=== asac_ is now known as asac [00:22] chromium is messing iup with zlib and libpng symbols, like mozilla, i can't use the system ones, at least not without further tweaks [00:27] Ugh. I still have no idea why KompoZer is so horribly broken in 8.10 compared to previous releases. [00:31] isn't KompoZer dead ? [00:33] well, I haven't heard much from the dev in a long time, so perhaps, but until 9.10 it's the best we have. [00:34] fta: isnt that just a define? [00:34] zlib/png [00:34] yes [00:34] ok [00:35] thats at least easy to switch then in theory [00:35] zlib: symbols are remapped to MOZ_Z_xx for libpng [00:35] libpng: problem with undefined references to lots of webkit_png_* symbols [00:35] in chrome? [00:35] yes [00:35] ok so they copied the mozilla code ;) [00:36] yes, parts of it [00:36] shouldnt that be solved in webkit package already? [00:37] or is png shipped in chrome itself and not by webkit? [00:37] both [00:37] cool ;) [00:37] third_party/zlib [00:37] third_party/WebKit/WebCore/platform/image-decoders/zlib [00:37] ok [00:38] in theory they ship their own widget thing right? [00:38] fta: did you sponsor gnomefreaks extensions yet? [00:38] just remember that he said a had a bunch ready ;O) [00:38] hm, no, i don't remember him asking me anything [00:39] http://paste.ubuntu.com/70237/ [00:40] your patch=? [00:40] yep [00:40] submit it ;) [00:41] i will, it's not complete yet [00:41] kk [00:42] well, maybe it is, but my build is still in progress [00:50] bug 28479 [00:50] Launchpad bug 28479 in firefox "firefox32 for amd64 => support 32-bit plugins on amd64" [Wishlist,Won't fix] https://launchpad.net/bugs/28479 [00:50] they want a firefox32 [00:55] asac, my notes so far: https://wiki.ubuntu.com/MozillaTeam/Chromium [00:55] not much [01:02] asac: where's the intrepid patches list? [01:02] mconnor, what for ? [01:02] http://bazaar.launchpad.net/~mozillateam/firefox/firefox-3.0.intrepid/changes is not easy to follow [01:03] fta: you know, so I can yell at you guys again [01:03] or something [01:03] remember, I hate freedom and everything [01:04] http://patches.ubuntu.com/by-release/extracted/ubuntu/x/xulrunner-1.9/1.9.0.3+nobinonly-0ubuntu1/ [01:04] http://patches.ubuntu.com/by-release/extracted/ubuntu/f/firefox-3.0/3.0.3+nobinonly-0ubuntu2/ [01:05] 1.9.1/3.1 are now in there yet, you have to use the branches [01:05] -now+not [01:05] s'ok, I just need to have a baseline [01:05] fta: thanks [01:06] mconnor, all those named bzXXX_ are supposed to be upstreamed [01:06] one day [01:07] uh [01:07] drop_bz418016.patch [01:07] you have jemalloc disabled? [01:08] hm, i don't think so [01:08] that patch kinda makes it look that way [01:09] anyway, I should do this after I finish reviewing blockers [01:09] i have it in 3.1 for sure, i'm spamed with zillions of /tmp/jemalloc.XXXX files [01:09] heh [01:10] $ lsof -p 9989 | grep -c /tmp/jemalloc [01:10] 283 [01:11] deleted but still opened so this is vicious [01:11] uh [01:11] weird [01:11] never heard of that [01:12] firefox-3 9989 fta DEL REG 8,1 35160076 /tmp/jemalloc.EfHUUh [01:12] firefox-3 9989 fta DEL REG 8,1 35160073 /tmp/jemalloc.32yXGC [01:12] firefox-3 9989 fta DEL REG 8,1 35160071 /tmp/jemalloc.sWt2vX [01:13] maybe a leaked filehandle or something [01:13] hmm [01:13] look for an upstream bug? [01:13] I doubt its just you [01:13] also, http://patches.ubuntu.com/by-release/extracted/ubuntu/f/firefox-3.0/3.0.3+nobinonly-0ubuntu2/ubuntu_no_app_updates.patch [01:13] why not just set the pref and be done with it? [01:14] and that, uh, doesn't change the pref? [01:15] not my patch [01:17] man... [01:17] this is not going to go quickly, is it? [01:18] ok, I'll do it tomorrow [01:18] where I can yell at asac directly [01:18] ok [01:18] (this is the overdue approval round for branding stuff, its so much cleaner than Fx2!) [01:19] mconnor: i am on holiday ;) [01:19] mconnor: the drop patch isnt applied. you have to look at the series file [01:20] asac: where is the series file? [01:20] mconnor: the pref isnt for app updates isnt enough ... users must not enable app updates [01:20] asac: so lock the pref [01:20] that should work just fine [01:20] mconnor: lock prefs is something that admins have to use [01:20] we used that in the past [01:20] no [01:21] why do you think that's true? [01:21] because you can only have one lock file? [01:21] er [01:21] sec [01:22] i think we have developed a patch that allows you to do locks in all pref files ... but the app updates patch probably predates that [01:23] pristine behaviour is that you can configure one general.config.filename ... and in that file you can have lockPref [01:23] mconnor: the series file in only in the branch or package source ... in debian/patches directory. [01:24] there also are all the patches so you dont need to switch to the patches.ubuntu.com site and just can navigate there. [01:24] link me please? [01:25] mconnor: http://bazaar.launchpad.net/~mozillateam/firefox/firefox-3.0.intrepid/files [01:25] there you can navigate the tree [01:25] go to debian/patches/ [01:27] asac: hmm [01:27] what's the branding stuff? [01:30] mconnor: we provide that for downstreams and oems that need a different branding (on a packaging level) [01:33] mmm, so you can package up another branding jar without having to rebuild the package, I assume? [01:33] mconnor: right ... so what we do is have firefox-3.0-branding package ... if you install that you get all firefox ... if you install abrowser-3.0-branding ... everything uses the example "alternative" branding [01:34] hmm [01:34] intriguing [01:35] mconnor: its ment to be the answer to the critique: [01:35] "mozilla provides a simple configure switch to use your own branding ... what do you do for packages?" [01:36] but we are still learning on this, so its definitly not perfect yet. [01:36] asac: yeah [01:36] mconnor, here is the content of the firefox-3.0-branding package: http://paste.ubuntu.com/70260/ [01:36] but we are searching for better way then just providing iceweasel or something [01:36] I mean, I think there's a nicer way to do that [01:36] i appreciate any feedback [01:36] fta: I can entirely predict that [01:37] which is a good thing i assume :) [01:37] I'm the one who added the /unofficial directory so people could see how easy it was to have alternate branding :) [01:37] we had to split browser.jar though [01:38] that should be "upstreamable" :) [01:39] welll ... [01:39] the split should not hurt [01:39] yeah [01:39] just like i said, we are still somewhat learning how to do this best for all the eventual cases [01:40] that sub jar and the icons could also be regrouped in a branding directory, that would help too [01:40] but keeping the branding in a separate chrome jar is probably part of the final solution ... though not 100% sure. maybe it can be better supported in a different way [01:41] hmm [01:41] I thought we kept stuff in a separate jar already [01:41] fair [01:41] mconnor: nope ... browser.jar [01:41] and some stuff directly in APP/GRE dir [01:42] huh [01:42] man [01:42] all of the things I don't think enough about :-/ [01:42] hehe [01:43] well. its understood that alternate branding is probably not a core feature ;) [01:43] on a product level, no [01:43] on a code level, it kinda is [01:44] it has been for a very long time, really [01:44] mconnor: there are also other minor issues, like the user-agent string [01:44] mconnor: some might want to use their own product in that id [01:44] mconnor: but most probably want to keep firefox in order to be not locked out of some sites [01:44] that should just work, aiui [01:44] well [01:44] mconnor: i think it encodes the appname [01:44] during build [01:44] yes [01:44] I think I remember where, even [01:45] but [01:45] we're trying to teach the world to sniff Gecko versions [01:45] yeah ;) [01:45] but the world is not enough :) [01:45] yeesh [01:46] mconnor: how about a campaign: "randomize your user agent parts" ;) ... and then complain to websites [01:46] hard to guess how to best teach MS to make hotmail work [01:47] they even stop providing full service if ubuntu is properly encoded [01:48] that's just them being dicks ;) [01:49] i wouldnt be shocked if they looked through the navigator.plugins array and downgrade their service when they find an openoffice doc plugin installed ;) [05:58] i really hate "at" [06:23] ok firegpg is set for intrepid in my PPA i am gonna rebuild all of them with target jaunty tonight or tomorrow === fabrice_sp_ is now known as fabrice_sp [06:58] fta: support for Jaunty isnt in PPAs yet, they say they uploaded but PPAs are not showing them anywhere. [10:41] gnomefreak: ok thanks [10:41] gnomefreak: if there are things to upload and i am not here, ask fta too ;) [10:47] asac: gentoo bug 234110 [10:47] Bug 234110 on http://launchpad.net/bugs/234110 is private [10:47] any clue why it would hang? [10:47] bah [10:47] https://bugs.gentoo.org/show_bug.cgi?id=234110 [10:47] bugs.gentoo.org bug 234110 in Applications "www-client/mozilla-firefox-3.0* USE=xulrunner - hangs on first start" [Normal,New] [10:52] hmm gentoo still has no cert ;)? [10:52] it has a cacert one [10:53] ok [10:53] brb, kernel upgrade, have a look and let me know :) [12:22] kernel upgrades appear to take ages on gentoo ;) [12:22] gtk [12:47] tell armin that i don't know ... only thing i could imagine is either cruft in the pkglibdir, a missing .autoreg or i dont know ;) [12:52] ok off for a while again ;) [12:58] asac: gentoo .. :P [12:58] asac: hf [13:13] dh_strip debug symbol extraction: ignored packages: [13:13] objcopy:./usr/lib/chrome/stiAjXMz[.debug_loc]: No space left on device [13:13] :( [17:33] <[reed]> sheesh, travel plans are going to be the death of me [17:33] ? [17:40] <[reed]> fta: finalizing UDS travel and such, as I finally got approval from one of my professors to move an exam :( [17:41] [reed], i need to finish mine too. i totally forgot [17:41] <[reed]> lol [17:43] a matter of direct flight more expensive than alternatives through 2 or more hops [17:52] <[reed]> yeah, the proposed flight for me is pretty hoppy, too [21:29] fta: 2 hops or two connections? [21:30] a-b-c is ok, a-b-c-d is pushing your luck :) [21:31] <[reed]> I'm a-b-c-d [21:31] <[reed]> where 'a' is my home airport, and 'd' is SFO [21:32] i have a-b, a-b-c and a-b-c-d, [21:32] a = paris [21:34] i have the choice for the connection(s), Toronto, Chicago/Frankfurt, Newark, Cinncinnati, Philadelphia, Detroit, ... [21:43] http://www.sofaraway.org/ubuntu/tmp/src.png [21:50] most are tests [21:51] [reed]: your home airport is weird though [21:51] <[reed]> yeah [21:51] <[reed]> mconnor: are you going to come to any part of UDS? [21:51] [reed]: tbd [21:52] depends on how today goes [21:52] <[reed]> what's special about today? [21:52] because today I'm looking at the intrepid patchset [21:52] <[reed]> ah [21:52] and if I find something that causes problems, I'm sure I'll get lynched [21:52] ;) [21:53] <[reed]> hehe [21:54] but intrepid is already out.. so why now? [21:55] because we're actually going to have a formal trademark usage agreement, rather than just ad-hoc stuff [21:55] oh [21:56] also, because we've always required that, even without contracts, as a prereq [21:56] that we haven't done it yet kinda sucks [21:56] just because the horse has left the barn doesn't mean you can't put the horse back in the barn, you just can't close the door and do any good [22:34] * mconnor grumbles a little [22:34] ? [22:35] seeing stuff in a patches list for a bug that was resolved invalid a year ago is slightly frustrating [22:35] especially when the patch was rejected [22:35] which one ? [22:35] sec [22:35] bz386904_config_rules_install_dist_files.patch [22:35] see https://bugzilla.mozilla.org/show_bug.cgi?id=386904 [22:36] Mozilla bug 386904 in Build Config "DIST_FILES and DIST_CHROME_FILES not implemented for install:: target in config/rules.mk" [Normal,Resolved: invalid] [22:36] the patch is wrong, and no longer needed in any case [22:36] bad asac, no donut [22:36] <[reed]> are you sure it's being used? there are patches in the repo that aren't used [22:36] <[reed]> also, make sure to look at both firefox and xulrunner patches [22:37] <[reed]> two different patch sets [22:37] its in the series file [22:37] <[reed]> ok [22:37] also, seriously, are you telling me how to do my job? :P [22:37] this isn't my first time at this here rodeo [22:38] and I haven't looked at the xulrunner patches because, for the most part, I don't care [22:38] <[reed]> no, just offering helpful suggestions in case you didn't know [22:39] [reed]: see, that's what you _think_ you're doing... :P [22:39] hm, we have that patch since trunk 20070703 [22:42] that was before the upstream bug got filed [22:42] and bsmedberg rejected the patch two weeks later [22:42] <[reed]> mconnor: just checking actual patches or ubufox too? [22:42] and it got left anyway [22:42] [reed]: ubufox is tomorrow [22:42] yay me [22:42] <[reed]> k [22:43] you're doing a bad job of being on sabbatical, btw [22:43] mconnor, i think asac just forgot about it, no evil intention here [22:43] fta: I'm not saying evil [22:43] I'm saying "this is frustrating" === jdstrand_ is now known as jdstrand [22:44] because he would have gotten mail explaining why that patch was wrong, and he did nothing about it [22:45] fta: btw, where are the build configs stored? [22:45] in debian/rules [22:47] ow [22:47] that's so... elaborate [22:47] :) [22:48] why do you guys specify so many build flags explicitly? [22:48] also, why do you disable reporter? [22:49] oh, you ship it with xulrunner [22:49] oy, ok [22:49] because we have apport pushing core dumps to launchpad, where they are retraced by a bot and published in a bug [22:51] not crashreporter [22:51] though, really, I still wish you wouldn't do that [22:51] but that's a separate discussion about how to not have linux distros hiding their crash data from us [22:51] what? crashreporter ? [22:52] <[reed]> I don't think they're hiding it... just think that we need to integrate better [22:52] we had several discussions about that [22:52] including at the last uds [22:52] hiding is pejorative [22:53] I mean that we don't have a unified view of crashes [22:53] because every distro does something different [22:53] and we can't roll that up easily [22:57] I also wish you guys ran unit tests [23:01] it's not something our builders can do. we need dedicated boxes. [23:01] i'm not canonical, but i think asac discussed that internally with the ubuntu QA team [23:02] unfortunately, nothing happened since [23:05] you have an hppa build? [23:07] yes, but not officially supported [23:07] i.e. best effort [23:07] ok [23:08] there's a newer version of the patch in the bug [23:08] https://bugzilla.mozilla.org/show_bug.cgi?id=436133 [23:08] Mozilla bug 436133 in Networking: Cookies "Cookies build failure on hppa" [Normal,Resolved: fixed] [23:08] which is on trunk already [23:09] at least that was r+ ;) [23:10] do all of these changes have launchpad numbers somewhere? [23:10] and the convention is just old? [23:10] er, new? [23:11] lpxxxx mean there's a bug in launchpad [23:11] means [23:11] yeah, and bz is bmo, but there's a bunch that have neither [23:12] and I don't have good details on those if I can't look stuff up [23:14] well, those without lp or bz are explained somewhere in debian/changelog [23:24] fta: except without actually being linked to the patch itself, in many cases [23:25] really, I should just give you guys a list and ask what's up [23:26] you should find a "add debian/patches/whatever" somewhere in debian/changelog, at least if it's from asac or me, we are really consistent about that [23:27] but a list of questions would be fine too [23:27] what emails should I use? [23:28] <[reed]> fta@ubuntu.com, asac@ubuntu.com, probably... and CC me if you want me watching/poking, too [23:28] yes [23:28] ok [23:29] http://people.mozilla.com/~mconnor/trademark-review/Ubuntu/Round%201/ is basically what I'm doing [23:29] (also, boy is omnioutliner nice for this) [23:31] ftbfs-with-branding-dir is no longer needed, i thought i already removed it [23:32] installer_shouldnt_copy_xulrunner got a r- from bs, the feature is needed but not that way [23:33] installer_use_stdout_for_missing_files got a r+ from bs and landed in trunk [23:33] [reed], do you have the bug # ? [23:34] <[reed]> looking [23:34] nspr_flags_by_pkg_config_hack is needed because we use system nspr and we don't have the nspr sources in the xul/ff tarballs [23:34] mconnor, ^^ [23:35] w00t, awesome [23:35] <[reed]> 460911 nor -- fta+bugzilla@sofaraway.org RESO FIXE Linu Make the installer complain about missing files on stdout [23:35] <[reed]> 460913 nor -- fta+bugzilla@sofaraway.org ASSI Linu Installer shouldn't copy xulrunner files into Firefox install directory [23:35] <[reed]> 460915 nor -- asac@jwsdot.com RESO INVA Linu Build problem when using --with-branding-dir [23:35] <[reed]> 460917 maj -- asac@jwsdot.com ASSI Linu New plugins only recognized after restarting Firefox [23:35] dont_depend_on_nspr_sources, same thing [23:35] ew [23:35] paste-nasty [23:36] (looks fine here) [23:36] mozilla bug 460911 [23:36] Mozilla bug 460911 in Installer: XPInstall Engine "Make the installer complain about missing files on stdout" [Normal,Resolved: fixed] http://bugzilla.mozilla.org/show_bug.cgi?id=460911 [23:36] mozilla bug 460913 [23:36] Mozilla bug 460913 in Build Config "Installer shouldn't copy xulrunner files into Firefox install directory" [Normal,Assigned] http://bugzilla.mozilla.org/show_bug.cgi?id=460913 [23:36] mozilla bug 460915 [23:36] Mozilla bug 460915 in General "Build problem when using --with-branding-dir" [Normal,Resolved: invalid] http://bugzilla.mozilla.org/show_bug.cgi?id=460915 [23:36] mozilla bug 460917 [23:36] Mozilla bug 460917 in Plugin Finder Service "New plugins only recognized after restarting Firefox" [Major,Assigned] http://bugzilla.mozilla.org/show_bug.cgi?id=460917 [23:36] fta: I'm using irssi, it doesn't like some things [23:39] regarding 460913, i need to work in this, i have it in ff3.0/ff3.1/prism/fennec/xul-explorer/instantbird/bluegriffon [23:40] bluegriffon? [23:40] the new nvu from glazou [23:40] ah [23:41] i have experimental packages for all those [23:43] fwiw, my goal here is to minimize the "special" patches distros end up with [23:44] because really, other than maybe build config stuff, and some alternate bookmarks, I don't see why everything else can't just live in mozilla.org Hg [23:44] i agree [23:45] hmm [23:45] Dave Townsend's comment in the plugin reload bug is apt [23:46] in 460917 [23:47] patch from asac, i don't know the details behind this one [23:49] ok [23:49] I mean, this is a process, not a binary decision [23:50] I'm not, despite what some would think, trying to do bad things :) [23:50] [reed], you guys should have a builder tracking the qt build, it doesn't even build [23:50] mconnor, i'm sure you're not evil :) [23:51] <[reed]> file a bug :) [23:51] [reed], i have [23:51] <[reed]> to get a qt build machine? [23:51] not, the build issue [23:51] -not+no [23:52] builders for new platforms are unlikely to happen until January or so [23:52] we're still working on getting builds on Mercurial fully automated [23:52] and stabilizing our unit test infrastructure [23:52] [reed], i just didn't have time to produce a patch so far [23:52] mozilla bug 463872 [23:52] Mozilla bug 463872 in GFX "Cairo-qpainter build is broken after latest cairo update" [Normal,Resolved: fixed] http://bugzilla.mozilla.org/show_bug.cgi?id=463872 [23:53] it's not fixed [23:54] also, its qt [23:54] so... meh ;) [23:54] <[reed]> then reopen? [23:54] i'm trying to have a build for both qt and gtk in the same (3.1) branch [23:55] some kde users are producing rogue packages [23:57] rogue packages? [23:57] using unofficial branding, I hope!