=== fta_ is now known as fta [01:18] fta: Mozilla/5.0 (X11; U; Linux x86_64; en-US; rv:1.9.1b2pre) Gecko/20081101 Shredder/3.0b1pre [01:18] ? [01:19] and ? [01:19] fta: sorry got distracted ;) [01:19] didnt want to summon you ;) [01:19] asac, b2~build1 < b2~hg and b2+build1 > b2 :( [01:19] asac, I have b2~rc1 left [01:22] mozilla bug 458612 [01:22] this is turning against ubuntu [01:22] Mozilla bug 458612 in GFX: Thebes "(Ubuntu) system fontconfig settings override GNOME font rendering screen settings" [Normal,New] http://bugzilla.mozilla.org/show_bug.cgi?id=458612 [01:29] people will sure complain about this [01:29] got a dozen complains already [01:37] fta: so does changing the priority fix it? [01:37] yes, but this has impacts on some other apps [01:37] fta: how can that be? [01:37] are there other confs in between? [01:39] i didn't try what's in the last comment, i simply removed the rule file overriding argb [01:39] is .fonts.conf the preferred gnome way to configure stuff? [01:44] i don't have enough background about why ubuntu added all those rules in fontconfig :( [01:44] fta: maybe its just "GNOME settings fail to apply for non-GNOME apps" ? [01:45] non gnome apps are ugly (ddd & google-earth come to mind) [01:45] fta: so can you confirm that evolution is ok, but non-gnome apps are ugly? [01:48] so 10-no-sub-pixel.conf would override 50-something.conf? [01:49] http://www.sofaraway.org/ubuntu/tmp/bad.png [01:49] no, 10 < 50 so it does not override [01:50] yes . but last comment changes 10 ... how can that help? [01:50] maybe fonts.conf mechanism is accept-first-only ;) [01:51] or maybe that changed in the meantime and nobody noticed that regression ;) [01:51] i just removed 10-no-sub-pixel.conf to fix ff3.1 [01:51] but in the end i think its really 53-monospace stuff [01:51] hmm [01:52] does that fix the other apps too? [01:52] but it's probably not the best thing to do [01:52] of course not [01:52] the other apps didn't change a bit (the gtk ones) [01:52] the motif ones turned ugly [01:53] well. apps not using modern fonts stuff can be ignored i think [01:55] fta: try to flip the order ;) [01:55] 3am, i need sleep [01:55] hehe ;) [01:55] yeah [01:55] ok [01:55] just an idea [01:58] something broke prism [01:58] badly [01:58] i didn't touch it recently, so it's xul 1.9 [01:58] fta: was it rebuilt yet? [01:59] i'm talking about prism 0.9.1+svn20081028r19379-0ubuntu1~fta1 [02:00] Error: uncaught exception: [Exception... "Component returned failure code: 0x80520012 (NS_ERROR_FILE_NOT_FOUND) [nsIXPCComponents_Utils.import]" nsresult: "0x80520012 (NS_ERROR_FILE_NOT_FOUND)" location: "JS frame :: chrome://webrunner/content/webrunner.js :: :: line 31" data: no] [02:00] Error: WebRunner is undefined [02:00] Source File: chrome://webrunner/content/webrunner.xul [02:00] Line: 167 [02:01] asac: are you ccing me on the upstream stuff? can you start with the ND items? [02:01] asac: I'd like to hash out the ND items before 3.1b3 [02:01] mconnor: aye, aye sir. [09:28] fta: Why do all mozilla packages have an additional gnome-support package? I can understand the need to have a dev [09:59] wiki, kde users don't want gnome libs on their system [10:11] have the QT libs in FF source yet? [10:11] s/have/are [10:22] no, i'm not done yet. i don't have much time to work on moz stuff [10:23] oh you are doing it, i thought upstream was doing it [10:25] they are, but the packaging needs a lot of work. i'm done with xul, not ff [10:27] ah ok [10:29] fta: last read, lots of stuff didn't work like printing in the QT port. Are they fixed ? [10:29] as i said, i'm not done with ff so i don't know yet [10:30] :) [10:30] i have to go now. cu. [10:31] bbye [11:21] i thought firefox uses %s when launching [11:21] maybe %u? [11:29] %u doesnt bring up a real site it brings up www.%u.bleh [12:44] gnomefreak: where is that? [12:47] wtf, b2~rc1 still calls itself 3.1b2pre [12:47] and jumped to official branding [12:51] is m-d broken? [13:21] fta: in which way would m-d be broken? [13:21] i think upstream uses official branding for beta milestones .... its just that we dont want that [13:50] fta: can you close changelog and upload 3.1 + 1.9.1 please? [13:50] i can also do it if you have no time ;) [14:17] asac, what i mean it that m-d ignores the tag for hg [14:17] strange because it used to work [14:29] got it, damn, i don't know how it could have worked before [15:50] : (malloc) Unsupported character in malloc options: 'O' [15:50] ??? [15:50] it's gone ? [16:33] fta2: fun [16:34] not really, if it's really gone, it will freeze the builders [16:35] when did this start? [16:38] what? the freeze or the Unsupported option? [16:42] the latter [16:43] not sure, 1st time i see it [16:45] what did you do in between ;) [16:45] http://hg.mozilla.org/mozilla-central/rev/36156fbf817d [16:45] * asac looks [16:45] nothing, as i said before, i'm busy with something else [16:47] fta2: oops [16:47] fta2: seems like the first hunk is an accident? [16:48] mozilla bug 464995 [16:48] Mozilla bug 464995 in jemalloc "avoid reads of size 1 in jemalloc" [Normal,Resolved: fixed] http://bugzilla.mozilla.org/show_bug.cgi?id=464995 [16:49] fta2: commented [16:50] thanks [17:22] asac, any reason nspr/nss are not pushed ? [17:26] yes ... same reason as always [17:26] soname [17:27] ? do you want soname or not? [17:41] fta2: how did you get the malloc error above? [17:41] during build? [17:41] get-orig in ff [17:41] * fta2 leaving, cu [18:04] does the .install files support regex.If so, will python regex work? [18:11] wiki: not sure. i think it only supports * [18:11] what about char classes ? [a-z] and all that [18:12] it kinda worked but ^ didn't [18:12] where do I get to know aboout this? which dh_ script handles this? [18:13] yes [18:13] install [18:14] dh_install [18:20] asac: supports wildcards :( [18:21] thanks anyway [18:21] like i said ;) [18:33] fta: we need more info on your regression [19:06] asac, "we" ? [19:22] james_w`: i have a bzr package like: http://paste.ubuntu.com/77184/ [19:22] james_w`: and a orig like: http://paste.ubuntu.com/77183/ [19:23] james_w`: bzr bd --merge --dont-purge fails because the produced tree is somehow nested: [19:23] james_w`: http://paste.ubuntu.com/77185/ [19:23] james_w`: how can that happen? [19:23] fta: can you explain to me how you see that error? [19:23] fta: mozilla says that change slipped in, but is ok and no need to stop b2 because of that [19:23] "pax_global_header" [19:24] if we see a regression we should do provide them with the info about the jemalloc problem we see [19:24] james_w`: whats that? [19:24] asac: dunno, something from your tar file [19:24] james_w`: no. thats only in the "merged" tree [19:24] ;) [19:25] james_w`: at least i dont see it in the orig list i pasted [19:25] http://paste.ubuntu.com/77183/ [19:25] yeah, it's from the tar file [19:25] you generated that list from "tar tzf"? [19:25] james_w`: yes [19:25] that's the difference then [19:25] hmm [19:26] whatever generated the tar file added a header that the python tarfile implementation doesn't understand, and believes to be a normal file [19:26] james_w`: i unpacked it ... its not in there either :( [19:26] james_w`: oh [19:26] james_w`: thats git archive [19:26] james_w`: scary [19:26] james_w`: so python cannot untar git archives ;) ... fun [19:26] bzr-builddeb then sees you have two files in the root, and thinks you have an odd upstream tarball that doesn't put things in a subdir [19:27] james_w`: why doesnt bzr-builddeb just use good old command line tar ;) [19:27] ok [19:27] now i am buffled [19:27] so i unpack it and pack it again ;) [19:28] lol [19:28] that works [19:28] wow [19:32] http://archives.free.net.ph/message/20060301.180552.3c03b2d3.en.html [19:34] I think special-casing that file would work [19:34] unless I can find out how to make dpkg-source do what I need [19:36] 1.9.1~b2~rc1+1+nobinonly <= 1.9.1~b2~rc1+nobinonly ??? [19:39] fta: dpkg --compare-versions 1.9.1~b2~rc1+1+nobinonly lt 1.9.1~b2~rc1+nobinonly | echo TRUE [19:39] TRUE [19:39] ^ ^ [19:39] james_w`: yeah. now that i know all is fine ;) [19:40] sebner: 1 < r [19:40] damn [19:40] asac: hm? [19:40] i should not work on this between meetings [19:41] heh [19:41] fta: why rc1? its +build1 ;) [19:41] then +build2 [19:41] etc [19:41] just like the tag [19:42] i wanted to do b2 once it's final, but b2+build > b2 [19:42] fta: it worked in the past because: [19:42] +build1+nobinonly < +nobinonly [19:42] and b2~build < b2~hgXXXXX [19:43] which is why i suggested +build1 [19:49] i seriously need holidays [19:49] next week that is ;) [19:49] beers in california :-D [19:50] fortunately, 1.9.1~b2~rc1+1+nobinonly < 1.9.1~b2+build1 < 1.9.1~b2+nobin [19:51] :) [19:51] finally some fortune :) [19:57] mozilla bug 466531 [19:57] Mozilla bug 466531 in JavaScript Engine "Crash/hang [@ mult][@ Balloc] when loading pages on PPC" [Critical,New] http://bugzilla.mozilla.org/show_bug.cgi?id=466531 [19:57] hm [19:58] see mozilla bug 463073 [19:58] Mozilla bug 463073 in Security: PSM "Deliver NSPR_4_7_3_RTM to Mozilla" [Normal,New] http://bugzilla.mozilla.org/show_bug.cgi?id=463073 [19:58] #14 [19:59] fta: is the first bug the same you get with the jemalloc glitch? [19:59] no, unrelated [20:00] k [20:00] fta: so what are the symptoms for the jemalloc thing? [20:06] initially, it was mozilla bug 452754 [20:06] Mozilla bug 452754 in jemalloc "Deadlock in jemalloc when using fakeroot" [Normal,New] http://bugzilla.mozilla.org/show_bug.cgi?id=452754 [20:07] fta: and now that doesnt work anymore? [20:08] now i'm getting : (malloc) Unsupported character in malloc options: 'O' [20:08] fta: ok. i think you can just drop that switch now [20:08] for: DEBIAN_XUL_VER := $(shell MALLOC_OPTIONS=O xulrunner-1.9.1 --gre-version) [20:08] previously it hanged in pagefile [20:08] now they did #define MALLOC_PAGEFILE [20:08] er [20:08] err [20:08] they removed that define [20:08] so probably we dont do pagefile anymore [20:08] which hopefully means that its ok now [20:09] O != 0 [20:09] fta: yeah [20:09] even then [20:09] that seems to be a MALLOC_PAGEFILE feature [20:09] now that isnt enabled anymore [20:09] at least worth a try [20:10] ix:~/bzr/firefox-3.1.head$ fakeroot xulrunner-1.9.1 --gre-version [20:10] 1.9.1b2pre [20:11] seems fine, but i need to try with xul b2 [20:11] fta: yes. [20:11] also remember that right after b2 they will backout that things again most likely [20:11] so you need to readd that ;) [20:11] crazy ... i know [20:12] you could also prepatch that backout [20:12] to avoid this [20:12] https://bugzilla.mozilla.org/show_bug.cgi?id=464995 [20:12] Mozilla bug 464995 in jemalloc "avoid reads of size 1 in jemalloc" [Normal,Reopened] [20:12] afaics, it's just a warning so i could just let it untouched === james_w` is now known as james_w [21:09] asac: last i see on screen is .:07:44:27:. >> asac << gnomefreak: where is that? can you please remind me what this is about? im thinking about qt support but not sure [21:13] kde is borkerd something bad [21:13] -r === thunderstruck is now known as gnomefreak [21:16] asac, what did you do why your last commit in 3.1 ? it's not a merge :( [21:16] -why+with [21:17] the fix was in 3.0.4, not in 3.1* [21:17] fta: because the rest wasnt merged either yet [21:17] i did merge -c 377 [21:17] i was doing it, now i'm all diverged :(( [21:18] ok court and work sucked today im going to rest and get ready for the shit tomorrow btw if you can avoid it dont hire or fire anyone :( [21:19] fta: you should directly push on commit ,) [21:19] or bzr bind [21:19] i ran update before ... nothing came. [21:20] told you i was waiting for xul to be built [21:20] it seems biggest issue is with mysql bullshit [21:20] fta: before pushing? why? [21:20] fta: so is your local tree now broken or not? [21:20] today it's not my day, i should stop everything right away [21:21] -it's+is [21:21] not mine either so dont feel too bad [21:22] gnomefreak: thought in US you can fire without any reason [21:23] asac: you can but moron thinks i fired him because hes black but the one person left is black :( [21:23] fta: if it helps you push overwrite [21:23] when the NAACP gets involved it becomes many bad days [21:24] NAACP? [21:24] oh minority group [21:24] dont reember whast it stands for off hand but its black american right org. [21:28] national african american something project manybe? [21:31] * gnomefreak wonders if there is a chance PA will be fixed next week [21:31] holiday next 4 or so days ;) [21:32] PA is rotten for me too. better than last week but still half broken. [21:32] fta: looks like deps need to be met [21:33] IIRC PA comes in 2-3 sources [21:33] for me, it installs just fine [21:34] so i dont see hhowit can break like this [21:34] its held back here [21:34] last time i tried it got all screwed up [21:34] at last, bzrtools is in, i had to ask in #bzr earlier today [21:35] held back are: [21:35] file-roller libpulse-browse0 libpulse0 pulseaudio-esound-compat pulseaudio-module-gconf pulseaudio-module-hal pulseaudio-module-x11 pulseaudio-utils xserver-xorg-input-evdev [21:36] The following packages have been kept back: [21:36] xserver-xorg-input-evdev [21:36] that's all [21:36] ok ill try iunsttalling them in a few [21:43] The following held packages will be changed: i never seen that before [21:44] asac, i wanted to re-enable system nss/nspr in 3.1b2, but apparently, you reverted the soname stuff so now i'm unsure [21:46] fta: does it fail to build with current version? [21:47] no but the urandom stuff is still there [21:49] fta: what doesnt work ? [21:50] remember the no more than 3 users bug? [21:51] ah [21:52] err [21:52] well 3 users? [21:52] anyway [21:52] fta: ,$(@:$(OBJDIR)/%.so=%.so) [21:52] what does that mean? [21:52] DSO_LDOPTS='-shared -Wl,-soname,$(@:$(OBJDIR)/%.so=%.so)' [21:52] is that the full line ? [21:52] ok [21:53] thats somewhat broken in our nspr build [21:53] upstream nspr has a SONAME as you might guess ... ours is empty [21:53] interestingly enough the nss one is ok [21:53] e.g. its just libnss3.so [21:53] here it hsould just be libnspr4.so (for example) [21:53] but its empty [21:53] thats basically the only blocker for the upload of the transition [21:54] nss is quite perfect [21:55] $ objdump -x libnspr4.so | grep SON SONAME libnspr4.so [21:55] thats in upstream tarball [21:55] if we drop the 81_soname patch (e.g. what i did) [21:55] its just empty [21:56] and the DSO_LDOPTS looks like that afaict [21:56] what is/are the target(s) of that rule ? [21:57] fta: hmm [21:57] fta: DSO_LDOPTS='-shared -Wl,-soname -Wl,$(notdir $@)' [21:57] at 1st glance, it just drops the dirname of the .so file [21:58] fta: i pasted none-sense [21:58] what i poosted comes from configure.in [21:58] what is important for us is autoconf.mk [21:58] in my build tree we have: ./config/autoconf.mk:DSO_LDOPTS= -shared -Wl,-soname -Wl,$(SONAME) [21:59] hmm [21:59] SONAME is nowhere defined :( [21:59] is was defined somewhere before [21:59] it [21:59] hah [21:59] i have it ;) [22:00] let me check [22:01] dpkg-buildpackage: host architecture i386 [22:01] fakeroot debian/rules clean [22:01] : (malloc) Unsupported character in malloc options: 'O' [22:01] /usr/share/cdbs/1/rules/buildcore.mk:72: parsing mozilla-central-3.1~b2+build1-source.tar.bz2 ... [22:01] yep, just a warning [22:02] its really my memory ;) [22:02] no i remember that last time i did that i ran in the same issue ;) [22:03] but believe it or not, i cannot remember if the fix i am now trying helped ;) [22:03] \o/ [22:03] i think it works [22:03] let me stresstest this a big [22:03] bit [22:08] damn, i know why the branding jumped to final [22:08] ifneq (,$(findstring ~b,$(DEBIAN_VERSION))) [22:08] ~b matches beta *and* build [22:09] hmmn, no, not anymore.. +build [22:09] fta: yeah. actually thats why i said that again the other day [22:10] but when you said that its ok, i thought i just overlooked that its not FORCE_OFFICIAL [22:10] and thought that would exclude it [22:10] it's the old code, in 3.0, beta was final [22:10] maybe we want FORCE_UNOFFICIAL ;) [22:10] fta: right [22:10] that code is still there, untouched [22:11] fta: can you move ~b to "milestone" branding? [22:11] ok i think nspr and nss branches are good now [22:12] i verified that removing a symbol from .symbols will fail with the diff [22:12] hmm still one lintian E ... but unrelated [22:12] lets see if thats easy to fix [22:15] hmm ... what is the DEBHELPER token for posinst/rm :( [22:20] old stuff i think [22:21] no i wondered what the right syntax is [22:21] added something no [22:21] w [22:29] fta: ok nspr.head and nss.head seem to be ok. can you also spin them ... install the binaries and see if you end up in any binary compatibility issues? (e.g. apps without respin break) [22:30] or wait one sec [22:30] ok nspr.head rev 54 [22:30] nss.head rev 90 [22:31] thats good crack i hope ;) [22:31] fta: can you test build and see if there are any binaries conflicts? [22:31] fta: i will do a xulrunner build and see if the dependencies are right there too [22:32] first just firefox (e.g. new libnss/nspr + old xul) [22:32] not really expecting that it works [22:32] ;) [22:33] ephy works too without respin [22:33] not much tht can be wrong then i guess [22:33] evolution + SSL works too [22:34] (without respin) [22:34] firefox-30 build succeeded ... even though xulrunner is not yet respun [22:36] oh [22:36] we have to bump symbols to current package version [22:36] (when we are happy with it) [22:39] https://blueprints.edge.launchpad.net/ubuntu/+spec/retaining-ubuntu-developers [22:41] fta: ok branches are ready now [22:43] fta: i guess you have comments for that spec [22:49] fta: let me know if you want to test nss/nspr still [22:49] i will take a break and then start to forward bugs. would appreciate if you could close xul 1.9.1 before that ;) [22:50] err s/bugs/patches/ [23:07] hm, reco, did i talk to myself? [23:08] asac, ^^ [23:19] fta: i think so [23:19] 23:39 < fta> https://blueprints.edge.launchpad.net/ubuntu/+spec/retaining-ubuntu-developers [23:19] 23:41 < asac> fta: ok branches are ready now [23:19] fta: i guess you have comments for that spec [23:19] asac, there's no blueprint for anything we do ? i see you have one for nm, but everything else is unrelated to our team field of activity [23:19] i would love to see something with qa & motu [23:19] also implement the unittests idea we discussed in pragues [23:19] btw, i would like to have those unit test packaged, so anyone interested could run them [23:19] 23:43 < asac> fta: i guess you have comments for that spec [23:19] i've done that for chromium [23:19] 23:49 < asac> fta: let me know if you want to test nss/nspr still [23:19] 23:49 < asac> i will take a break and then start to forward bugs. would appreciate if you could close xul 1.9.1 before that ;) [23:19] (at first because there's nothing else to see, but now i think there's something useful there) [23:19] 23:50 < asac> err s/bugs/patches/ [23:19] * Disconnected (Network is unreachable). [23:19] fta: yeah. [23:19] fta: point is we have enough specs [23:20] fta: and the bug spec was never fully drafted [23:20] yep, i've just been told the same thing in -devel for multiarch [23:20] for multiarch? [23:20] how is that related? [23:20] it's not [23:20] ah [23:20] fta: we == mozilla [23:20] yep [23:20] ok [23:20] thought you understood we == ubuntu ;) [23:21] no moz [23:21] well umoz [23:21] j [23:21] y [23:21] ;) [23:21] lies [23:23] see what happens when you use simple-patchsys [23:23] i just foudn myself editing patches directly instead of code :) [23:24] * asac runs cdbs-edit-patch [23:24] sigh ... cdbs-edit-patch just ran autotools stuff :(