[00:00] micahg: but you wants to determine the path at runtime, right? [00:00] want* [00:00] chrisccoulson: will that then drop the build-dep on xul? [00:00] micahg - it will [00:00] chrisccoulson: excellent another rdepends gone :) [00:01] nthykier: yes, I was wondering if anything that's not compiled is taking it [00:01] nthykier: woohoo [00:01] micahg, i'm completely lost on how to do a commit-specific compile. what exactly am i compiling? firefox? xulrunner? [00:01] bdrung_: good news? [00:01] LLStarks: depends [00:01] LLStarks: what do you want to compile? [00:01] nthykier: i found a working solution: we have to only set MOZILLA_FIVE_HOME [00:02] bdrung_: that's what a lot of the wrapper scripts do [00:02] awesome - THAT'S our solution [00:02] long term solution: fix upstream eclipse to find the correct place [00:02] micahg, i have my pushlog from yesterday and i'd like zero in on the offending commit. [00:02] i have a cloned mozilla-central hg repo, but i have no idea what to do with it. [00:02] LLStarks: if you're using upstream build system, it's firefox [00:02] nthykier: i will commit our solution to git repo [00:02] bdrung_: go ahead [00:03] LLStarks: hg help bisect [00:03] bdrung_: btw, upstream would not accept it - they allow us to provide a hook to find the library [00:03] nthykier: really? [00:03] nthykier: let's find the code first [00:04] bdrung_: if the xulrunner property is not set, they will reflect a class and check the value of the property after that class has been reflected [00:05] bdrung_: but if we can set that MFH variable, then we do that - because it is easier [00:05] It is a PITA to execute programs via Java [00:06] micahg: what path should we use? /usr/lib/xulrunner-1.9.2 --gre-version? how to detect the xulrunner version on compile time? [00:06] bdrung_: yep [00:06] oops [00:06] no [00:07] micahg, i'm a bit of noob and i'd like to handle this in a more insane manner by compiling each commit in the pushlog range. [00:07] /usr/lib/$(xulrunner-1.9.2 --gre-version) [00:07] /usr/lib/xulrunner-`xulrunner-1.9.2 --gre-version` [00:07] bdrung_: ^^\ [00:07] micahg: how to detect 1.9.2 on compile time? [00:07] LLStarks: that's what hg bisect helps you do [00:07] * micahg looks for the blog post [00:08] LLStarks: http://harthur.wordpress.com/2009/12/08/regression-range-finder-for-firefox-nightly-builds/ [00:08] i'm on that page [00:08] :) [00:08] nothing's happening though [00:09] ah one sec [00:09] here we go. [00:09] i have no idea what's being compiled or tested [00:09] micahg: /usr/lib/xulrunner-`xulrunner-@VERSION@ --gre-version` [00:10] micahg: i want to replace @VERSION@ on compile time [00:10] bdrung_: idk offhand, FF postinst script uses the version of the package [00:11] bdrung_: I don't think that's as big a deal as we'll only be changing that at most every 6 months [00:11] bdrung_: and we can fix that later too [00:13] micahg: can i assume a three number version (X.Y.Z)? [00:13] bdrung_: ATM [00:13] bdrung_: 4 [00:13] bdrung_: 4.0 most likely won't be [00:14] micahg: for @VERSION@ [00:14] bdrung_: yes, idk if they'll jump to xul 2.0 for FF 4.0 [00:14] bdrung_: in that case it wouldn't be, but your solution is probably good for the next year [00:14] micahg: other trick: search for xulrunner-* [00:14] that is enough - 3.6 will be out before then [00:15] eclipse 3.6 that is [00:15] first xulrunner-X.Y.Z.A - if it not exists xulrunner-X.Y.Z and otherwise xulrunner-X.Y [00:15] goddammit [00:15] bdrung_: can you see which xulrunner*dev is installed? [00:15] mozregression is tracing [00:16] ]AttributeError: Bisector instance has no attribute 'prevAppInfo' [00:17] micahg: pkg-config --modversion libxul [00:19] so frustrating [00:19] bdrung_: well, that's no so clear either [00:19] actually, ATM, yeah, you can do that [00:19] X.Y.Z [00:20] bdrung_: ^^ [00:21] bdrung_: micahg: I am off for now, unless you need me for something [00:21] nthykier: nope, thank you [00:22] you should thank bdrung_ - he is doing all the work and found the solution as well :P [00:23] nthykier: I intend too :) [00:32] micahg: do you know the corresponding eclipse bug number? [00:32] bdrung_: bug 559229 [00:32] Launchpad bug 559229 in eclipse "eclipse-platform: /etc/eclipse.ini references non-existent xulrunner path" [Undecided,Confirmed] https://launchpad.net/bugs/559229 [00:33] bdrung_: there's also bug 553779 [00:33] Launchpad bug 553779 in eclipse "xulrunner window causes swt exception" [Undecided,New] https://launchpad.net/bugs/553779 [00:33] bdrung_: thanks [00:35] you're welcome [00:39] in 12 minutes i will see if my fix work [08:45] asac, should i just push chromium or what? the ffe bug is not moving [08:53] Raise the Sun, Raise the soldiers moral o/ [09:03] fta great news tech.slashdot.org/story/10/04/13/0141208/Google-to-Open-Source-the-VP8-Codec [09:16] BUGabundo_remote, indeed, the missing http:// is disturbing [09:17] haaha [09:17] I filed it [09:17] no one touched it :( [09:17] that's why I hate to file bugs on google projects [09:17] they don't get fixe [09:17] and mozilla is even weired [09:18] they just talk talk talk talk talk and nothing [09:18] remember that session bug? [09:18] its still ongoing discussion on the FF bug [09:19] BUGabundo_remote, iirc, the last bug you filed were fixed, or at least assigned [09:19] which one? [09:19] I filed a bunch of them [09:20] and have one more, opened here to press SEND [09:20] :) [10:08] fta: WTH is M5 omnibar? code.google.com/p/chromium/issues/detail?id=41146 [10:09] ahhh [10:09] mileston [10:10] oh damn [10:10] its a 'feature' not a bug [10:10] FAIL [10:11] bye bye chromium, welcome back Firefox [10:14] http://code.google.com/p/chromium/issues/detail?id=40865 [10:28] BUGabundo_remote, if i understand it correctly, it's still a work in progress. it looks weird without http:// but if it's there when pasted, i won't mind [10:39] but its not! [10:39] you can't copy it [10:40] like my link 1h ago [10:40] only some app do fill in the gap [10:40] but it's a very bad UI regression [10:43] remember it's trunk [10:44] for 3 days? [10:44] and with 8 dupe bugs? [10:44] with them saying it's a feacture? [10:45] incl. a week end [10:45] ill wait a few more days [10:45] if not reverted, I'll drop ch [10:45] I use URL links a lot! [10:53] asac, i give up, there's no way i can sanely maintain chromium since it's frozen in the repo :( [10:54] :( [11:02] fta2: sorry? [11:38] fta: asac: why does html5test.com/ say both chromium and FF don't don't have mp3 support? === BUGabundo_remote is now known as BUGabundo_lunch [12:17] hey fta2, how is the chromium change going? [12:23] 11:53 < fta2> asac, i give up, there's no way i can sanely maintain chromium since it's frozen in the repo :( [12:23] seems fta needs some love :( [12:23] * asac hugs fta2 [12:23] chrisccoulson: ^^^ [12:24] * chrisccoulson hugs fta2 too [12:24] fta2 - if you have any issues, please let me know :-) [12:56] chrisccoulson, i didn't have time to work on this. if you do, please feel free (on the .head branch) [12:57] chrisccoulson, and i don't see how i can upload. See bug 561510 [12:57] Launchpad bug 561510 in chromium-browser "FFe and new upstream as security updates for chromium-browser" [Undecided,Incomplete] https://launchpad.net/bugs/561510 [12:58] fta2: take one step back ;) [12:58] fta2 - ok, i will try and push that along. the current process of requiring a FFe doesn't really fit well with the chromium support model [12:58] fta2: stefan offered to delegate FFe to me and you refused it ;) [12:59] but we will take over [12:59] chrisccoulson: thanks. its really a non issue. [12:59] lets close the bug and just upload [12:59] asac - ok, so you can ACK that? [12:59] excellent [12:59] seems that ubuntu-release has no plan what to do that [12:59] thanks :) [12:59] chrisccoulson: yes, i can ack that [13:02] done [13:02] fta2: just upload closing that bug. i take the blame if someone complains [13:02] or chrisccoulson do that ;) [13:02] chrisccoulson: i think for the search provider the idea is to not add a patch which would be a pain to maintain, but rather a sed command in rules [13:02] did fta2 align with you on that? [13:03] asac - no, i wasn't aware of that, but that makes sense if it is possible [13:03] chrisccoulson: should be [13:03] cool. i'm just doing fennec now [13:04] that looks fairly trivial to do [13:04] great [13:06] asac, i didn't refuse to delegate to you or to anyone, just to delegate to the mozilla delegates, it's not a mozilla product [13:08] well. there is no delegate group for chromium ;) [13:08] anyway. not a problem. lets get it just in before final freeze and then get the standing FFe etc. [13:12] is micahg not about? [13:12] directhex, he's usually around later in the afternoon [13:12] asac: fta: how about rename this channel to ubuntu-browsers ? [13:12] and create a oficial team in ppa to run the oficially supported browsers? [13:15] in case i forget, can someone ask him to sync gluezilla? it fixes xulrunner 1.9.1 requirement (and lack of xulrunner dependencies) [13:15] directhex: no diff for ubuntu? [13:15] kk [13:15] asac - you can do that anyway can't you? [13:16] asac, same source package test-built on sid & lucid with appropriate 1.9.1/1.9.2 binary dep in the output [13:16] chrisccoulson: synching? [13:16] asac - yeah [13:16] i have no archive powers ;) [13:16] so i would need to poke ;) [13:16] asac - oh, i thought you did ;) [13:16] like everyone else [13:16] no ... just MIR [13:16] i always felt like not picking up more responsibility would be a good idea [13:17] ... powers always come with new tasks ;) [13:17] it is :) [13:17] i try to get other people to issue pokes, so the archive admins don't get cross with me [13:17] yeah, i always poke pitti when i need something doing ;) [13:17] lol [13:18] hmm, ff 3.7 freezes badly when i visit http://www.webreference.com/authoring/languages/html/HTML5-Client-Side/ [13:19] (i wanted to see the "Browser Feature Test Results" box) [13:19] directhex: same here ;) [13:19] anyway, this sync is needed for .net apps using windows gui toolkit and embedded browser control to not segfault when xul 1.9.1 isn't installed [13:20] i already streess archive admins and ubuntu-release enough [13:20] fta will test in a few secs [13:20] need dgb of that? [13:21] actually, i can upload it myself can't i... [13:21] i'll do it tonight [13:21] directhex: yes. [13:21] just upload with ubuntu1 ;) [13:21] or just upload with the debian version [13:21] i think that is considered bad ... but still ;) [13:21] * asac break [13:22] fta where is it crashing exaclty? [13:22] page loaded fine for me, on FF3.7 daily [13:22] asac, there's a script. but i need my .changes [13:22] Browser Feature Test Results Session Storage: supported [13:22] Global Storage: supported [13:22] Local Storage: supported [13:22] Database Storage: not supported [13:22] BUGabundo_lunch, freezing, not crashing (and unkillable) [13:23] damn, Database Storage not supported :( [13:23] and 3.6? [13:23] sec [13:24] Browser Feature Test Results Session Storage: supported [13:24] Global Storage: supported [13:24] Local Storage: supported [13:24] Database Storage: not supported [13:24] both in safe-mode [13:24] so no addon interfires [13:30] fta_ chromium: [13:30] Browser Feature Test Results [13:30] Session Storage: supported [13:30] Global Storage: not supported [13:30] Local Storage: supported [13:30] Database Storage: supported === Loki_ is now known as Loki === BUGabundo_lunch is now known as BUGabundo_remote [13:34] nobody has any outstanding items for fennec do they? [13:34] if not, then i'm just going to upload [13:42] asac - i get a FTBFS with videolink against xulrunner192 btw. i'm not sure if that's actually a xulrunner bug: [13:42] /usr/include/xulrunner-1.9.2.3/nsPresContext.h:889: error: 'nsRunnableMethod' is not a template [13:42] chrisccoulson: need to see the full error [13:43] asac - the only other line is: [13:43] In file included from event_state_manager.cpp:9: [13:43] asac - http://paste.ubuntu.com/413625/ [13:45] chrisccoulson: videolink is using MOZ_INTERNAL_API [13:45] such things need to get dumped from the archive imo [13:45] asac - ok, i'm happy with that [13:45] it has no rdepends anyway [13:45] yeah. lets dump it ... until upstream fixes to use frozen api [13:46] chrisccoulson: or try to drop the MOZILLA_INTERNAL_API define first [13:46] maybe it just works then [13:46] if not, remove [13:46] * asac has to run out and get some stuff [13:46] ok, i'll try that first, but i suspect it probably won't work at all [13:56] BUGabundo_remote, thanks. Global Storage is a moz hack, deprecated. so no worry for ch. I guess i have to use local storage then, i wanted db storage :( [14:09] chrisccoulson, the idea for the patch is to make it dynamic in d/rules, so we can upload the same src package to debian without it. so something like "ifneq (,$(filter Ubuntu,$(DEBIAN_DIST))) perl -i~ -pe 's/foo/bar/' file endif" should work [14:10] i wanted to do it that way but i lacked time to properly build and test it :( [14:12] fta2, how do i restart a bisect? [14:12] LLStarks, what do you mean? [14:12] fta2 - ok, no worries. i will try and look at that this afternoon [14:12] the bisect i did is lying to me and won't let me double check the work it did [14:13] it's all "lol trust me. you don't need to compile the commit i think is bad. just trust me." [14:13] eh? [14:14] when bisect finds the first bad commit, it doesn't forward the repo to that commit [14:14] it remains on the last good commit [14:14] and i can't get the repo to move up the first bad commit [14:14] and thus i can't double check [14:16] sorry, i don't know anything about that. I usually bisect manually. [14:16] (hence all my daily builds) [14:37] chrisccoulson: we really need to fix the ffox dailies for 3.6 [14:37] thats where all the rework for oopp thing landed [14:37] and we currently dont see how bad it is ;) [14:37] at least the last snapshot i use has severe problems ;) [14:38] asac - ok, i will try and fix that in a bit [14:38] if we dont fix that now we cant provide feedback to mozilla and we cant complain if next security update blows things up for us :( [15:00] chrisccoulson: hi! I was wondering how the progress was going on bug #559881 for lucid and jaunty. Also, is tbird3/lucid affected by #559918? [15:00] Launchpad bug 559881 in nss "libnss3-1d 3.12.6-0ubuntu0.9.10.1 breaks ssl/fips support in firefox" [High,Fix released] https://launchpad.net/bugs/559881 [15:00] bug #559918 [15:00] Launchpad bug 559918 in thunderbird "Thunderbird cannot initialize the security component when libnss3-0d 3.12.6 is installed" [High,Fix released] https://launchpad.net/bugs/559918 [15:03] jdstrand - i'm going to do those this afternoon [15:03] and thunderbird 3 doesn't have the same issue [15:03] chrisccoulson: cool, thanks [15:04] chrisccoulson: I'm also interested in getting nss 3.12.6 in hardy and jaunty. do you have thoughts on this? [15:04] (to fix CVE-2009-3555) [15:04] The TLS protocol, and the SSL protocol 3.0 and possibly earlier, as used in Microsoft Internet Information Services (IIS) 7.0, mod_ssl in the Apache HTTP Server 2.2.14 and earlier, OpenSSL before 0.9.8l, GnuTLS 2.8.5 and earlier, Mozilla Network Security Services (NSS) 3.12.4 and earlier, multiple Cisco products, and other products, does not properly associate renegotiation handshakes with an existing connection, which allows man-in-the-middl [15:04] jdstrand - we will need to eventually, but there's not much benefit at the moment [15:05] i think the fix also relies on quite a large xulrunner patch [15:05] chrisccoulson: not for firefox, but there is for other users of nss [15:05] ah, ok [15:05] i don't mind really then [15:05] (I thought we were going all embedded for ff anyway) [15:06] we have done in lucid [15:06] chrisccoulson: yes. whenever security support goes away for 3.0/3.5 is what I mean [15:07] chrisccoulson: maybe the plan of attack should be to wait on hardy and jaunty until after firefox uses its own embedded nss [15:07] after which, we can upgrade the system nss [15:07] chrisccoulson: does that sound reasonable? [15:07] yeah, that sounds good. i'm not entirely sure what the timeframe for that is though [15:08] jdstrand - i've also been trying to think of ways we can coordinate testing of security updates in the future [15:08] do you think something like http://iso.qa.ubuntu.com/ targetted for mozilla updates would be useful? [15:08] chrisccoulson: sure, but they keep threatening it. I'm not interested in fixing this right away-- it's a protocol change that we'll need to do gradually, and so waiting on 3.6 seems reasonable [15:10] chrisccoulson: maybe. how it was done in the past is asac would do a bunch of testing, then give me the ok (sometimes asking for additional testing). I would then do my list of tests and publish. in the more distant past, people would use and give feedback on the packages in ubuntu-mozilla-security iirc. there might even be an announcement that people would get to know it was available [15:11] chrisccoulson: this is more or less what I follow https://wiki.ubuntu.com/MozillaTeam/QA [15:12] though I don't test all the plugins-- just adobe flash and java, and then verify opening of a slew of mimetypes (including PDFs) [15:12] ok, that looks good as a testing sequence. how do users communicate their testing results though? [15:12] oh [15:12] chrisccoulson: I think they communicated with asac in this channel, but I'm not sure [15:12] it's at the top of the page [15:12] oh, there you go :) [15:12] http://mozilla.qa.ubuntu.com/ [15:12] ok, so that looks good [15:13] so, we should resurrect this then :) [15:13] chrisccoulson: I'm not sure, but it seems that lately this hasn't been happening as much, probably because mozilla had several 'fire drills' in a row [15:13] chrisccoulson: sounds good [15:13] chrisccoulson: incidentally, my tests are based on lp:qrt-regression-testing/scripts/test-firefox.py [15:14] cool, thanks. i will take a look at those [15:14] I wish it was automated, but it isn't-- the script just walks you through stuff [15:14] but will test epiphany-browser in hardy-jaunty [15:15] (also) [15:15] ah, there is even a thunderbird test plan in there-- cool [15:15] (that wasn't always there) [15:17] https://litmus.mozilla.org/ looks interesting. wouldn't it be neat if we could get our builds integrated into that? [15:18] it would. perhaps i might look at things like this for next cycle [15:25] chrisccoulson: I'll update our CVE tracker for nss based on our discussion and plan [15:42] chrisccoulson: I'm assuming miro from the PPA worked fine for you with only xul192? [15:52] micahg - it seems to be working fine here [15:53] chrisccoulson: k, I guess I'll deal with the fallout if there is any, probably like you said with apport being turned off, most people won't notice [15:53] I'm testing vlc now [15:55] micahg - thanks [15:55] we dropped another rdepends today (videolink) [15:55] chrisccoulson: BTW, is packagekit done? [15:56] micahg - i think packagekit just needed rebuilding. when was it last built? [15:56] needs a loog (chrisccoulson) [15:56] 5 weeks ago [15:57] ok, packagekit still needs a rebuild then [15:57] chrisccoulson: can you test that and take care of it, it's in main [15:57] jdstrand: chrisccoulson: we want to regularly announce security ppa [15:57] and ask for more folks to join and where to escalate issues [15:57] we need to resurrect this to make us better sleep at night ;) [15:58] asac: FYI, we can't push 3.6.4 to security PPA until we figure out why our builds don't work [15:58] asac - yeah, i was thinking about something like http://iso.qa.ubuntu.com/, before i realised that we already have http://mozilla.qa.ubuntu.com/ [15:58] micahg: yes. chrisccoulson is working on fixing the dailies soon ;) [15:58] we could make better use of this [15:58] chrisccoulson: we already have that [15:58] micahg: re NSS, I just filed bug #562332 [15:58] asac: it might be fixed by the xulrunner patch [15:58] Launchpad bug 562332 in nss "please upgrade to 3.12.6" [Medium,Fix released] https://launchpad.net/bugs/562332 [15:58] chrisccoulson: however, it turned out to be really hard getting folks submitting stuff there [15:59] if we could get there it would be best [15:59] jdstrand: great [15:59] chrisccoulson: however, geting folks using -security ppa all the time is the first step [15:59] micahg: in summary-- we want to, but are going to wait until the embedded firefoxs hit hardy - jaunty [15:59] then asking when we push to beta or something to test and submit to mozilla.qa would be great [15:59] asac - do you know what users found difficult about submitting there? tbh, i've not used it yet, but it looks like it could be useful [15:59] micahg: that'll help ease testing [15:59] jdstrand: ok, that should be later this month hopefully [16:00] for some definition of 'hopefully' :P [16:00] well, I made progress last night on porting some hardy packages [16:00] I already started hitting things that bug me about all the embedded stuff [16:00] jdstrand: what do you mean embedded stuff? [16:01] eg, the recent libpng CVE-- I had to check the mozilla source to see if it was fixed. I couldn't see anything anywhere stating it was fixed (it was) [16:01] anyway, my complaining won't change anything, so I'll stop [16:03] jdstrand: did you try searching for it in the upstream tracker? [16:03] chrisccoulson: its not difficult to submit there ... its just hard to get enough folks doing that [16:03] so in the end i settled on getting stuff early on the security ppa [16:03] and getting users use that so they start complaining [16:03] if something breaks [16:04] if you dont hear anything, its probably fine [16:04] micahg: I just checked the source changelogs and didn't see anything. I might have missed it, but it only illustrates the problems with embedded libs, that's all. if it used a system libpng, I knew it was fixed. since it doesn't, I had do do additional stuff to see if it was vuln or not [16:04] chrisccoulson: if you want to try to resurrect, go ahead :) [16:04] its not really mutually exclusive ... we have to announce the security updates anyway [16:04] includin a request to submit the qa thing is probably good [16:05] asac - i will probably try to resurrect this, but that's something for me to look at next cycle now [16:05] but yeah, we need to announce the updates first [16:08] yeah [16:08] chrisccoulson: we can discuss all the great and perfect ideas i had during UDS ;) [16:08] while drinking a beer [16:08] asac - sounds good :) [16:08] maybe you want to pick some of those up ... or not [16:09] can I come too? [16:13] asac, i'm not sure what i'm supposed to do now? upload or provide the -imho useless- info asked for the FreezeExceptionProcess? [16:13] asac: was DEB_MIN_SYSDEPS unintentionally applied to NSS and NSPR? [16:19] fta: you started it by refusing to accept that mozilla delegates can approve it ;) ... [16:19] imo you should just upload [16:19] now [16:19] unless they vetoed on the comment i made [16:19] they didn't, until now [16:21] fta: just upload then [16:21] ok [16:22] lose this bug and never open a new FFe bug [16:22] we sent the mail you prepraed to technical-board [16:22] i think it will just get approved and then we dont need to bother about it [16:22] if someone complains, blame me ;) [16:23] ok [16:31] asac, you meant you already sent the email to TB? [16:32] no i didnt [16:32] we can do that today [16:32] i can send as you seem to be unhappy about paperwork ;) [16:32] which i totally understand ;) [16:43] chrisccoulson: Bug 520166 is fixed according to reporter [16:43] Launchpad bug 520166 in firefox "Manage Content Plugins doesn't work with Firefox 3.6" [High,Fix released] https://launchpad.net/bugs/520166 [16:43] chrisccoulson: please add a fat sticky note to your desk that we need to update the plugin finder database before release [16:43] ;) [16:44] asac - ok, no problem [16:44] whats involved to update that? [16:44] lets not discuss that right now ;) [16:44] its simple, but i need to explain and probably we need IS love to get you permission to change it [16:44] chrisccoulson: vlc is ready if you want to review [16:44] asac, done [16:44] cool ;) [16:44] ok, no problem [16:45] chrisccoulson: bug 558981 [16:45] Launchpad bug 558981 in vlc "vlc fails to build against xulrunner-1.9.2" [High,Confirmed] https://launchpad.net/bugs/558981 [16:45] chrisccoulson: add a fat note ... we can do that after freeze is enforced [16:45] micahg - thanks, i will look at that in a bit. i need to get xulrunner working on ia-64 first though [16:45] chrisccoulson: did you see my note? [16:45] micahg - i did. but the fix for the test-suite is mostly trivial. the hard part is testing it [16:45] which is what i'm going to try this afternoon [16:45] chrisccoulson: ok [16:51] asac, someone is cloning the chromium ppa: https://edge.launchpad.net/~towolf/+archive/crack/+packages?field.name_filter=&field.status_filter=&field.series_filter= [16:54] fta: binary copies? [16:54] or respinning? [16:54]