[00:09] fta: did you try to drop courgette? [00:09] fighting with it [00:13] committed another license checkpoint [00:15] hey guys, how do you name xulrunner betas and alphas? [00:16] s/name/version/? [00:16] multiple ~s [00:17] yeah, hm. [00:17] Pavlov: NEXTVERSION~b1 [00:17] e.g. 1.9.2~b1 [00:17] right [00:17] why? [00:17] fennec debian versioning hell [00:18] you want to wrok on feenec? [00:18] i am [00:18] or is fennec in debian? [00:18] for maemo, actually [00:18] cool [00:18] you put that in debian? [00:18] i think fennec needs 1.9.2 [00:18] it does [00:18] thats not in debian (and not yet in ubuntu, but soon) [00:18] no, sorry, i'm actually working _on_ fennec [00:18] you can use our dailies [00:19] great [00:19] for mozilla? [00:19] yeah [00:19] nice to meat you! [00:19] trying to get our debian package versions right [00:19] you too [00:19] learning all about ~s and such ;/ [00:19] the hard way! [00:19] we have a fennec package ;) [00:19] nice [00:19] not the latest. but i would suggest to help on that ;) [00:20] 1.0~a2... [00:20] or are you doing a package for your internal stuff? [00:20] we have a repo for maemo [00:20] https://code.edge.launchpad.net/~mozillateam/fennec/fennec.head [00:20] we should share the packaging imo [00:20] yeah we need to make a source deb at some point [00:20] right now they're built from our build system [00:21] if we get the branch up to speed we could easily include it as a daily build in our daily repo [00:21] for everything: hardy/intrepid/jaunty/karmic/lucid [00:22] https://edge.launchpad.net/~ubuntu-mozilla-daily/+archive/ppa [00:22] asac: we're trying to get our versioning stuff "right" [00:23] the problem right now is that we've been releasing xulrunner packages like: 1.9.2b6pre-datetime [00:23] now we're trying to release 1.9.2-datetime [00:23] and well, one is earlier than the other [00:23] argh [00:23] thats too late then [00:23] right [00:23] look at our xulrunner version [00:23] so i figure we need to bump the epoch [00:23] oh ... thats even worse [00:23] is it? [00:23] how many consumers do you have? [00:24] enough that i don't want to tell them to reinstall [00:24] at best you could wipe everything ;) [00:24] moving ahead of us is a mess :) [00:24] heh [00:24] well. then you need an epoch. [00:24] so we have a mozilla version to ubuntu version script [00:24] oh realy? [00:24] that is a perfect bi-dirctional mapping [00:25] ooooh [00:25] yes [00:25] * gavin likes the sound of that [00:25] that would be huge huge help [00:25] asac: almost perfect... [00:25] its in mozilla-devscripts ... let me check [00:25] micahg: alḿost? afaik its not lossy [00:25] prism... [00:25] well. prism doesn use that script ;) [00:26] its rather new for our automagic extension dependency stuff [00:26] asac: ah, right [00:26] so the script is perfect :) [00:26] * micahg needs to fix prism.. [00:26] asac: do you have al ink to the right place? [00:26] * asac looks it up [00:26] ;) [00:26] http://bazaar.launchpad.net/~mozillateam/mozilla-devscripts/mozilla-devscripts/annotate/head:/src/moz_version.py [00:27] its a lib [00:27] but we could make a command line tool out of it [00:27] just say what you want ;) [00:27] bdrung: ? [00:28] bdrung: moztodebian version ... thought we had a wrapper for that too ;) [00:28] http://bazaar.launchpad.net/~mozillateam/mozilla-devscripts/mozilla-devscripts/annotate/head:/src/moz-version [00:28] asac: yes. i added that. please check if it works correctly [00:29] bdrung: what Pavlov needs is moztodebian [00:29] where is that? [00:29] asac: i can write that easily [00:29] that would be awesome [00:29] that would be fantastic and you would get hugs from Pavlov ;) [00:29] many hugs [00:29] asac: how should i call this moz-version parameter? [00:30] gavin: wonder if we can just use these in the build system? [00:30] bdrung: either --to-moz .. --to-deb ... or make two high level wrappers [00:30] Pavlov: once they are done you could copy them in worst case [00:30] yeah [00:31] asac: high level wrappers? [00:31] soo ... epoch is bad because that would kind of kill the ecosystem ;) ... what we should do is do something ugly for 1.9.2 so you can go back to normal when 1.9.3 arrives [00:31] any suggestions? [00:31] bdrung: like moz-to-debian ... debian-to-moz [00:31] can we do 1.9.2+foo? [00:31] or something? [00:31] what was your current version? [00:32] yeah. so you can do CURRENT_VERSION+REALVERSION [00:32] asac: no, i put it into moz-version [00:32] so 1.9.2fuckedup+1.9.2~b2~hg20090920 [00:32] for the snapshot from 20th sep 2009 [00:32] pre beta2 [00:32] bdrung: if that works [00:33] --to-moz --to-deb [00:33] asac: the infrastructure is there [00:33] or whatever you think is better ;) [00:33] asac: is this part of the in package mozclient? [00:33] asac: our last real thing was xulrunner_1.9.2b6pre-20091231145857_armel.deb [00:33] asac: and how to call the one char parameter? [00:34] dpkg --compare-versions 1.9.2fuckedup+1.9.2~b2~hg20090920 lt 1.9.3~a1~xxx && echo yes [00:34] yes [00:34] so that works [00:34] Pavlov: yeah probably - DEB_VERSION = $(shell moztodebian MOZ_APP_VERSION) or whatever [00:34] Pavlov: so you build native packages ;)? [00:34] * asac wonders if that would be compared as if it was debian revision [00:34] * asac checks [00:35] the version is.. [00:35] dpkg --compare-versions 1.9.2b6pre-20091231145857+1.9.2~b6~20091231145857 lt 1.9.3~a1~xxx && echo yes [00:35] Version: 1.9.2b6pre-20091231145857 [00:35] asac: what do you think about "moz-version --to-deb 1.2.3 --epoch 2"? [00:36] if you want epoch ... but i would like to hide that feature ;) [00:36] asac: i can implement it but do not add a parameter until it is required [00:36] dpkg --compare-versions 1.9.2b6pre-200912311458571 lt 1.9.2+ && echo yes ? :) [00:36] yeah [00:36] oh ;) [00:37] yeah [00:37] so good [00:37] you can do that [00:37] 1.9.2+1.9.2~b6~20091231145857 [00:37] ;) [00:37] ugly [00:37] but better [00:37] well [00:38] this would be for 1.9.2 itself [00:38] no beta [00:38] can put a date if you think that is good [00:39] 1.9.2+~20100107133804 ? [00:39] let me check something [00:39] we might have a 1.9.2.x, or some other things, i guess [00:40] so yeah. until 1.9.2 is final, use that or whatever you want [00:40] after that, why dont you pocket copy our xulrunner packages daily? [00:40] 1.9.2 is final [00:40] is it? [00:40] or rc2 [00:40] its rc2 [00:40] er, rc1 [00:40] right [00:40] so its not yet;) [00:41] otherwise you could directly use our package versions [00:41] we have 1.9.2+nobinonly... ;) [00:41] but before final its 1.9.2~rc2... [00:41] well [00:41] the way our release process usually works is that we call the rcs the final version [00:42] at least internally [00:42] right. [00:42] we can call the debian version rc1 i guess [00:42] so ... rc1 is a bit special because it gets its own version right? [00:42] so for stable updates we do: [00:42] it doesn't [00:42] 1.9.2+build1 [00:42] 1.9.2+build2 [00:42] 1.9.2+build3 [00:42] right [00:42] so right now we have: [00:42] and whatever is last gets released [00:42] 1.9.2b6pre-20091231145857 [00:43] and 1.9.2-20100107133804 [00:43] yeah. so you could release as 1.9.2+build1 now [00:43] ok [00:43] if thats ok [00:43] yeah [00:43] and if you dont like +nobinonly (which we add for final builds if we spin one) [00:43] you can say: +final [00:43] right [00:43] asac: what if there are multiple rcs? [00:43] but if you are fine to just sticking with +build3 [00:43] then its fine [00:43] micahg: build1 is rc1 usually [00:43] build2 is rc2 [00:43] gavin: how do you feel about just adding a +rc1 to the version on the rc branch? [00:44] except for the prerelease rampup time [00:44] er, rel branch? [00:44] sounds good to me [00:44] Pavlov: be careful [00:44] rc > final [00:44] build1 < final [00:44] * Pavlov grumbles ;p [00:44] if you stick with rc3 then its fine [00:44] oh you can ro [00:44] we will [00:44] rc1 [00:44] rc2 [00:44] release [00:44] :-P [00:44] someone should have worked on this versioning stuff a bit more;p [00:44] in case ;) [00:45] yeah. but you are lucky it seems [00:45] we have to fix our 1.9.3 people [00:45] but no one is really on that channel [00:45] yeah. just wipe that [00:45] so we can fix it to use the script you guys have [00:45] hmm Fennec went from FENNEC_1_0_RELEASE to FENNEC_1_0rc2_RELEASE... [00:46] micahg: they usually add both tags [00:46] micahg, no, _RELEASE is fake, it's a moving tag [00:46] and then if it needs a respin move the _RELEASE tag [00:46] but i would have expected [00:46] ah [00:46] yeah each RC will have a tag, and the one we ship will also have RELEASE [00:46] FENNEC_1_0_BUILD1 ;) [00:46] tats what is used everywhere else [00:46] yeah, mobile is a bit different [00:47] but then: imo the pre-final RCs are different from the RCs for stable updates [00:47] I can ask our build guy about that [00:47] then FENNEC_1_0_BUILD2, then FENNEC_1_0rc2_BUILD1, then FENNEC_1_0rc2_BUILD2 [00:47] i really think that the pre release rcs are different [00:47] gavin: http://pastebin.mozilla.org/695534 ? :) [00:47] you even have a 3.0rc1 or something in application.ini [00:47] in RCs as in stable updates you have just moving tags [00:48] Pavlov: looks good to me [00:49] hope you do well ;) ... in the end i would still suggest that you sync our daily packages ;) [00:49] for 1.9.3 [00:49] etc. [00:49] well [00:49] we build our own packages [00:49] daily [00:49] for maemo [00:49] asac: what do you think about this: http://paste2.org/p/599114 [00:49] i would assume you guys aren't building daily maemo5 packages [00:50] we are building daily sources [00:50] ah [00:50] you just sync those and spin them [00:50] think about it ;) [00:50] still some time ahead [00:50] there are risks for dowstreams ... but also a big win ;) [00:51] but you dont have a build system for debs i guess? [00:51] Pavlov: how do you build fennec deb? are you spinning whole xulrunner? or are you just copying the build system and use --with-libxul-sdk? [00:52] bdrung: err. install .xpi file? e.g. without unpacking? [00:52] oh ;) [00:53] misread [00:53] asac: better explanations are welcome [00:53] bdrung: thats good. somewhat similar to the deprecated thing we talked about ... did that work out btw? [00:54] Pavlov: btw, are you X-compiling for arm? or do you have native builders? [00:54] asac: i removed the depracated variable, the packages still build and should work (e.g. there was thunderbird in the list and install-xpi creates a link for thunderbird) [00:54] asac: we're using scratchbox [00:55] bdrung: ok. [00:56] asac: i need this install-dir parameter for gears (it plays with the file after installation) [00:56] gears? doesnt that ship a proper .xpi? [00:56] why doesnt the auto detection work? [00:58] asac: it wan't some files in /usr/share and some in /usr/lib [00:58] asac: i assumes that the xpi is installed in /usr/share (but 0.18 installs it in /usr/lib -> ftbfs) [00:59] asac: we use --with-libxul-sdk and a two-stop build (i.e. MOZ_BUILD_PROJECTS) [00:59] two-step* [01:00] gavin: what i am wondering about is that we manually pack up the xulrunner build system and ship that as part of xulrunner-dev [01:00] and now that you probably need something like that maybe we can do something better upstream [01:00] builds xulrunner, then fennec, then fennec's build scripts package both into debs [01:00] asac: all xul extension build (some after patching) [01:00] so basically fennec build unpacks the build system for us at the beginning [01:01] gavin: yeah. so you still have the full source tree ;) [01:01] yes [01:01] the build system packup as part of xul sdk is quite nice. i would think it just needs to be standardized so all xulapps and extensions could use that [01:02] s/could/would/ [01:02] probably would require a single script or something that everyone wuld includ ein top level dir of its xulapp [01:02] this also would help stop xulapps from patching xulrunner ;) .. because they just dont have all the source at hand [01:02] asac: build system packup for extension? isn't xpi-pack enough? [01:02] and adding the full tree would make a slink build heavy wait [01:03] bdrung: native extensions [01:03] we already have that [01:03] http://mxr.mozilla.org/mobile-browser/source/installer/Makefile.in#115 [01:03] fennec just calls into the existing xulrunner packaging stuff [01:03] gavin: thats packaging at the end [01:03] i am talking about checking out the fennec tree [01:03] and then run a script to get the build system from the xulrunner sdk [01:04] or is that what you posted? [01:04] no [01:04] I misunderstood what you meant [01:04] if you currently read build instructions on how to build fennec or xulapps you always need the full tree [01:05] i want them to just use the build system [01:05] but not as a hard copy, rather on the fly [01:05] from the sdk [01:05] is that better worded ? [01:05] yeah I see what you mean [01:06] at best it wouldnt need to copy it but just include /usr/lib/xulrunner-devel/...sdk/build-system.mk [01:06] we don't really benefit from doing any of that work, though [01:06] and the build system isn't that self-contained [01:06] it helps the xulrunner eco system ;) [01:06] gavin: its not that bad anymore ;) [01:07] basically you can unpack the build system nowadays and ./configure --application=yourcoolapp [01:07] there are some rough edges [01:09] well yeah, that's what we do, it's just done manually :) [01:09] scp /usr/lib/xulrunner-devel-1.9.1.6/sdk/build-system.tar.gz rookery.canonical.com:public_html/tmp/ [01:09] -> http://people.canonical.com/~asac/tmp/build-system.tar.gz [01:09] right. i think it just needs some infrastructure and then evangilism so all the xulapps start to ship a full xulrunner copy [01:10] tar tzf build-system.tar.gz | pastebinit [01:10] http://pastebin.com/f5c4fdf59 [01:10] * gavin made the mistake of extracting that to ~/ [01:10] ouch [01:10] tar tzf | xargs rm ;) [01:11] take care :) [01:11] imo would need some magic default .mk that all xulapps could ship on top level [01:12] and then allow that whole thing to work even if its under build-system/ [01:12] (so it doesnt clutter the whole source tree) [01:13] hmm, did we sort out the version script thing? [01:13] too many version numbers melt my brain :( [01:13] asac: btw, m-d 0.19 is faster than v0.18 [01:14] Pavlov: from what i understood you are going to use 1.9.2+rc1 ;) [01:15] oh, we are [01:15] for today;) [01:15] i am not yet sure how you do the dailies afterwards then :) [01:15] hehe [01:15] so i would still suggest to go for OLDVERSION+NEWVERSION [01:15] and then just adapt whatever great new versioning scheme we decide on [01:15] for NEWVERSION [01:15] asac, all fine without courgette now. at last [01:15] and from 1.9.3 on just use OLDVERSION [01:16] 1.1~a1~datetime seems ok [01:16] or for xulrunner, 1.9.3~a1~datetime? [01:16] our script will produce a good version for the thing that is in application.ini [01:16] Pavlov: yes. thats basically it. [01:16] how does it deal with releases and other stuff [01:16] or do you set some env var to add +foo to things? [01:17] i would suggest this: [01:17] if you have version 1.9a1pre ... the script would cut of the pre and remember that its a daily [01:17] yep [01:17] the you run the version through the moz-to-version thing we produce [01:17] and if its remembered daily, append ~DATETIME [01:18] we use [01:18] hgDATEtTIME [01:18] ;) [01:18] 1.9.3~a1~hg20091227r36701+nobinonly- [01:18] oh now we use r [01:18] rather than time [01:18] thats good [01:18] i think you should adopt that ;) [01:18] heh [01:18] i think the problem is that hg doesnt always move ahead [01:18] right [01:18] thats why we wnet to r rather than t [01:19] well, those are local [01:19] asac: do the xulapps check a directory in /usr/share for extensions? [01:19] not necessarily global [01:19] Pavlov: its better to use a reproducible version [01:19] so we use versions that we know we can recreate the source for [01:19] yeah those are hex [01:19] heh [01:19] if you use build system local time then you cannot do that [01:20] problem is that you have xulrunner also checked out [01:20] so you either build xulrunner independently (like we do) [01:20] or you declare it a DEP [01:20] and have a DEP file with the current revision in fennec or something [01:21] so just reproducing the same fennec tree will allow you to also reproduce xulrunner [01:21] makes sense ? ;) [01:21] (its late) [01:22] ok now you distracted me from what i wanted to do this night initially ;) [01:23] fta: i think we are close: [01:23] http://pastebin.com/f48ffd9fb [01:23] unless i messed up something completely (not that anyone would spot that in this 1Million line copyright file ;) [01:24] hmm "Public Domain" isnt whitelisted it seems [01:25] maybe that should be case insensitive ;) [01:25] fixed [01:25] fta: waiting for a fresh stripped tarball ... to see if stripped the same stuff you stripped now [01:28] http://people.canonical.com/~asac/tmp/copyright.full [01:33] i stripped everything you asked except sdch/open-vcdiff [01:34] yeah. those are still in the probs [01:36] -rw-r--r-- 1 fta fta 87406668 2010-01-07 04:10 chromium-browser_4.0.292.0~svn20100107r35689.orig.tar.gz [01:36] -rw-r--r-- 1 fta fta 80988593 2010-01-08 01:25 chromium-browser_4.0.293.0~svn20100108r35757.orig.tar.gz [01:36] -7% [01:37] cool ;) [01:37] hope it was not 07 to 08 ;) [01:37] but our stripping [01:41] why do you have build-tree/src/base/third_party/purify/* twice with different Copyright? [01:44] asac: hey, lets say we pulled xulrunner in to our fennec package [01:44] what would be the best way to deprecate the xulrunner one? [01:48] Pavlov: you want to deprecate xulrunner? [01:48] heh [01:48] ok [01:49] do you want all folks that just have your xulrunner to automatically get fennec? [01:49] if so, you add transitional packages to fennec [01:49] but ... are you sure maemo doesnt have xulrunner? [01:55] yes [01:55] it has xulrunner, but not installed like xulrunner [02:01] _Tsk_: is there anything we can do about mozilla 534651 [02:01] Mozilla bug 534651 in Build Config "make install should install the SDK for Thunderbird 3" [Normal,New] http://bugzilla.mozilla.org/show_bug.cgi?id=534651 [02:02] 03:00 < StevenK> asac: gyp looks okay, aside from the OMG-my-eyes-are-bleeding debian/rules. [02:02] fta: ^^ [02:02] so ... go ahead [02:02] finish this stuff [02:03] then he will cry in front of chromium ;) [02:03] fta: already prepared him ;) [02:03] feeding him copyright.full and the pastebin [02:03] fta: can you upload that now or do you want him to upload [02:04] what? gyp to lucid? [02:04] he would do the changelog bump to -0ubuntu1 [02:04] yes [02:04] what version do you have? [02:04] 0.1~svn770-0ubuntu1 [02:04] right [02:04] upload that with the bug closed [02:04] ok [02:07] what is in debian/rules? [02:07] he said he didnt like it ;) [02:07] done [02:07] oh get-orig-source [02:07] 4 lines of python, + get orig [02:08] ok now the usual its past 22pm reminder ;) [02:08] * asac off [02:08] off too [02:08] ++ [02:08] did you do everything right or do i need to wait [02:08] fta: ? [02:08] hm [02:08] e.g. ensure that its in the archive? [02:08] time for one more dumb question? [02:08] its 3am ;) [02:09] "Successfully uploaded packages." [02:09] fta: well. sometimes it gets rejected because of something bad ... especially at 3am ;) [02:09] if we did merge fennec and xulrunner, we'd need some kind of version that has both things in them :/ [02:09] bogus checksum etc. [02:09] oh hm thats not true [02:09] nevermind [02:09] get some sleep [02:09] :) [02:09] Pavlov: yes. i already asked above: [02:09] 02:48 < asac> heh [02:09] 02:48 < asac> ok [02:09] 02:49 < asac> do you want all folks that just have your xulrunner to automatically get fennec? [02:09] 02:49 < asac> if so, you add transitional packages to fennec [02:10] 02:49 < asac> but ... are you sure maemo doesnt have xulrunner? [02:10] but lets talk tomorrow ;) [02:10] ok [02:10] thanks again for your help [02:10] asac, got the NEW email [02:11] good [02:11] fta: good night!! [02:11] asac, thx, you too === _Tsk__ is now known as _Tsk_ [08:44] <_Tsk__> micahg: how do you feel about making a patch ? [08:44] _Tsk__: if I knew what to patch, I would be happy to try :) [08:45] <_Tsk__> can you come to irc.mozilla.org in #maildev and ask that specific question to standard8 [08:46] _Tsk__: k, I'll have to do it a little later though, it's too late for me right now to concentrate on that (almost 3AM) [08:46] thanks _Tsk__ [08:46] <_Tsk__> elcome === _Tsk__ is now known as _Tsk_ [09:11] _Tsk_: one last Q, you know what tz Standard8 is in? [09:12] <_Tsk_> BST [09:13] ugh, so I better ask now I guess, otherwise when I get up, he'll be gone for the weekend [09:16] <_Tsk_> he stays late - but yes that would be wise === _Tsk__ is now known as _Tsk_ [09:45] BST ... its 9:45 for BST (if thtas british) [09:45] ah ... guess its late for micahg ;) [09:46] asac: yes :) [09:46] wanted to catch standard 8 [09:46] making a patch? [09:46] whats that about? [09:46] fixing tbird? [09:46] and try to figure out this devel thing for tb3 [09:46] yeah [09:46] glorious ;) [09:46] what bug? [09:46] ah -devel packages [09:46] cool [09:47] asac: wanna hop in the mail-dev channel? [09:47] i am not in there? [09:47] now i am ;) [10:05] asac: if I tell debuild not to purge, if it completes, can I still see the full dir structure? [10:06] debuild -nc is fine [10:06] you might need to remove the build-stamp [10:06] otherwise it skips it and doesnt see that you changed something [10:06] or the install-stamp if you want to just test make install [10:06] or nothing if you just want to test the debian packaging install stuff [10:07] so yes. its highly recommended to develop dewbian packaging by running debuild -nc rather than a full builds [10:07] saves you hours ;) [10:07] or even days [10:08] just in the end remember to do a full spin to verify ... and remember to copy stuff to your bzr tree before it gets purged next time by accident [10:08] that's the thing you gave me before that I couldn't remember [10:09] ok, I'll try this sat night [10:09] ccheney: heya [10:09] ccheney: status update? [10:09] I have to get some sleep now and probably won't get to it when I get up tomorrow [10:09] short day [10:30] crimsun: hey. old nick ;)? [12:01] morning [12:06] asac, http://code.google.com/p/chromium/issues/detail?id=308103 [12:06] asac, http://code.google.com/p/chromium/issues/detail?id=30810 [12:06] BUGabundo_work, lo [12:08] create symlinks? [12:08] is that done on first run? [12:08] thats messy [12:09] that's their wrapper, i don't use it [12:09] i have mine [12:11] why does redhat have the wrapper in their svn? [12:11] does redhat create symlinks in the system lib dir? [12:11] or is that in profile? [12:12] no, upstream is doing a deb and a rpm [12:12] i don't think fedora ships this wrapper either [12:12] i will ask [12:15] so what does it do? need root access? [12:21] seems like it :P [12:22] that should get removed imo [12:22] something like that is not really something ever to be done again ;) [12:23] someone needs to see that and stop it ;) [12:23] they don't use sudo or anything so it most likely doesn't work at all [12:23] hmm. i think they ship it like the firefox tar.gz [12:23] so you can unpack in your home [12:24] but not when used through the deb or rpm [12:35] it's a bad idea but i don't see what else they could do [12:35] they have only 1 binary for a bunch of distros [12:54] yeah [12:54] ship more ;) [12:54] in-sorce [13:14] asac, i can't find an obvious culprit for the chromium regression. i hope it's not one of our changes from yesterday.. [13:22] hello fta [13:22] chromium unbroken ? [13:22] nope [13:23] :\ [13:23] then i'll pin Ch down ! [13:23] http://code.google.com/p/chromium/issues/detail?id=31809 [13:23] star it [13:23] but don't add a me too [13:25] fta: i'm not new to BTS :o [13:25] well glad i have my old debs locally [13:26] i'll downgrade to one of those if need be [14:20] fta: thoought there is a bug upstream [14:21] are all those using our packages? [14:21] fta: you can spin chromium with tarball from two days ago to test [14:21] if its our thing [14:21] e.g. run the cleanup stuff etc. on the old tarball [14:22] it's all from the daily ppa, but it's the only widely used daily so i'm not sure [14:23] could also be the gl build-deps [14:24] did you try it? (you don't use the nvidia driver iirc) [14:24] asac, ^^ [14:25] bla [14:25] bug 504149 [14:25] Launchpad bug 504149 in xorg "[lucid] after update keyboard and mouse do not work in X" [Critical,Invalid] https://launchpad.net/bugs/504149 [14:53] asac, too many changes. i've created two up-to-date tarballs, one with the get-orig-source from 2 days ago, and one with the new rule [14:53] but there's also the gl build-deps [14:54] not sure which one try 1st [14:54] i need the gl deps to build [14:56] try the last yo uknow that worked [14:56] run the removal and see if it works [14:56] if so we can rule out its the stripping [15:09] asac, http://code.google.com/p/gyp/issues/detail?id=133 => fixed [15:17] asac: was working on openoffice yesterday and ran into some sort of build breakage due to what appears to be lucid :-\ [15:18] fta: cool [15:21] asac, i didn't check, is it enough? [15:22] will check after release meeting [15:22] have to do the report now :( [15:22] ccheney: pleaes dont upload ooo before a2 [15:22] you will certainly bust our arm images with that [15:23] ccheney: ooo is painful, so consider to continue 30% of day on the backporting ;) [15:23] much more fun [15:24] asac: slangasek asked for OOo upload specifically because images are already oversized [15:24] alternate cd for i386 or amd64 iirc [15:26] asac, *sigh* http://paste.ubuntu.com/353513/ [15:26] please provide context [15:26] ;) [15:26] what tarball etc. [15:26] asac: hi.. [mac_v here] reminding Bug #386900 :) could you target it to a milestone which would ensure it is fixed for Lucid? [15:27] Launchpad bug 386900 in network-manager ""Auto eth0" in notifications is confusing" [Wishlist,Confirmed] https://launchpad.net/bugs/386900 [15:28] asac: i just heard how to hopefully fix the OOo issue so i will be rebuilding 3.1.1 instead of 3.2.0 if it works and uploading it so hopefully it won't hurt arm, plus there is another arm patch to include from doko. [15:31] does that fix the oversize issues? [15:31] hopefully is really not enough [15:31] if you upload that and it breaks it means no image for a2 for arm [15:31] i ping slangasek. lets see [15:32] bug * Bug:431963: linux-fsl-imx51: io/fs errors when launching gdm on imx51 with sata [15:33] bug 431963 [15:33] Launchpad bug 431963 in linux-fsl-imx51 "io/fs errors when launching gdm on imx51 with sata" [High,Fix committed] https://launchpad.net/bugs/431963 [15:33] * vish wonders if asac noticed previous message ... [15:37] asac, x64 host, 32bit builder, v8 is confused. seems it's a regression [15:41] asac: yea it fixes the issue due to not needing old icu library anymore from what slangasek told me [15:41] asac: the old library is currently using 7MB on the disk from what he said [16:22] asac, fresh rev without yesterday's stripping, NOK [16:22] good [16:22] well, sort of :) [16:22] if its the same issue we could rule out the stripping, but trying with a good checkpoint would be better to verify === stevel_ is now known as stevel [16:46] bdrung: the dailies broke after yesterday's updates [16:47] micahg: dailies of what? [16:47] bdrung: xul192, prism, tb3 [16:48] I think it's based on the new m-dev in lucid [16:48] hmm [16:48] nm [16:48] it must be something else [16:49] micahg: do you updated the m-d in the ppa? [16:49] but prism I know is due to some of the changes [16:49] prism is due to its own changes [16:49] the other 2 I'm not sure [16:49] micahg: do you have a link to the ppa? [16:49] since it was only on lucid [16:49] https://edge.launchpad.net/~ubuntu-mozilla-daily/+archive/ppa/+packages [16:50] hmm [16:50] maybe there isn't an issue afterall [16:50] * micahg was tired when I discovered it [16:51] probably just need to wait for tomorrow so there's an updated m-dev [16:51] bdrung: don't worry about it [16:53] bdrung: can we update the packages to require a certain version of m-dev if that's the case? === yofel_ is now known as yofel [16:55] micahg: the build failure thunderbird-3.1 can be due to m-d, but the other package have probably other reasons. in m-d i have only changes the extension part (xpi.mk and such tools) [16:55] bdrung: I think TB31 is upstream [16:55] k [16:55] * micahg already has to fix that [16:56] they updated their configure file and it doesn't like our changing the binary name [16:57] what about this in prism: with open("debian/control") as f: ^ SyntaxError: invalid syntax [16:57] hardy and intrepid use python 2.5 [16:58] ok ... so i promissed to do something after releas meeting ;) ... what was that? [16:59] asac: do we have to make everything in dailies backport properly? [16:59] prism is failing on some python code that's valid in 2.6 but not 2.5 [17:01] hmm. [17:02] what python code is that? [17:02] with open("debian/control") as f: ^ SyntaxError: invalid syntax [17:02] its probably again itchy developers jumping on latest api crap without rason ;) [17:02] it's something in m-dev [17:02] micahg: can you check if that is fixable? if you dont figure i can check, but maybe a good excersize for python :-P [17:03] m-dev? [17:03] mozilla-devscripts [17:03] in our package? [17:03] then tell that bdrung [17:03] bdrung: dont use python 2.6 in m-dev [17:03] we need backports [17:03] 18:02 < asac> its probably again itchy developers jumping on latest api crap without rason ;) [17:03] please [17:03] ;) [17:03] thx [17:04] i am running around evangelising folks to not use latest API stuff to improve the linux ecosystem ... so mozillateam should try to be good example ;) [17:05] great that we spotted that [17:05] thx fta for pushing devscripts to daily [17:06] it's supposed to be automatic [17:06] it didn't work? [17:06] branch nick: mozilla-devscripts.daily [17:06] timestamp: Fri 2010-01-08 05:25:25 +0100 [17:06] message: [17:06] * Merge with mozilla-devscripts #302 [17:07] fta: it worked ;) [17:07] read a bit back [17:07] like last 25 lines [17:07] asac, oh, it was not a request, lol. nm then [17:07] it just revealed a bug in python code [17:07] fta: it was a kudo [17:08] :) [17:08] if i say thx i _always_ mean it that wy [17:08] way [17:08] would never be that impolite/sarcastic [17:36] asac: sorry, i tried to avoid python 2.6 (but that slipped through) [17:37] no problem ;) [17:37] i should have thought about that and ask to test it [17:38] guess its a one minute fix for you ;) [17:39] asac: yes, will fix it [17:39] asac: after finishing moz-version [17:39] thx [17:39] i think next daily run is 4am UTC [17:39] so getting it fixed by then might allow to verify on sat or sun [17:40] not sure if md goes up before prism [17:40] that's no problem [17:40] it looks like it goes up first [17:45] micahg: sure it goes up first? or finishes first? [17:45] ah [17:45] good question [17:45] * micahg checks the logs from the bot [17:45] the latter is expected but would mean prism doesnt catch it ;) [17:46] it should be > 1h difference on i386 etc. [17:48] uploaded first [17:48] good [17:48] prism is uploaded last [17:55] fta: what bug was that ld thing? [17:55] 18:51 < fujimitsu> Inconsistency detected by ld.so: dl-minimal.c: 138: realloc: Assertion `ptr == alloc_last_block' failed! [17:57] got it [17:57] fta: unping [18:02] asac: what are your thoughts on the firefox kde integration? is there a chance it will go into lucid? anything I can do to help? [18:03] we need upstream bugs with indidivual patches [18:03] otherwise we wont get approval ... i am pretty sure [18:03] so if you have an individual bug list we can review that and then decide on a case by case base [18:04] to get some credits we need to shepherd the stuff into trunk [18:04] at least actively working on that. [18:05] just pulling in incomplete patches without helping out wont give us approval i am sure [18:08] asac: pushed moz-version. please test it thoroughly (moz-version --compare VS dpkg --compare-version) [18:11] no test cases? ;) [18:12] didnt we do that before? [18:12] i thought we just add the debtomoz and moztodeb mapping feature [18:12] i really thought we had a complete test for the --compare at some point [18:12] where did that go? [18:14] asac: i have a local test for --compare. but i want a test for the new parameter. [18:15] asac: take two moz version, compare them, then convert them into deb version and compare them again [18:16] asac: 2. the other way around [18:18] hmm. [18:18] cant we make a direct test case for tomoz and todeb first [18:18] that indirect test case looks reasonable to cover more [18:18] but definitly isnt the thing to start ;) [18:18] asac: here my local script http://paste2.org/p/600229 [18:18] how do you interpret *? [18:18] can you commit that to mozilla-devscripts ? [18:19] like in tests/ and tests/data [18:19] or something [18:19] asac: you can to moz-version -> deb version -> moz-version and compare both moz versions [18:19] feels like the right thing to do [18:19] * = int.max [18:19] heh [18:19] ok [18:19] i would think we should error [18:19] rather than that [18:19] but ok [18:19] error? [18:20] most likely happens becaues of the underlying code [18:20] yes. for compare there is a reasonable meaning [18:20] for todeb there is none [18:21] at least a warning to stderr ;) [18:21] or spit out the result, but give error exit code [18:22] but thanks so far ;) ... lets commit the current test anad we can check ;) [18:22] k [18:24] asac: how can i import the module from another dir? import ../src/moz_version does not work [18:51] asac: am i allowed to apply this patch? http://paste2.org/p/600264 === vish is now known as mac_v === mac_v is now known as vish === asac_ is now known as asac [19:24] asac, http://codereview.chromium.org/524075 \o/ [19:30] asac: we seem to have a problem with the upstream fennec xul deb builds being pulled in by our FF app on moblin [19:31] * micahg guesses he should see what it installs where... [19:35] * micahg can't figure out what's causing it to conflict... [19:35] hmm [19:36] maybe it's a different build [19:55] hi moziall team ;) [19:55] what is the difference between xpt and xpi [19:56] esp. when looking into creating packages [19:57] my understanding is: xpi is an extension, and xpt is a component, but my understanding is also that extensions can contain components === _Tsk__ is now known as _Tsk_ [20:10] asac: ^ === _Tsk__ is now known as _Tsk_ === _Tsk__ is now known as _Tsk_ === _Tsk__ is now known as _Tsk_ [21:54] asac: done [21:54] asac: it's now ready for release. objections? [21:55] <[reed]> thekorn: you are correct === _Tsk___ is now known as _Tsk_ === _Tsk__ is now known as _Tsk_ [22:58] asac: icedove ignores the extension in /usr/lib/mozilla/extensions/{...}/. why? [23:19] http://ffextensionguru.wordpress.com/2009/12/28/hide-menu-bar-in-firefox-3-6/ [23:19] will this be in ubuntu too? [23:19] i guess it's currently just in windows