every respin. [10:26] ArneGoetje: so this time bump it to .3, right? [10:26] ArneGoetje: and how to run? [10:26] asac: from temp/, build a new tarball, so that the top level directory in the tarball is rosetta-lucid/ [10:26] asac: yes [10:27] when you have the tarball, cd to /srv/language-packs.ubuntu.com/langpack-o-matic [10:27] asac: the run ./import $tarball lucid 1> ../logs/lucid.log 2>../logs/lucid_errors.log [10:27] does the tarball name matter? [10:28] asac: no [10:28] ok [10:28] and then? [10:28] asac: then check the logs and tarballs in ../lucid/sources-update/language-pack-$languagecode/data/mozilla.tar.gz [10:29] asac: if everything is fine, run dchroot -d /srv/language-packs.ubuntu.com/langpack-o-matic/packages upload [10:29] asac: done. [10:29] hmm ok [10:29] ArneGoetje: i dont see any langpacks on the builders [10:29] does that mean it just takes a few hours to build all? [10:30] asac: are they in the queue? [10:30] which queue? [10:30] asac: or maybe built already? [10:30] https://edge.launchpad.net/ubuntu/+source/lucid/+queue [10:30] https://edge.launchpad.net/ubuntu/lucid/+queue [10:31] no idea [10:31] ArneGoetje: does that automatically produce update langpacks? [10:32] its odd... i thought we push 100 langpacks or so [10:32] how can it be that they already finished? [10:32] i see the -nl is already build [10:32] asac: the updates are not that big [10:32] asac: those are not full-exports from rosetta, just deltas [10:32] what does it mean? [10:33] ok [10:33] so langpack-o-matic checks later if there were any changes and decides to not upload all? [10:33] asac: yes [10:33] where can i see what was uploaded? [10:33] ArneGoetje: how does it determine that? [10:34] now that we have searchplugins its more than just the .po files [10:34] does it produce all sources and then compare the content of those sources? [10:34] ArneGoetje: i need to run everything as langpack from above? [10:35] asac: no, it goes through the language codes in the rosetta tarball. [10:36] and compares the tarballs? [10:36] asac: only those languages which actually have updates in the tarball get built. [10:36] ok [10:36] makes sense [10:36] thanks [10:36] enjoy your holiday [10:36] and sorry for the disturbance ... i really didnt expect you to come back ;) [10:37] asac: IOW, when we implement the searchplugins stuff for all languages, we need to get new -base packages, means a full-export from rosetta. And then need to change the code in po2xpi to detect changes in the searchplugins folder, since they are independent from the rosetta-tarballs. [10:38] asac: currently it doesn't do that [10:38] asac: but languages, which ave upstream XPIs, get updated in any case. [10:40] asac: if you have trouble or questions, ping me later, I will probably be online again after 4 hours, when my kids are asleep. [10:40] i dont get that. [10:40] i thought rosetta tarballs means the .tar.gz produced by rosetta [10:40] i thought mozilla.tar.gz is included in each of that [10:41] asac: but from tomorrow on I will be travelling with my family for the rest of the week and won't always have internet connectivity. [10:41] so if the searchplugins change ... the mozilla.tar.gz changes ... the rosetta tarball changes [10:41] asac: no. [10:41] ArneGoetje: thanks. i think its ok this week [10:41] i just need to change the existing searchplugins ... or maybe add one more (or does that not even work?) [10:42] asac: the rosetta tarballs only include the .po files of those languages which have changes. [10:42] hmm [10:42] asac: the mozilla.tar.gz tarballs are produced by po2xpi [10:42] langpack-o-matic should do a full run and then compare the output rather than the input [10:42] ArneGoetje: i know. thats why i wondered if it works ata ll if i rerun [10:42] asac: if you are volunteering to write the code to do that, go ahead. ;) [10:43] why does it work if i rerun now? [10:43] in my book the .po files didnt change [10:43] -> searchplugins wont get included [10:45] fta: https://launchpad.net/bugs/522078 [10:45] Ubuntu bug 522078 in chromium-browser "Crashes reproducibly when trying to edit events in Google Calendar" [Undecided,New] [10:45] asac: it doesn't matter if the po files changed. If they are present, the language will get updated. [10:45] mdz says its not crashing from beta channel ppa, but from archive [10:46] asac: and somehow those languages with upstream xpis also get rebuilt... need to check the code in po2xpi to figure out how this works and why [10:47] ah ok [10:47] ArneGoetje: so it means we might not have gotten all langpacks? [10:47] err searchplugins [10:47] have you checked that at least those languages were included? [10:48] asac: I have only checked de and zh-hant, not all languages [10:48] ok. [10:49] i will see [10:49] enjoy your holiday [10:49] asac: ok, thanks [10:51] fta: how can we verify that our release really matches the beta channel? === BUGabundo_remote is now known as BUGabundo_lunch [12:15] asac, http://code.google.com/p/chromium/issues/detail?id=34725 [12:17] fta2: so go for in-source? [12:17] what are we still using from system? [12:17] gtk, nss/nspr? anything else? [12:17] on jpeg? [12:18] lets upload with libxslt also dropped then [12:18] sigh [12:18] GYP_DEFINES += \ [12:18] use_system_bzip2=1 \ [12:18] use_system_zlib=0 \ [12:18] use_system_libjpeg=1 \ [12:18] use_system_libpng=1 \ [12:18] use_system_sqlite=$(USE_SYSTEM_SQLITE) \ [12:18] use_system_libxml=0 \ [12:18] use_system_libxslt=0 \ [12:18] $(NULL) [12:18] USE_SYSTEM_SQLITE=0 [12:18] so just bzip2, jpeg and png [12:18] right. [12:19] they dont have their own apng? [12:19] i don't think so [12:19] mm [12:19] wonder if we should just set everything to 0 ;) [12:19] probably [12:20]