[05:14] <vish> could someone add the aurora channel to the topic?
[07:47] <chrisccoulson> vish - we should probably actually set up a dashboard somewhere with all the PPA's displayed, and just link that in the topic
[07:47] <chrisccoulson> the topic is already very long ;)
[07:47] <vish>  yea..  :)
[08:53] <vish> chrisccoulson: https://wiki.ubuntu.com/MozillaTeam/PPAs  could be a good place, it looks quite outdated too..
[13:05] <m_conley> chrisccoulson: hello, how are you?
[14:06] <chrisccoulson> hi m_conley, i'm good thanks. how are you?
[14:07] <m_conley> chrisccoulson: all is well. :)  I was wondering if you'd seen:  https://bugs.launchpad.net/unitylauncher-extension/+bug/805194
[14:07] <ubot2> Ubuntu bug 805194 in unitylauncher-extension "New message count incorrect TB 5.0" [High,Confirmed]
[14:07] <m_conley> chrisccoulson: has the Unity Launcher functionality been uploaded yet?
[14:08] <chrisccoulson> m_conley, yeah, it was uploaded to oneiric on friday
[14:09] <m_conley> chrisccoulson: hm.  Just thinking about that last comment.  Maybe that user isn't on the Oneiric alpha.
[14:09] <m_conley> chrisccoulson: alright, cool, I'll let him know. Thanks. :)
[14:42] <bhearsum> anyone else having issues with Tb not properly opening links in Fx? (Aurora PPAs for both)
[14:47] <chrisccoulson> bhearsum, i'm not using the aurora build of thunderbird. which ubuntu release are you on though?
[14:48] <bhearsum> 11.04
[14:49] <chrisccoulson> bhearsum, does it just not open links at all, or open them in the wrong application?
[14:49] <bhearsum> it doesn't open them at all
[14:49] <bhearsum> i think it's been happening ever since i started using the Tb PPA
[14:49] <bhearsum> the aurora one, that is
[14:49] <chrisccoulson> bhearsum, what sort of links are they? (ie, are they http, https, or something else)
[14:50] <bhearsum> http/https mainly
[14:50] <bhearsum> huh, that's funny....they don't work out of Gnome Terminal, either
[14:50] <bhearsum> but they do work out of Konversation
[14:51]  * bhearsum digs around in the control center
[14:51] <chrisccoulson> heh, that's odd :)
[14:51] <bhearsum> yeah
[14:51] <bhearsum> must be something GNOME/Unity-wide, though
[14:52] <bhearsum> if i change my Preferred Web Browser to Chromium, that works
[14:52] <bhearsum> but setting it to Aurora doesn't
[14:52] <m_conley> bhearsum: hm - this bug has been brought to my attention a few times, but I've never been able to reproduce
[14:52] <bhearsum> i do see my HD indicator light up when i click it, so _something_ is happening
[14:52] <m_conley> bhearsum: when you click, does something pop up in the error console?
[14:53]  * bhearsum looks
[14:53] <bhearsum> nothing
[14:53] <m_conley> Ok, then this sounds like a different bug.
[14:53] <chrisccoulson> interesting
[14:53] <m_conley> I've also occasionally had TB opening up links in FF with http://%u/
[14:54] <chrisccoulson> i have a 11.04 machine here behind me, i'll try installing aurora on there
[14:54] <chrisccoulson> m_conley, that bug should be fixed now
[14:54]  * m_conley checks his Oneric machine..
[14:54] <bhearsum> huh, /etc/alternatives/x-www-browser points to /usr/bin/seamonkey
[14:54] <m_conley> hurrr
[14:54] <chrisccoulson> bhearsum, yeah, you can ignore that ;)
[14:54] <bhearsum> oh
[14:54] <chrisccoulson> we don't use alternatives. they are pretty broken anyway
[14:55] <bhearsum> ah, ok
[14:55] <bhearsum> is there somewhere else i can look to see exactly what it's trying to execute?
[14:55] <chrisccoulson> bhearsum, possibly strace, but that will be quite slow
[14:56] <bhearsum> oh, there's not something in gconf or something?
[14:56] <bhearsum> oh, here's something
[14:56] <bhearsum> [] bhearsum@voot:/etc$ firefox https://bugzilla.mozilla.org/show_bug.cgi?id=670929
[14:56] <bhearsum> Couldn't load XPCOM.
[14:56] <ubot2> Mozilla bug 670929 in Release Engineering "release_sanity.py should halt if buildbot isn't in $PATH" [Normal,New: ]
[14:57] <chrisccoulson> bhearsum, oh, interesting
[14:57] <bhearsum> ubot2: stop stalking me! ;)
[14:57] <ubot2> bhearsum: Error: I am only a bot, please don't think I'm intelligent :)
[14:57] <chrisccoulson> i wonder if that is fallout from dropping the shell wrapper script
[14:57] <bhearsum> certainly sounds like an LD_LIBRARY_PATH problem
[14:57] <chrisccoulson> ah, bingo
[14:57] <chrisccoulson> thunderbird still sets LD_LIBRARY_PATH, whereas firefox doesn't
[14:58] <bhearsum> and "LD_LIBRARY_PATH=. firefox https://bugzilla.mozilla.org/show_bug.cgi?id=670929" works as expected
[14:58] <chrisccoulson> so i bet the firefox binary tries to load the thunderbird libxpcom.so
[14:58] <ubot2> Mozilla bug 670929 in Release Engineering "release_sanity.py should halt if buildbot isn't in $PATH" [Normal,New: ]
[14:58] <chrisccoulson> ah, that's bad
[14:58] <bhearsum> (in /usr/lib/firefox-7.0a2)
[14:58] <chrisccoulson> strace would verify that though ;)
[14:58] <chrisccoulson> but i bet that's it
[14:58] <bhearsum> i'll try strace'ing my terminal, that should be a bit quieter
[14:59] <chrisccoulson> oh, i forgot, it happens from the terminal too
[14:59] <chrisccoulson> so that doesn't make sense anymore ;)
[14:59] <bhearsum> doesn't my tests with and without LD_LIBRARY_PATH confirm the issue?
[15:01] <chr1sccoulson> bhearsum, ok, i moved to my natty desktop now
[15:01] <bhearsum> ahh
[15:02] <chr1sccoulson> time to upgrade to aurora :)
[15:02] <bhearsum> i'm going to try to repro in upstream aurora, too
[15:03] <chr1sccoulson> ah, my internet connection hates me
[15:04] <chr1sccoulson> trying to download aurora on this machine whilst uploading aurora and nightly builds from my other machine ;)
[15:04] <bhearsum> hah
[15:04] <bhearsum> how do i set my default browser to a custom installed one?
[15:05] <chr1sccoulson> bhearsum, that's with an upstream build isn't it?
[15:05] <bhearsum> the original problem?
[15:05] <chr1sccoulson> the custom installed build is just a stock upstream build isn't it?
[15:05] <bhearsum> right
[15:06] <chr1sccoulson> ah, i'll need to think about that
[15:06] <bhearsum> i just want to make sure Thunderbird is trying to use it to open links instead of the known-to-broken PPA one
[15:06] <chr1sccoulson> bhearsum, basically, you need to create a desktop file in ~/.local/share/applications/ with an Exec= line pointing to your custom install
[15:07] <bhearsum> ah, that's easy enough
[15:07] <chr1sccoulson> then you need to set the default for "x-scheme-handler/http" and "x-scheme-handler/https" in ~/.local/share/applications/mimeapps.list to point to the name of the new desktop file
[15:08] <chr1sccoulson> our build of firefox does this automatically if you choose it as the default browser
[15:08] <bhearsum> ahh, yeah, i see it in there
[15:08] <chr1sccoulson> (but that's because we build with --enable-gio)
[15:09] <bhearsum> do i need to restart anything for it to take effect?
[15:09] <chr1sccoulson> i don't think so
[15:09] <bhearsum> ah, it works with the upstream build
[15:10] <bhearsum> should i file this in launchpad?
[15:10] <chr1sccoulson> yeah, can do
[15:10] <bhearsum> huh, ubuntu-bug gave me a traceback
[15:10] <chr1sccoulson> yeah, feel free to just file it normally
[15:10] <bhearsum> k
[15:10] <chr1sccoulson> i think our apport hook is broken with the PPA builds
[15:11] <chr1sccoulson> but i've not fixed it because i'm going to completely rewrite it at some point ;)
[15:11] <bhearsum> looks like it, based on the traceback
[15:11] <bhearsum> hm, now my Aurora PPA build won't launch at all!
[15:12] <chr1sccoulson> does it give you an error if you launch it from the console?
[15:12] <bhearsum> yeah, the same XPCOM one
[15:12] <bhearsum> works fine if i run it from /usr/lib/firefox-7.0a2, of course
[15:13] <chr1sccoulson> bhearsum, could you run that with "strace -f firefox" and attach the output to the bug?
[15:13] <bhearsum> yup
[15:13] <chr1sccoulson> thanks
[15:14] <chr1sccoulson> hmmm, so, aurora seems to work here :/
[15:16] <bhearsum> =\
[15:17] <bhearsum> https://bugs.launchpad.net/ubuntu/+source/firefox/+bug/809384
[15:17] <ubot2> Ubuntu bug 809384 in firefox "Firefox Aurora PPA build fails to load XPCOM" [Undecided,New]
[15:19] <chrisccoulson> bhearsum, oh, that's odd:
[15:19] <chrisccoulson> stat("/home/bhearsum/firefox", {st_mode=S_IFDIR|0755, st_size=4096, ...}) = 0
[15:20] <chrisccoulson> bhearsum, is /usr/bin/firefox a symlink to /usr/lib/firefox-7.0a2/firefox ?
[15:20] <bhearsum> lrwxrwxrwx 1 root root 28 2011-07-11 13:16 /usr/bin/firefox -> ../lib/firefox-7.0a2/firefox
[15:23] <chrisccoulson> bhearsum, have you modified PATH in your environment at all?
[15:25] <bhearsum>  /home/bhearsum/bin is in it, but i think that's normal
[15:25] <bhearsum> (and there's no firefox in it, either)
[15:28] <chr1sccoulson> bhearsum, oh, it's the /home/bhearsum/firefox folder which breaks it
[15:28] <chr1sccoulson> i get the same issue here
[15:29] <chr1sccoulson> if i create an empty /home/chr1s/firefox folder, then it breaks :/
[15:29] <bhearsum> oh, hah!
[15:29] <bhearsum> that's a crappy bug
[15:29] <chrisccoulson> bhearsum, ok, i'll take a look at that after our team meeting
[15:29] <bhearsum> by which i mean, hard to deal with
[15:30] <bhearsum> (indeed, if i remove that folder, everything works again)
[15:30] <chrisccoulson> good catch though, that wouldn't have been too good if we pushed that in to the distro :)
[15:31] <bhearsum> np :)
[15:33] <micahg> vish: ^^
[15:46] <chrisccoulson> bhearsum, ok, i see what the problem is not
[15:47] <chrisccoulson> mozilla::BinaryPath::Get is quite broken, and it's being used at startup now since the changes to lazy load libxul
[15:47] <chrisccoulson> i'll write a patch and send it upstream once i'm done
[15:48] <bhearsum> oh, sweet
[15:48] <bhearsum> why didn't this affect upstream, then?
[15:48] <chrisccoulson> bhearsum, your custom build worked btw because you launch it with the full path (it only fails if it's loaded from PATH, so "firefox" will fail, but "/usr/bin/firefox" works)
[15:48] <bhearsum> ah
[17:03] <chrisccoulson> oh, i broke all the nightly builds, https://edge.launchpad.net/~ubuntu-mozilla-daily/+archive/ppa/+build/2622207
[18:51] <micahg> chrisccoulson: http://groups.google.com/group/mozilla.dev.planning/browse_thread/thread/51e2b5b4a209b4a3, I wonder if this would help us with space issues: http://groups.google.com/group/mozilla.dev.planning/browse_thread/thread/51e2b5b4a209b4a3
[19:45] <m_conley> chr1sccoulson: ping
[19:59] <m_conley> chrisccoulson: ping
[19:59] <chrisccoulson> hi m_conley
[19:59] <chrisccoulson> sorry about the confusing names, i've been back and forwards between machines all day ;)
[20:01] <m_conley> chrisccoulson: hey - quick question for you.  So it turns out that js-ctypes doesn't really like it when I invoke the nsPromptService from async callbacks. I need the prompt in order to get auth details from the user for address books that require user/pass.
[20:01] <m_conley> chrisccoulson: jorendorff in #jsapi came up with a patch, but it's blocked on a few other patches
[20:02] <m_conley> chrisccoulson: I've found a workaround that involves spawning my own window dialog with user/pass prompts.  My question to you:  should we prefer to backport jorendorffs patch, or should we go with my workaround?
[20:02] <chrisccoulson> m_conley, it might be easier to backport the changes. how big are they?
[20:02] <m_conley> chrisccoulson: jorendorff's patch:  https://bugzilla.mozilla.org/show_bug.cgi?id=671027
[20:02] <ubot2> Mozilla bug 671027 in js-ctypes "Assertion failure: !cx->outstandingRequests, at js/src/jsapi.cpp:6053 with reentering ctypes function callbacks" [Normal,New: ]
[20:05] <chrisccoulson> m_conley, yeah, i just read the scrollback on #jsapi
[20:06] <m_conley> chrisccoulson: thoughts?
[20:06] <janimo`> hello, can you point me to an update page describing patching mozilla packages? The links under Patching on this page are 404 https://wiki.ubuntu.com/MozillaTeam/KnowledgeBase
[20:07] <chrisccoulson> m_conley, it looks fairly trivial to backport
[20:07] <m_conley> chrisccoulson: cool, alright, i'll stick with the prompt service then.  The patch indeed fixes the problem.
[20:07] <chrisccoulson> m_conley, ok, i'll start working on that tomorrow. i've got all your other bits to backport anyway :)
[20:08] <chrisccoulson> janimo`, i'm not sure if there are any wiki pages describing what you want
[20:08] <m_conley> chrisccoulson: oh, yes, those.  :)
[20:09] <chrisccoulson> m_conley, do you think we'll be able to get people testing your extension this week?
[20:09] <m_conley> chrisccoulson: for read capability?  Sure.
[20:09] <chrisccoulson> (once i've got all the changes backported)
[20:09] <chrisccoulson> cool
[20:10] <m_conley> chrisccoulson: I'm eager to hear about all the things I forgot to implement.  :p
[20:10] <chrisccoulson> heh :-)
[20:17] <janimo`> chrisccoulson, what is the sequence of commands when patching configure.in ? I edit-patch for the actual patch, then edit-patch 99_configure.patch and in the subshell autoreconf but that fails to run
[20:23] <chrisccoulson> janimo`, there's no need to run autoreconf
[20:23] <janimo`> how do all configure files get regenerated?
[20:27] <chrisccoulson> janimo`, they're generated at build time. we don't even ship a pregenerated configure script
[20:27] <janimo`> chrisccoulson, ok, but some patches (in nspr at least) touch configure.in
[20:28] <janimo`> and there's a 99_configure patch which lead me to believe it was an autoreconf variation as done in many packages
[20:28] <janimo`> so at least for nspr the configure scripts are not upstream ones but patched
[20:28] <chrisccoulson> janimo`, oh, you're looking at nspr?
[20:29] <janimo`> chrisccoulson, trying to fix ARM FTBFS
[20:30] <janimo`> and generally hate it when after what should be a simple fix I get to fight autotools and this whole regeneration of config files. The number one thing I dislike about packaging, by far
[20:30] <chrisccoulson> i think 99_configure.patch is just the result of autoconf2.13 run
[20:31] <janimo`> as sometimes autoreconf works but now it even fails with undefined macro AC_OUTPUT_FILES
[20:31] <janimo`> whereas I think I have a new enough autoconf
[20:31] <janimo`> 2.68 namely
[20:33] <chrisccoulson> you need 2.13
[20:45] <janimo`> chrisccoulson, thanks, I assumed when seeing 2.13, 'or newer' was implied .