[00:08] my netbook is usable again! at last [00:09] YAY [00:10] same problem as for audio (wrong perms) but with the video card :P [00:10] udev-extras was missing from UNR [00:12] http://www.itrmanager.com/articles/92695/internet-acces-mobile-continue-gagner-terrain.html [00:12] 0.4% [00:13] d'oh! (firefox-3.6:9299): Gdk-WARNING **: XID collision, trouble ahead [00:19] eeh [00:19] I know [00:19] I told you about those [00:19] when I loose keyboard [00:26] hmm. libxcb bug? [00:26] (collision?) [00:29] Add diagnostics for XID collisions [00:29] This should help with diagnosing crashes caused by over-eager XID [00:29] reuse in Xlib, see bug 581526. [00:29] Error: Launchpad bug 581526 could not be found [00:29] gnome bug 581526 [00:29] Gnome bug 581526 in gdk "XID table corruption from reuse of XIDs, resulting in leak, incorrect window destroyed status ("unexpectedly destroyed"), and crash" [Critical,Unconfirmed] http://bugzilla.gnome.org/show_bug.cgi?id=581526 [00:31] mozilla bug 467744 [00:31] Mozilla bug 467744 in Widget: Gtk "crash @gdk_window_new from moz_drawingarea_create_windows after GdkWindow 0x???????? unexpectedly destroyed" [Critical,Reopened] http://bugzilla.mozilla.org/show_bug.cgi?id=467744 [00:33] http://git.gnome.org/cgit/gtk+/commit/?id=339298b638ae76c546717f2136970b93438295a9 [01:02] it's when hovering flash [01:02] menus [01:03] not for me [01:03] but then again, I haven't tested it much [01:03] in the url i pasted above [01:07] yeah. can happen randomly [01:07] whever xids are created for windows === micahg1 is now known as micahg [02:58] * BUGabundo $ bed ; echo command not found. please try apt-get install sleepdisorder === micahg1 is now known as micahg [09:14] hi [09:47] Launchpad will be going offline for maintenance in 14 minutes. [09:47] ;) [12:31] fta: ffox rc3 build2 [12:46] !info xulrunner [12:46] xulrunner (source: xulrunner): XUL + XPCOM application runner. In component universe, is optional. Version 1.8.1.16+nobinonly-0ubuntu1 (jaunty), package size 279 kB, installed size 1020 kB [13:15] gwibber from karmic crashes [13:52] * asac lunch [14:14] transition started https://edge.launchpad.net/~mozillateam/+archive/ffox35 [14:14] now lunch [14:23] grr [14:23] [181607.641134] scim-bridge[21407]: segfault at 10 ip 00007f5413f949c6 sp 00007fff810e7e18 error 4 in libscim-1.0.so.8.2.3[7f5413f27000+d6000] [14:23] [182087.850012] evolution[32523]: segfault at 71233c20 ip 00007f9392f14e59 sp 00007fffcbbde6d0 error 4 in libgtk-x11-2.0.so.0.1702.0[7f9392e5f000+43e000] [14:23] [182719.297984] evolution[2317] general protection ip:7f6a8159d335 sp:7fff7ed8a710 error:0 in libc-2.9.so[7f6a81523000+168000] [14:24] http://paste.ubuntu.com/202881/ [15:09] hi folks, the art team are planning on shipping an incremental update to the Human theme in Karmic, but the use of a non-zero trough border in the Murrine engine seems to be causing a problem. Can someone with a free minute take a look at the report? https://bugs.launchpad.net/libgtk/+bug/327863 [15:09] Ubuntu bug 327863 in murrine-themes "non-zero GtkRange::trough-border value produces strange boxes in Firefox" [Undecided,Invalid] [15:10] I've tested the latest 3.5rc2 build in the repository and the issue still exists. No other applications exhibit problems like this, so it's most likely a xulrunner/firefox issue [15:59] psyke83: let me check [16:01] asac: thanks. The theme update isn't in the repository yet, I can give you a link to a tarball (installable via gnome-appearances-properties) if you wish to test [16:04] psyke83: is that ffox 3.5 only? [16:04] asac, nope, it's reproducible with firefox 3.0.x as well [16:04] can i see the problem without installing the new stuff? [16:04] ah ok. [16:04] asac, nope [16:04] hmm so this was reported on 1st jan [16:05] why cant i see that with the current stuff then? [16:05] psyke83: so yes, please gimme a link [16:05] asac: kwwii hasn't uploaded it yet, one sec [16:08] asac: if you're logged into ubuntuforums: http://ubuntuforums.org/attachment.php?attachmentid=118788&d=1245856072 [16:08] otherwise: http://connogriofa.googlepages.com/Humanity_v0.8.tar.gz [16:08] thoughti can drag that onto the preferences dialog :/ [16:09] asac: the Install button should do the trick [16:09] I didn't test drag'n'drop, may be a bug [16:10] note that the theme is renamed to "Humanity" in that tarball [16:12] install button crashes the dialog :( [16:13] psyke83: it installed it sucessfully, but i dont see any humanity now [16:14] oh found it [16:14] its greyed out [16:14] odd, works fine here on Karmic. If you have problems: sudo mv ~/.themes/Humanity /usr/share/themes [16:15] yes it works [16:15] i see the problem [16:16] psyke83: what does "through-border" mean? [16:17] asac, the trough-border is a property of the GtkRange widget - in other words, the spacing around the scrollbar [16:17] trough as-in, pig trough ;) [16:18] psyke83: the space between the throbber and the border? [16:19] throbber == grip [16:19] asac: yes, that's [16:19] *that's it [16:19] hmm [16:19] but the main problem is below the listbox [16:20] here https://bug471789.bugzilla.mozilla.org/attachment.cgi?id=355041 [16:20] there is the CC List: [16:20] if you look here: https://bugzilla.mozilla.org/show_bug.cgi?id=471789 [16:20] there are not CC folks filled in [16:20] Mozilla bug 471789 in Shell Integration "Strange boxes created by GTK" [Normal,Verified: worksforme] [16:20] there is the scrollbar [16:20] asac: yes, but those gtkentry boxes sometimes have horizantal scrollbars when the text overflows, so I think that an "invisible" scrollbar is getting drawn [16:20] and the listbox with all the CCed names [16:20] the main problem is below that listbox and not the scrollbar [16:21] ah [16:21] asac, an example: http://ubuntuforums.org/showthread.php?s=c5d40fe278ee41a121e651ac3ee607b4&t=1130582 [16:21] reduce your firefox window to a smaller size and pay attention to the CODE boxes [16:21] i have never seen a horizontal scrollbar in the comment box though [16:22] i dont see it on the vertical side where a scrollbar would be [16:22] CODE boxes? [16:22] where is that? [16:23] sorry, the window size isn't important. Look at "Part C", at the code boxes [16:23] the code box underneath "i386 users:" and "amd64 users" where the text scrolls [16:24] contrast that to the other code boxes which do not need a horizontal scrollbar - it seems that the trough-border is being drawn in the cases where a scrollbar isn't necessary (in code boxes with short lines) [16:26] psyke83: do you refer to this screenshot: https://bug471789.bugzilla.mozilla.org/attachment.cgi?id=355040 ? [16:26] * asac looked at the bugzilla one [16:27] asac: I didn't file the bug, so that's not my testcase. I'll navigate to that site and test [16:27] psyke83: just say me where you are looking at ;) [16:28] asac: here: http://ubuntuforums.org/showthread.php?s=c5d40fe278ee41a121e651ac3ee607b4&t=1130582 [16:28] skip to "Part C" [16:28] (of post #1) [16:28] oh ... ok ;) [16:28] sorry, I though I sent the link ;) [16:28] *thought [16:28] not sure if you did. maybe further above ;) [16:29] the testcase image isn't appropriate, because that website requires a login to display the affected elements. At least this forum post can be seen by both of us easily enough [16:39] psyke83: do you only see this on horizontal scroolbars? [16:40] asac: it appears so, yes [16:45] asac: since this is a quirk only in Firefox, I was going to add a hack to the gtkrc (setting the trough-border to 0 for GtkEntry cases in Firefox); unfortunately, you can't change the GtkRange properties in Firefox without changing *all* the elements (including the actual browser scrollbars) [16:48] *I meant GtkRange cases in Firefox [19:02] hmm. thought we had a bug for ubufox being broken in ffox 3.5 [19:02] ah bug 347972 [19:02] Launchpad bug 347972 in ubufox "Does not work with Shiretoko Web Browser (Firefox 3.5)" [Undecided,Triaged] https://launchpad.net/bugs/347972 [19:02] so tell me why launchpad doesnt find it if i search for 3.5 in ubufox package [19:02] pfft [19:07] my public tab is empty in gwibber :( [19:07] yes everything is empty here [19:07] (karmic version) [19:07] also it crashes [19:07] ;) [19:07] guess its webkit or something [19:07] fta: bug 391735 [19:07] Launchpad bug 391735 in malone "bug search returns no hit when searching for a version like "3.5"" [Undecided,New] https://launchpad.net/bugs/391735 [19:08] i cant believe it, do you see the same? [19:08] this is really annoying ;) [19:09] tse [19:09] i get bug 308397 [19:09] Launchpad bug 308397 in ubufox "localizable general.useragent.locale pref shipped by ubufox breaks mozilla OS usage stats (ubuntu not counted)" [Critical,Fix committed] https://launchpad.net/bugs/308397 [19:09] 1 hit: bug 308397 [19:09] but on that bug there isnt even any 3.5 === rickspencer3 is now known as rickspencer3-afk [19:44] BUGabundo, is your public tab empty in gwibber? [19:45] nope [19:45] but I haven't fecthed the last updates fta [19:45] it's empty since it 1st appeared [19:46] let me check for update [19:51] asac, d'oh! http://launchpadlibrarian.net/28293831/buildlog_ubuntu-intrepid-amd64.xulrunner-1.9.2_1.9.2~a1~hg20090624r29543%2Bnobinonly-0ubuntu1~umd1~intrepid_FAILEDTOBUILD.txt.gz [19:53] fta: which commit did this? [19:54] r29543 [19:54] ~ mozilla 427715 [19:54] Mozilla bug 427715 in Security: PSM "nsCryptoHash apparently being called while NSS is in shutdown state [@ NSSRWLock_LockRead_Util]" [Critical,Resolved: fixed] http://bugzilla.mozilla.org/show_bug.cgi?id=427715 [19:56] fta: guess they are on it given that their trees are burning ;) http://tinderbox.mozilla.org/showbuilds.cgi?tree=Firefox [19:56] yep, i didn't even bother trying to patch it [19:57] i 1st thought it was our system nss, but it's different this time [20:05] asac: fta: did you guys know http://mozillaca.com/ ? [20:05] a laconica for mozilla devs [20:06] nope [20:06] but i have enough with identi.ca [20:07] your last dent appeared while i was fragging a bunch of aliens, making me lag and die [20:10] err back [20:10] so did you ? [20:24] asac: re bug #327863, do you think that this issue will get resolved in Karmic without upstream intervention? [20:24] Launchpad bug 327863 in murrine-themes "non-zero GtkRange::trough-border value produces strange boxes in Firefox" [Undecided,Invalid] https://launchpad.net/bugs/327863 [20:24] psyke83: please dont hide it. I will escalate it [20:24] e.g. dont work around [20:25] if you know a workaround keep that in mind and maybe comment in the bug how that can be done, but hiding that now, will remove the pressure from us fixing it upstream [20:26] asac: I'm not going to add any hacks to the gtkrc, don't worry. I think that Ken Wimer may be apprehensive to use the theme update as-is if it exposes this bug... [20:26] asac: I would like for the theme to be pushed into the repository as-is, and having the bug exposed will encourage the bugfix to come quicker ;). [20:27] if it's not fixed by release, we can push an update to the theme that sets the trough border back to 0 (the current default for the older version of the theme) [20:28] psyke83: yes. thats what i would think is best. i will check with kwwii if you want [20:28] asac: I'll send him an excerpt of this log so he's kept informed [20:28] thanks for taking a look at this issue, I appreciate it [20:29] psyke83: yeah. ping me in case you have problems getting this theme update in [20:31] asac: will do, thanks [20:35] hm, FIREFOX_3_5rc3_RELEASE [20:35] asac, fta ping? [20:35] and SEAMONKEY_2_0a3_RELEASE.. [20:35] fta: didnt you get my ping earlier today? [20:35] NCommander, yep? [20:35] asac, not sure, i was busy with work [20:35] fta, asac, care to look at the new TB backtrace :-/ [20:35] 13:31 < asac> fta: ffox rc3 build2 [20:36] (we have a memory corruption issue in fontconfig thats been kludged around ATM) [20:36] NCommander: yes [20:36] i am eager ;) [20:36] https://bugs.edge.launchpad.net/ubuntu/+source/thunderbird/+bug/385325 [20:36] Ubuntu bug 385325 in thunderbird "[armel] thunderbird-bin crashed with SIGSEGVI" [Medium,Confirmed] [20:36] asac, BTW, where was that se of fontconfig patches again, maybe I can get lucky and someone else has alrady fixed fontconfig for me [20:37] NCommander: you mean the behdad tree? [20:37] asac: ahhah [20:37] NCommander: http://cgit.freedesktop.org/~behdad/fontconfig/ [20:37] asac, thats it; I figure its a worthwhile shot to see if if that fixes it [20:37] NCommander: try tag 2.6.99.behdad.20090601 [20:37] yes definitly worth a try (doesnt cost much to spin that somewhere i guess) [20:37] asac, thanks, hopefully I'll get lucky [20:38] http://cgit.freedesktop.org/~behdad/fontconfig/snapshot/fontconfig-2.6.99.behdad.20090601.tar.gz [20:38] asac, as it stands, we now have a issue in the JS library [20:38] is a tarball ... not sure if the packaging needs a dist-tarball though [20:38] asac, nope, but I can work with this [20:42] http://isthetreegreen.com/ [20:44] Your browser doesn't support cross-domain XMLHttpRequests. :-( [20:44] LOLOL [20:44] actualy its running noscript [20:44] LOL [20:45] argh [20:45] launchpad search [20:45] its soooo annoying ;) [20:45] lol [20:45] use google [20:46] Mook_sb, http://launchpadlibrarian.net/28299655/buildlog_ubuntu-hardy-i386.songbird_1.3.0~a~svn20090624r14060-0ubuntu1~usd1~hardy_FAILEDTOBUILD.txt.gz [20:46] bug 291760 [20:46] Mook_sb, only on hardy [20:46] Launchpad bug 291760 in network-manager "network-manager roams to (none) ((none)) - background scanning" [Undecided,Confirmed] https://launchpad.net/bugs/291760 [20:46] fta: thanks, looking [20:46] asac, so any ideas? [20:46] system gstreamer too old? [20:46] NCommander: i am not there yet ;) [20:47] asac, ah ;-) [20:47] chatting on three fronts besides this one [20:47] fta: sounds like it [20:48] NCommander: hmm #2 0x40a64cf0 in FcPatternObjectAddString (p=0x0, object=1, s=) at fcpat.c:664 [20:48] still optimized out [20:48] google finds http://lists.kde.org/?l=kde-multimedia&m=121380268225018&w=2 which says added in -base 0.10.20 [20:48] asac, no, not that backtrace [20:48] asac, the one after it :-P [20:48] fta: so your claim of base >= 0.10.7 probably isn't going to be happy :) [20:49] (fontconfig got built with optimization by accident on that run) [20:50] Mook_sb, hardy has 0.10.18 [20:50] fta: yeah, that's what it appeared to install; so you're right, too old [20:50] damn [20:51] ah good you made the debugging printfs [20:52] asac, that's what fixed it unfortunately [20:52] asac, which suggests either race condition or memory corruption [20:53] asac: another crazy FF discussion bug 387822 [20:53] Launchpad bug 387822 in firefox-3.0 "Clicking on Firefox address bar doesn't select all text" [Undecided,Won't fix] https://launchpad.net/bugs/387822 === micahg1 is now known as micahg [20:53] NCommander: so now that backtrace is reproducible? [20:54] asac, hrm? [20:54] asac, the new one is, which is what I got when I added debug printf *sigh* [20:54] NCommander: yes. but did you run it multiple times? [20:54] NCommander: or just once? [20:55] Oh, yes [20:55] Its reproducable, hangs in the same spot [20:55] asac, the behdad fontconfig causes the old trace to be reproduced [20:56] NCommander: + print ("Encountered an FcTypeVoid\n"); [20:56] did it actually built? [20:56] Yes, it did [20:57] Eek [20:57] How it did is a different question [20:57] you can see some of the printf's in the trace I put [20:57] yes. i think there is something busted. try to do it again ;) [20:57] hmm [20:57] print() must be defined somewhere [20:58] where do you see the output? [20:58] or which print are you seeing ? [20:59] returning [20:59] The value of the pointer returned [20:59] Yeah [20:59] I should probably make my printfs a little clearer [20:59] but I didn't plan to post the patch [21:00] yeah. so i guess after fixing the printf ;) .... and it still crashing there a valgrind might give more hints [21:01] asac, no valgrind on ARM [21:01] :-) [21:01] don' you love this architecture? [21:03] NCommander: you use gcc-4.4 i guess? [21:03] asac, mostly, I remember I ended up building bits with 4.3 (of TB) due to a broken compiler [21:04] desparate try, but maybe gcc 4.3 is better [21:04] for fontconfig? [21:04] for all tbird [21:04] * NCommander wants to find a good place to hurl [21:04] NCommander: did you do a clean respin? [21:04] (after switching to gcc-4.3)? [21:04] asac, not yet, that takes hours [21:04] I was going to let it run overnight [21:04] there are odd warnings [21:05] asac, why do you think 4.4. is responsible? [21:06] i dont thihnk. just want to rule out that some 4.4 regression kills us here (and that you have a half 4.4 half 4.3 build) [21:07] ###!!! ASSERTION: id differs from id in atom table!: 'd_val == idval', file xpcwrappednativejsops.cpp, line 1046 [21:07] things like that look kind of odd to me [21:07] asac, I was able to find someone else who was able ot trip a similar assertion [21:07] "This is pretty much always bad. It usually means that native code is [21:07] making a callback to an interface implemented in JavaScript, but the [21:07] document where the JS object was created has already been cleared and the [21:07] global properties of that document's window are *gone*. Generally this [21:08] indicates a problem that should be addressed in the design and use of the [21:08] callback code.K" [21:08] asac, http://groups.google.com/group/mozilla.dev.general/browse_thread/thread/afdae3bc27a25eeb [21:13] mozilla bug 426563, [21:13] mozilla bug 426563 [21:13] Mozilla bug 426563 in XPCOM "arm xptc stubs need to explicitly place symbols in text section" [Normal,Resolved: fixed] http://bugzilla.mozilla.org/show_bug.cgi?id=426563 [21:13] "Using GCC 4.2, the assembly generated from the arm xptc stubs doesn't [21:13] necessarily have any code emitted before the SharedStub symbol, causing it to [21:13] end up in the wrong section -- often the debug section, which doesn't get [21:14] mapped and so causes a segfault. [21:14] " [21:14] NCommander: does that patch apply to tbird? [21:14] It applied fuzzy [21:14] Testbuilding now [21:14] mozilla bug 476903 [21:15] Mozilla bug 476903 in XPCOM "Setting nsNavHistoryQueryOptions.sortingMode to any value raises exception on Fennec" [Normal,Reopened] http://bugzilla.mozilla.org/show_bug.cgi?id=476903 [21:15] * asac searches for ARM fixes in xpcom [21:15] oh [21:15] mozilla bug 476903 [21:15] ubottu: wake up [21:15] Sorry, I don't know anything about wake up [21:15] mozilla bug 476903 [21:15] mozilla bug 339783 [21:15] Mozilla bug 339783 in Build Config "stop shipping sherlock plugins" [Normal,Resolved: fixed] http://bugzilla.mozilla.org/show_bug.cgi?id=339783 [21:16] ugh [21:16] I think I'm going to have to take a look at that handwritten ASM [21:16] It might be an OABI/EABI issue again [21:16] mozilla bug 476903 [21:17] Mozilla bug 476903 in XPCOM "Setting nsNavHistoryQueryOptions.sortingMode to any value raises exception on Fennec" [Normal,Reopened] http://bugzilla.mozilla.org/show_bug.cgi?id=476903 [21:17] wake up ;) [21:17] ah ;) [21:17] strange [21:17] the commit says: "summary: b=476903; additional ARM xptc marshalling fixes (avoid unsigned char dependency); r=bsmedberg" [21:17] so take that patch too if it applies [21:17] NCommander: ^^ [21:17] there are two patches actually [21:18] yay. how much i love xptcinvoke ;) [21:18] but i really think this could be it ;) [21:18] having bad xptcinvoke will cause massive corruption most likely ;) [21:19] so: https://bug476903.bugzilla.mozilla.org/attachment.cgi?id=363974 https://bug476903.bugzilla.mozilla.org/attachment.cgi?id=368529 [21:19] NCommander: ^^ [21:19] take those two on top of what you applied above [21:22] 368529 does not apply [21:22] neither apply [21:23] ok lest see what landed before that in bonsai [21:23] hg history doesnt go back to branchpoint of 3.0 [21:24] asac, this is TB2, remember [21:24] TB3 is known to work fortunately [21:24] God help us if we had to port the entire stack [21:24] http://bonsai.mozilla.org/cvsquery.cgi?treeid=default&module=all&branch=HEAD&branchtype=match&dir=&file=mozilla%2Fxpcom%2Freflect%2Fxptcall%2Fsrc%2Fmd%2Funix%2Fxptcinvoke_arm.cpp&filetype=match&who=&whotype=match&sortby=Date&hours=2&date=all&mindate=&maxdate=&cvsroot=%2Fcvsroot [21:24] NCommander: right [21:24] thats why we have to look in bonsai [21:24] the branch for ffox 3 was done long ago and isnt in mercurial [21:25] so look what isnt applied from the list on that link [21:25] and apply all that ;) [21:25] then continue with the patches from above [21:25] (this MUST make them apply) [21:25] 2006-12-05 19:00 looks promising [21:25] Ugh, that sounds like overkill [21:25] When was TB2 branched?! [21:25] (i.e, I don't know if I need code from say 1999 ;-)) [21:25] NCommander: the branch was created for ffox 1.5 [21:25] NCommander: there are just a few commits [21:25] let me check [21:26] I just want a time line to work again [21:27] NCommander: http://bonsai.mozilla.org/cvsgraph.cgi?file=mozilla/xpcom/reflect/xptcall/src/md/unix/xptcinvoke_arm.cpp&rev=1.13 [21:27] so seems 1.8 was the first commit [21:27] that didnt go directly in the tbird 2 branch [21:28] http://bonsai.mozilla.org/cvsquery.cgi?treeid=default&module=all&branch=MOZILLA_1_8_BRANCH&branchtype=match&dir=&file=mozilla%2Fxpcom%2Freflect%2Fxptcall%2Fsrc%2Fmd%2Funix%2Fxptcinvoke_arm.cpp&filetype=match&who=&whotype=match&sortby=Date&hours=2&date=all&mindate=&maxdate=&cvsroot=%2Fcvsroot [21:28] thats what happened on tbird 2 branch (after the branching) [21:28] so that matches [21:28] that everything after 1.7 from the first line isnt in yet [21:29] seems 1.8 was cherry picked to 1.8 branch too [21:29] * NCommander feels like puking [21:29] so everything after that is a candidate [21:30] I don't like playing grope the patches [21:30] That's how we got sidetracked with the NSPR patch [21:30] so mozilla bug 322806 [21:30] Mozilla bug 322806 in XPCOM "ARM crash [@ XPTC_InvokeByIndex]" [Critical,Resolved: fixed] http://bugzilla.mozilla.org/show_bug.cgi?id=322806 [21:31] and mozilla bug 339783 [21:31] Mozilla bug 339783 in Build Config "stop shipping sherlock plugins" [Normal,Resolved: fixed] http://bugzilla.mozilla.org/show_bug.cgi?id=339783 [21:31] hmm [21:31] that bug sounds wrong :( [21:32] asac, to my knowledge, TB2 works fine on old ABI ARM systems [21:32] (Debian port arm) [21:32] what a mess. wrong bug id in commit [21:32] NCommander: yeah [21:32] Ugh [21:32] so look at http://bonsai.mozilla.org/cvsquery.cgi?treeid=default&module=all&branch=HEAD&branchtype=match&dir=&file=mozilla%2Fxpcom%2Freflect%2Fxptcall%2Fsrc%2Fmd%2Funix%2Fxptcinvoke_arm.cpp&filetype=match&who=&whotype=match&sortby=Date&hours=2&date=all&mindate=&maxdate=&cvsroot=%2Fcvsroot [21:32] Who the heck makes a library platform specific [21:32] and pick the commit 1.9 [21:32] Its like [21:32] and 1.13 [21:33] i think 1.10 is not relevant for tbird 2 ... seems to be a follow up commit on some more general changes they did to 1.9 branch [21:33] I hate to be dense, but why do you think its an xpcom issue, it looks like its a JS issue [21:33] * NCommander knows more about Mozillas underpinnings than I care to, but not so much on the relationship [21:33] ^ between modules [21:34] NCommander: well. if xpt is broken you are doomed to get bad corruption [21:34] point very well taken [21:34] because thats basically implementing how functions are called [21:34] so before xpt is known to work well, it doesnt make much sense to look furthre [21:34] Is there an xpt test suite [21:34] Or can I keep dreaming? [21:35] also almost all arm specific (and other minority arch) fixes i saw were either alignment or xpt issues [21:36] so if you look at current arm patch we have its: [21:36] xulrunner/xpcom/reflect/xptcall/src/md/unix/xptcinvoke_arm.cpp [21:36] and xulrunner/xpcom/reflect/xptcall/src/md/unix/xptcstubs_arm.cpp [21:37] and looking at the commit messages from bonsai it sounds really like there are issues ;) [21:38] so i would think applying: [21:38] http://bonsai.mozilla.org/cvsview2.cgi?diff_mode=context&whitespace_mode=show&subdir=mozilla/xpcom/reflect/xptcall/src/md/unix&command=DIFF_FRAMESET&file=xptcinvoke_arm.cpp&rev1=1.8&rev2=1.9&root=/cvsroot [21:38] and http://bonsai.mozilla.org/cvsview2.cgi?diff_mode=context&whitespace_mode=show&subdir=mozilla/xpcom/reflect/xptcall/src/md/unix&command=DIFF_FRAMESET&file=xptcinvoke_arm.cpp&rev1=1.12&rev2=1.13&root=/cvsroot [21:38] yes, there is a series of xptcall tests [21:38] not sure if http://bonsai.mozilla.org/cvsview2.cgi?diff_mode=context&whitespace_mode=show&subdir=mozilla/xpcom/reflect/xptcall/src/md/unix&command=DIFF_FRAMESET&file=xptcinvoke_arm.cpp&rev1=1.9&rev2=1.10&root=/cvsroot is 1.9 specific [21:38] I wonder if I could just take xptcinvoke_arm.cpp from TB3, plot it into TB2 and hope for the best [21:39] *plop [21:39] NCommander: i wouldnt think you can. they probably changed a bit of api or something [21:39] though I thought there was more than http://mxr.mozilla.org/seamonkey/source/xpcom/reflect/xptcall/tests/ [21:40] NCommander: 1.10 commit was: Bug 361533 arm port is broken XPTC_PUBLIC_API / XPTC_InvokeByIndex weren't updated for arm [21:40] Launchpad bug 361533 in evolution-exchange "evolution-exchange-storage crashed with SIGSEGV in strcmp(): crash when clicking plugin in plugin Manager (dup-of: 145287)" [Undecided,New] https://launchpad.net/bugs/361533 [21:40] Launchpad bug 145287 in evolution-exchange "evolution-exchange-storage crashed with SIGSEGV in strcmp()" [Medium,Triaged] https://launchpad.net/bugs/145287 [21:40] patch by romaxa@gmail.com r=timeless moa=timeless (this is ports only) [21:40] that suggests that there were changes on 1.9 that probably touched more than just that file [21:40] grumble [21:40] NCommander: does ffox 3.0 work? [21:40] there was an XPTC_Invoke change, yes [21:40] if so, really try just these two: [21:40] Bug 361533 arm port is broken XPTC_PUBLIC_API / XPTC_InvokeByIndex weren't updated for arm [21:40] oops [21:40]