[00:14] fta: i think we will get issues with firefox-3.5 branch tomorrow [00:14] i committed take-over for 3.6 [00:14] if you renamed the src package, the bot will sure have troubles [00:14] yeah [00:14] its now "firefox" [00:14] ;) [00:14] bloody mess i tell you [00:15] hmm [00:15] so firefox-3.6.head is now firefox [00:15] asac: ff3.5 branch broke in dailies today [00:15] can you fix that at least? [00:15] micahg: i saw [00:15] or wait [00:15] the conf needs to be updated [00:15] the mozclient should be ok in the 3.6.head branch [00:16] micahg: i think starting tomorrow it doesnt really matter ;) [00:16] asac: why? [00:16] oh, right [00:16] we could move the 3.5.head branch either to be a firefox-old source [00:16] or drop it [00:17] asac: drop from daily? [00:17] as its now, the upload of binaries will fail because the 3.6 provides higher versioned transitional packages :) [00:18] if we dont want to rework the branch in the same way as i did with 3.6 now [00:18] just with name firefox-old [00:18] but lets wait [00:18] want a respin now? [00:18] only thing i am sure about is that there will be a firefox stable release channel ;) [00:18] hmm [00:18] fta: can you just try the 3.6 branch? [00:18] yes [00:18] fta: maybe just for lucid/karmic? [00:18] ;) [00:18] nm [00:18] hm [00:19] if thats difficult, just push it. [00:19] no more xul? [00:19] it worke dhere [00:19] yes [00:19] xul is universe now [00:19] i tried to keep the packaging still working [00:19] not sure if that can be maintained though [00:19] ;) [00:19] 'vpattern' => '^3\.6(\~|\.)', [00:19] useless then? [00:20] unless we make a second ppa where we build it [00:20] fta: what does that do? [00:20] fta: xulrunner is quite important ;) [00:20] asac: I think we should keep building it for devs testing stuff, but maybe a new ppa [00:20] otherwise every rdepend we cannot remove would have to build against firefox [00:20] which is what we really dont want [00:20] micahg: 3.5? [00:21] asac, it's a test to make sure the tarball is not x-3.7 when x is supposed to be 3.6 [00:21] asac: at least 3.6 and 3.7 [00:21] I think we should keep doing 3.5 until we migrate all releases to 3.6 [00:21] fta: hmm. maybe something similar based on the previous version ;) [00:21] e.g. dont bump from 3.5 to e.6 [00:21] 3.6 [00:21] micahg: well, our dailies have .head [00:21] micahg: we are preparing the transition there [00:21] for lucid ... and later everywhere else [00:22] right [00:22] basically already everywhere elkse [00:22] so next step is also to add -3.0 transitional packages to 3.6. head [00:22] its a tough topic ;) [00:23] running [00:23] what i would love to do is to have firefox-previous -next -dev [00:23] so we could make firefox-previous out of that [00:23] nice, i can control my new radio from my webbrowser and from anywhere [00:23] heh [00:24] better than a not controllable laptop ;) [00:24] asac: indeed, sounds good, problem with firefox-previous is what to do when branch is EOL? [00:24] or we just keep it rolling so -previous is never EOL [00:25] yes. thats my main concern. [00:25] we dont know if there will be -previous in future [00:25] k [00:25] there are two approaches in general i think: [00:26] 1. if -previous is dead, it means that -previous is the same branch as firefox [00:26] upstream branch i mean [00:26] ;) [00:26] still with the different profile etc. [00:26] BTW, we can't have firefox-dev branch since that would conflict with package names, firefox-trunk is probably better [00:26] micahg: somewhat true. so back to what we had initially ;) [00:27] though i will drop firefox-dev before the first archive upload [00:27] ah, right [00:27] the idea is that noone can build depend against it ;) [00:27] but right. would be funny if someone previously having the -dev package installed suddenly tracks the -dev branch [00:27] ok, so I guess we can do firefox-dev, but firefox-trunk seems to make more sense since firefox-next is also -dev [00:28] i want to wait for a upstream decision [00:28] k [00:28] if they give their release channels a real name we can use that [00:28] like if they start doing -beta channel etc. [00:29] micahg: ok. so 3.5 only failed in lucid [00:29] i think its really the same issue we have for thunderbird [00:29] bash or dash or something [00:29] hmm. only amd64 [00:29] nm it was 3.6 [00:29] we have a patch for that [00:29] yes, 3.6 was half baked last night [00:29] we have a patch for the shell issue? [00:29] yes [00:30] it seems random [00:30] where is it? [00:30] kees submitted it with his new xulrunner hardening branch. i found it by accident. he didnt understand why [00:30] ah, ok, I'll look there [00:30] let me check [00:30] he has a bug about that [00:30] bug 507744 [00:31] Launchpad bug 507744 in xulrunner-1.9.1 "build with PIE to gain remaining ASLR support" [Medium,In progress] https://launchpad.net/bugs/507744 [00:31] yep, I saw, I was waiting for you on it [00:31] feel free to merge it. cleanup changelog documentation a bit maybe [00:31] e.g. document the xulrunner patch explicitly [00:31] then put the same patch in firefox-3.5 and tbird and firefox-3.6 and xurlunner-19.2 ;) [00:31] asac: k [00:32] actually the changes he submitted need to go in all branches ;) [00:32] + the patch from xulrunner also needs to go to all firefox .head branches [00:32] k, I need to add dh_xulrunner to 1.9.2 and 1.9.3 still also [00:32] so I can both at the same time [00:32] let me commit the patch to 3.6 [00:32] otherwise it will fail [00:33] fta: already pushed? [00:33] ;) [00:33] maybe hold back :( [00:33] FAILED [00:33] *ERROR* firefox != firefox-3.6 [00:33] good ;) [00:33] that's vpattern [00:33] i have to pick a patch and test that it applies [00:34] no, not vpattern, i have to rename the package in my conf file [00:34] my keys are source package names [00:36] hmm [00:36] yeah its branch name now [00:36] and source package name is in theory indpendent [00:38] ok applied that patch to 3.6.head for now [00:38] fta: i think we are ready from this side ;) [00:40] respining [00:41] asac: when should I clean up the changelog in ff3.6.head? [00:42] micahg: already tried to clean it up a bit ... needs more love [00:42] asac: k, when are you uploading to lucid? [00:44] asac: is there a bug for NM "For All Users" nor working in lucid with WiFi? [00:44] works for karmic, but I never manage to get it workign in +1 [00:45] micahg: wanted to see a green daily build [00:45] then doing the official branhding and uploading it [00:45] when upstream releases [00:45] assuming that its ok now [00:45] asac: k, can I fix the changelog tonight [00:46] sure. just clean it up and remove redundant stuff [00:46] asac: k, I'll also add a changelog entry for all the bugs fixed [00:46] yes, adding a 3.6 released header to the changelog sounds good [00:47] micahg: ensure that the bugs closed are also filed against firefox source package [00:47] asac: ? [00:47] otherwise they dont get closed [00:47] ah, k [00:47] micahg: like what we do for security updates on top [00:47] * firefox 3.6 release [00:47] - fix LP: #.... [00:47] - fix LP: #... - this bug [00:47] etc. [00:48] asac: oh, ok, I didn't remember that's how it's done [00:48] firefox 3.6 release (FIREFOX_3_6_RELEASE) [00:48] I mean about the busg [00:48] micahg: check the stable branches [00:48] yeah. i like that way [00:48] you can also use + instead of - ;) [00:48] k, I messed up TB then ;) [00:48] heh [00:49] i think a typical structure of a changelog is: [00:49] * UPSTREAM RELEASE x.y.z (UPSTREAM_TAG) [00:49] - fix LP: #1 - microsft has majority of market share [00:49] . [00:49] . [00:49] [ Master Luke ] [00:49] * packaging change 1 [00:49] * packaging change 2 [00:49] etc. [00:49] asac: good to know [00:50] asac, done [00:50] fta: how is the new "consistency" check? [00:50] unchanged [00:50] e.g. ensuring that no bad version gets uploaded? [00:50] hmm [00:50] ok [00:51] so that still works? [00:51] * asac happy [00:51] asac: if the patch works in older versions, should I upstream it? [00:51] firefox is building in daily ;) [00:51] http://paste.ubuntu.com/359835/ [00:52] micahg: do we see the issue in xulrunner trunk builds? [00:52] asac: yes [00:52] if so, yes. remember to give credits to kees like in changelog [00:52] and CC asac@jwsdot.com ;) [00:52] asac: where do I put that in an upstream patch? [00:52] * asac might be able to read bug mail soon again ;) [00:53] you just submit the patch ... and say in the submit comment that you forward that patch; credits go to Kees... [00:53] maybe remember them when you ask for the commit [00:53] asac: k [00:57] fta: can you disable 3.5 daily for now? [00:57] or want to keep it failing ... i dont mind if noone complains about ppa usage [00:57] ? [00:58] as you want, i don't mind either way [00:58] ok. lets keep it running for a few days and see if we get to fixing it [01:08] asac: ff3.5 worked last night [01:09] good [01:10] i think it failed on lucid [01:10] * micahg hopes to have more time next week,,, [01:19] asac: do we need all the packages updated by Feature freeze? [01:20] TB, SM, Lightning, Sunbird [01:21] fta: oh. can you also push karmic? [01:22] ;) [01:22] i wanted to test the upgrade path more or less extensively [01:22] if not i can push that to my sandbox [01:22] micahg: good question. in general yes. [01:22] FF and TB need to happen this week ;) [01:22] heh, ok [01:22] hehe [01:22] TB will be kind of half baked at first [01:23] without -dev package etc. [01:23] but debian icedove folks seem to make progress on that [01:23] asac: I wanted to suggest branching tb.head before my changes and overwritting [01:24] the final result is ok for most things, but the commits aren't so good [01:24] yes [01:25] i wanted to redo that with merges etc. [01:25] and copying your files [01:25] did the build work? [01:25] no [01:25] I missed a few things in the rules file when I merged them === asac_ is now known as asac === micahg1 is now known as micahg [08:46] morning [08:49] morning [08:49] you guys need squad members at that upcoming meeting? [08:51] Has anyone thought about making a dh_xulrunnerdeps to emit substvars for packages that link against xulrunner libs? [08:54] in english? [08:59] A debhelper tool to generate ${xulrunner:Depends} variables with appropriate versioning. [09:10] ah [09:10] no clue. [10:26] RAOF: thought we did that now [10:26] in lucid [10:29] asac: Really? [10:29] I'd find that quite useful - where is it! :) [10:34] apt-file search dh_ doesn't find anything xul-ish. [10:38] I guess I could write one. It can't be that hard, right? :) [10:40] let me check [10:43] Aha! I had an out-of-date apt-file cache. [10:45] Sweet. An idea so good, someone had it first! [10:54] asac: http://blog.mozilla.com/tglek/2010/01/19/chromium-vs-minefield-cold-startup-performance-comparison/ [10:54] Firefox commandline: firefox -profile /mnt/startup/profile/firefox -no-remote file://`pwd`/startup.html#`python -c ‘import time; print int(time.time() * 1000);’` [10:54] Chromium commandline: chromium-browser –user-data-dir=/mnt/startup/profile/chrome file://`pwd`/startup.html#`python -c ‘import time; print int(time.time() * 1000);’` [10:54] I can't run FF [10:55] and Ch doesn't work much better either [10:55] but at least it opens [10:56] heh [10:56] BUGabundo_remote: i will be here in a minute [10:56] BUGabundo_remote: did you upgrade to latest daily firefox-3.6 yet? [10:56] that should give you startup speed etc. [10:56] BUGabundo_remote: use this url:http://people.canonical.com/~asac/tmp/startup.html [10:56] for that test [10:59] Darn. dh_xulrunner doesn't pick up libmozjs dependencies. Ah! Because it's taken from Debian. [11:02] I guess I'll re-add the handcrafting and write a patch for dh_xulrunner. [11:03] asac: I'm using 3.7 [11:05] asac: in Ch I get urls like http://xn--user-data-dir%3D-bg6i/tmp/chrome :( [11:05] and file:///home/bugabundohttp://people.canonical.com/~asac/tmp/startup.html#1264071912437 [11:06] $ chromium-browser –user-data-dir=/tmp/chrome file://`pwd`http://people.canonical.com/~asac/tmp/startup.html#`python -c 'import time; print int(time.time() * 1000);'` [11:08] firefox-3.7 -profile /tmp/firefox -no-remote file://`pwd`http://people.canonical.com/~asac/tmp/startup.html#`python -c 'import time; print int(time.time() * 1000);'` [11:08] bugabundo@BluBUG:~$ psx fox [11:08] 1000 25618 0.0 0.0 7404 744 pts/7 D+ 11:08 0:00 grep --color=auto fox [11:10] BUGabundo_remote: did you have firefox-3.6 installed? [11:10] wonder if your daily update today went well [11:11] BUGabundo_remote: remove the file://`pwd` [11:11] from the url [11:11] just the http:/.... [11:29] BUGabundo_remote: would appreciate if you can check that firefox 3.6 still works nicely ;) ... we shuffled the packaging quite a lot ;) [11:29] so if you had issues during upgrade i would want to know [11:44] hallo, anyone aware of a PPA problem? [11:45] Tallken: let me know what probably you see [11:45] asac, I was gonna paste a pastebin URL [11:45] we did some package shuffeling, so some issues might come from that [11:45] right [11:45] please post what your issue is [11:45] is firefox-3.5 supposed to bring firefox-3.6 ? [11:46] yes [11:46] then who has both firefox-3.5 installed and firefox-3.6 will run into troubles? [11:46] we made the firefox 3.6 by default transition the last few days [11:46] Tallken: yes, you need to remove firefox-3.6 [11:47] we didnt add a transition for those packages, because they never entered the archive [11:47] and putting more packages in is overkill [11:47] ok, makes sense [11:47] i might upload a special package to the daily ppa to ensure a good transition [11:47] with just that transition [11:47] but that way you have to hear ppl like me complaining :p [11:47] already thought about that [11:47] will probably do that [11:47] Tallken: thats ok. [11:47] ;) [11:47] :p [11:47] Tallken: let me know if you have issues after removeing firefox-3.6 [11:47] packages [11:48] yo should end up with firefox and firefox-3.5 (empty/transitional) packages [11:48] and your default browser should be 3.6 then [11:49] apt-get is sooooooo annoying when this happen [11:49] I want to remove a package and he wants to install first the ones that failed [11:49] using aptitude [11:50] ......downgrading to previous version of firefox...... [11:50] ......removing firefox-3.6 [11:52] and finally, using apt-get to dist-upgrade, yay :D [11:52] asac, thank you for your help [11:52] if anyone else comes here, mention aptitude will do a better job [11:52] removing firefox 3.6 [11:52] just refuse the first solution it presents [11:53] and then it will offer to downgrade the packages back to the old firefox-3.5 (really 3.5) [11:53] and then the system is in consistent state and you can remove firefox-3.6 [11:56] Tallken: ok. i think we should make that special package for 3.6 users [11:57] hum [11:57] or, i'm thinking about this now [11:57] just tell ppl to install firefox-3.6 first now [11:57] I assume it's empty now? [11:58] then, when installing firefox-3.5 there won't be any file collision [11:58] the problem here was that apt-get would kill the install because it wanted to replace a file which belonged to another package [11:58] so, I assume if firefox-3.6 gets installed first, no files will be assigned to firefox-3.6 (assuming the firefox-3.6 package has now no files) and firefox-3.5 will complete with no issues [12:00] Tallken: no. thats not it [12:00] firefox-3.6 package is still in that ppa, from the previous version [12:00] we dont ship an empty package in the new package [12:00] because that would mean we would have to push that to real archive ... and we never had that package there [12:00] then you'll run into troubles :p [12:00] dpkg: erro ao processar /var/cache/apt/archives/firefox_3.6~hg20100120r33527+nobinonly-0ubuntu1~umd2~karmic_i386.deb (--unpack): [12:00] a tentar reescrever '/usr/lib/firefox-3.6pre/components/libnkgnomevfs.so', que também existe no pacote firefox-3.6-gnome-support 0:3.6~hg20100117r33523+nobinonly-0ubuntu1~umd1~karmic [12:00] yes [12:00] it says: error processing [12:00] sure [12:00] ...trying to rewrite ... which also exists in package [12:00] right [12:01] its half a bug, half a missing transitional package for the ppa-only [12:01] DO NOT remove the idicater-applet (the one with your name) than try to add it back. You will lose your gnome-panels [12:02] Tallken: please dpkg -r firefox-3.6-gnome-support [12:02] and try again [12:02] Tallken: do you get more cnflicts? [12:02] * asac currently fixes packaging [12:02] force it will fix it [12:02] asac, no, no, no, I'm good now :) [12:02] asac, what I was saying was just discussing the thing [12:02] ^^^ ugly way [12:03] as I said, I used aptitude to handle it "automatically", it downgraded to the previous version (in which firefox-3.5 was firefox-3.5) and then I was able to remove [12:03] I was just wondering any other possible solutions, but personally I've everything correct now [12:03] Tallken: ok committed. so that error shouldnt happen anymore [12:04] :) [12:04] the gnome-support one [12:05] then if that error won't happen, ppl will just have two Firefox 3.6 instaled at the same time without noticing it :p [12:06] bugabundo@BluBUG:~$ python -c 'import time; print int(time.time() * 1000);' [12:06] 1264075331753 [12:06] bugabundo@BluBUG:~$ python -c 'import time; print int(time.time() * 1000);' [12:06] 1264075332732 [12:06] bugabundo@BluBUG:~$ python -c 'import time; print int(time.time() * 1000);' [12:06] 1264075333176 [12:06] bugabundo@BluBUG:~$ python -c 'import time; print int(time.time() * 1000);' [12:06] 1264075333533 [12:06] bugabundo@BluBUG:~$ python -c 'import time; print int(time.time() * 1000);' [12:06] 1264075333883 [12:06] bugabundo@BluBUG:~$ python -c 'import time; print int(time.time() * 1000);' [12:06] 1264075334310 [12:06] hum... er? [12:07] asac: how can that simple test have so much diff results?= [12:07] BUGabundo_remote: you just print the time [12:07] who will the browser test be coeheren ? [12:07] since current time moves on - as we all know [12:07] asac: yeah, I still haven't manage to get any browser to open with it :( [12:07] you dont run the browser test with that command line [12:07] you just run python to print hte time [12:07] BUGabundo_remote: then dont do it ;) [12:08] i told you how ;) [12:08] anyway, thank you! idleing [12:08] $ firefox-3.7 -profile /tmp/firefox -no-remote http://people.canonical.com/~asac/tmp/startup.html# python -c 'import time; print int(time.time() * 1000);' [12:08] prints nothing [12:08] chromium-browser –user-data-dir=/tmp/chrome "http://people.canonical.com/~asac/tmp/startup.html#"`python -c 'import time; print int(time.time() * 1000);'` [12:09] ELAPSED 704 [12:09] YAY [12:09] firefox "http://people.canonical.com/~asac/tmp/startup.html#"`python -c 'import time; print int(time.time() * 1000);'` [12:09] yes [12:09] asac: do you know what app to file a bug against for update-apt-xapi [12:09] that works [12:09] gnomefreak: no. dont know what that is [12:10] dpkg -S update-apt-xapi [12:10] gives you the package the file is shipped in [12:10] thanks checking [12:10] so its apt-xapian-index [12:11] aahhh -profile /tmp/firefox is causing probs [12:11] * Tallken wonders why firefox-3.6 now that it is in the place of firefox-3.5 has the antialiasing or something different [12:11] Tallken: font issues? [12:11] no, it's prettier :p [12:11] really? [12:11] mine isnt comming up yet but thanks i will file bug against that [12:11] cool [12:12] asac: 3.7 in safemode ELAPSED 4822 [12:12] I had already noticed Firefox-3.5 had prettier fonts than Firefox-3.6 [12:12] way worse the Chromium [12:12] I'd assumed it was some flag which was only enabled in final builds or something [12:12] BUGabundo_remote: be sure chromium is closed before running it ;) [12:12] also dont use -profile [12:12] BUGabundo_remote, ppl want the speed while browsing, not the startup time, unless the difference is huge [12:12] firefox has probably system extensions so it takes ages longer [12:13] e.g. run with existing profiles [12:13] or ensure that no extension whatsoever is on your system [12:13] also no plugin [12:13] if you want to do a "fresh-profile" test [12:13] ok be righnt back i have to fix this. [12:14] asac: new profile : ELAPSED 885 [12:14] Tallken: I *do* want speed at start [12:14] I keep opening new windows, and closing browsers, on remote machines [12:14]