[01:19] asac: i was wondering about CVE-2008-4066 for icedove/iceape. Do we need something like https://bugzilla.mozilla.org/attachment.cgi?id=356697 backported? Just wondering, because i couldn't find the original bugnumber (448166) for CVE-2008-4066 (second part of MFSA 2008-43) in the series file [09:05] white: https://bugzilla.mozilla.org/attachment.cgi?id=356697 [09:05] which bug number is that? [09:06] the patch looks familiar [09:24] white: ok found the bug number in your line ;) [09:25] white: yes, thats why stransky has asked for approval1.8.0.next.... e.g. it will be in next patchset unless we hit regressions [09:26] white: also mozilla bug 316859 (regression) and the original escape patch (so three for this one in total) [09:27] Mozilla bug 316859 in XPCOM "undefined symbol in components/libhtmlpars.so" [Critical,Resolved: fixed] http://bugzilla.mozilla.org/show_bug.cgi?id=316859 [09:28] hi [09:44] fta: your Ubuntu 8.10 PPA's firefox-3.1 requires libasound2 1.0.18 but Ubuntu 8.10 only has 1.0.17; is this intentional? Should users of your PPA install 1.0.18 manually? === iaindalton is now known as iain|away [10:04] iain|away: you sure you are using intrepid lines for PPA? [10:05] if so, then fta probably added libasound2 to his archive ... which makes it hard to not pull it in [12:06] iain|away, asac_: i didn't do anything special with libasound2, you may be using the wrong apt line [12:11] yeah .... i assume so too [12:16] hi, can someone see me (hear me)? [12:16] there is someone there? [12:25] sigh [12:25] those folks that just quit [12:25] i answered again for nothing in query [12:34] oh PPA's are now signed ;) [12:36] yep, i µblogged about that yesterday [12:37] hmm ... probably too much traffic in my gwibber backlog ;) ... i didnt see that particular entry of yours [12:40] i noticed because the lp guys used me as an example to explain the feature to users. so yesterday, my xchat was blinking like a Christmas tree [12:41] asac_, http://identi.ca/fta [12:48] nice ;) [13:01] sigh sound is gone again [13:03] hmm ... now everything is busted ... typing this from console now [13:03] radeon ... missing symbol :/ [13:04] * asac_ runs update;upgrade in hope [13:04] pulse 0.9.14 coming [13:04] yay [13:05] could this be what i want? [13:05] i doubt this will cure radeon driver though [13:09] hah ... X is back ;) [13:57] hi. i'm having issues trying to get a package to link with xulrunner-1.9-dev. it links fine with libxul-dev, but with xulrunner-1.9-dev (linking using flags from mozilla-gtkmozembed-embedding), ld throws a hissy fit about undefined references and whatnot. what should i do to get it to link? [13:57] hyperair: yes. [13:57] hyperair: if its standalone app use the standalone glue [13:58] if its xpcom compnent or plugin use the dependent glue [13:58] asac_: yes i'm using the standalone glue. -lxpcom_glue or something [13:58] it uses gtkmozembed stuff [13:58] standalone glue == libxul-embedding(-unstable) [13:58] and all the gtk_moz_embed_* functions are undefined [13:58] dependent glue == libxul(-unstable) [13:58] asac_: tried with that, doesn't work eitehr [13:58] libxul-unstable won't link without -R [13:59] libxul-embedding-unstable won't link at all [13:59] hyperair: use standalone glue ... you also need some code magic [13:59] asac_: what kind? [13:59] https://developer.mozilla.org/en/XPCOM_Glue [14:00] let me check [14:00] hyperair: https://wiki.ubuntu.com/XulrunnerGecko [14:00] its a bit outdated but should give you the basics [14:00] hyperair: and gives example patches on how various other apps had to be fixed [14:01] look at the yelp example for instance [14:01] if you are using unstable headers use 1.9.0.* as max version (not 1.9.*) [14:02] yeah i'm using the unstable headers [14:02] i mean the package requires headers from unstable/ [14:02] thats what i ment yes. [14:02] 1.9.0.* is your maxversion then [14:03] e.g. for GREVersionRange [14:03] (or if yoiu link against 1.9.1 use 1.9.1b1 and 1.9.1.* as bounds [14:03] i see [14:04] important is the glue initialization code as well as +#include [14:04] the rest is more or less the same [14:04] where does gtkmozembed_glue.cpp come from? [14:04] it's also in /usr/include? [14:04] yes [14:05] i see [14:05] thats also in unstable i think .... as all gtkmozembed is unstable ;) [14:05] right [14:05] okay.. [14:05] i'll give it a go [14:05] enjoy [14:06] meh. i've spent over 24 hours working on this package. the last thing i'd do is enjoy, lol [14:06] heh ... success is in reach ;) [14:06] and finally you found where to ask :-P [14:06] thats a win [14:06] i sure hope so [14:06] heh yeah [14:07] some guys at #ubuntu-motu guided me [14:07] i found some blog posts by you, so if all else failed, i was going to start emailing and/or pinging you on IRC =p [14:07] i know ... its just that you left so i couldnt answer there [14:07] ah i did? [14:08] you left -motu channel ... yes :) [14:08] hmm when was that [14:08] hyperair: not sure ... some time between you asking and me opening my screen this morning ;) [14:08] could be 1 or 6 hours :) [14:08] i was having some connectivity issues. everytime i enter a certain lecture, my wireless goes haywire. almost as if somebody's aireplaying me off the net [14:08] must be during that time [14:09] sure... no problem. i didnt complain ;) [14:09] i figured i was throwing up a lot of spurious disconnect/reconnect/nick collision messages, so i quit [14:09] =p [14:09] heh. ok [14:09] i dont see QUITS/JOINS [14:13] hmm [14:13] must be when i went to sleep then [14:13] but that's over 12 hours ago [14:13] maybe not [14:14] or maybe i was shifting location [14:14] anyway, what's the difference between xpcomglue and xpcomglue_s? [14:24] hyperair: the xpcomglue includes the standalone pieces [14:25] asac_: what standalone pieces? [14:26] hyperair: http://mxr.mozilla.org/mozilla-central/source/xpcom/glue/standalone/ [14:28] what does gtkmozembed startup code look like minus the glue? [14:28] i can't seem to figure out where to put the XPCOMGlue startup stuff [14:29] hyperair: do that before you do anything else with xpcom [14:30] dont understand your question i guess [14:30] hyperair: depends on how its done [14:30] look for gtk_moz_embed_push_startup (); [14:30] i can't find it [14:30] or gtk_moz_embed_set_comp_path (MOZILLA_HOME); [14:30] which is strange [14:30] also not there [14:30] gtk_moz_embed_set_pat [14:31] hyperair: maybe your app uses stuff like XRE_* ? [14:31] XRE_InitEmbedding or something like that [14:32] but that isnt good ... if you use gtk_moz_embed symbols you really should push/pop startup stuff ;) [14:32] err [14:32] @_@ [14:33] found it? [14:33] okay, a list of functions starting with gtk_moz_embed which are used... can_go_back, can_go_forward, get_js_status, get_link_message, get_location, go_back, go_forawrd, load_url, new, reload, set_chrome_mask, stop_load [14:34] all prefixed with gtk_moz_embed_ [14:34] are any of them relevant? [14:34] hyperair: and XRE_ ? [14:34] none [14:34] NS_Ini ? [14:35] (as a substring) [14:35] NS_InitXPCOM2 (getter_AddRefs (sm), directory, nsnull); [14:35] would that be it? [14:35] no wait [14:35] shit [14:35] that's in gecko.m4 [14:36] how do gnome-panel applets get started? [14:36] i mean this looks like an ELF executable [14:36] hyperair: gnome-panel applets get started by gnome-panel ;) [14:37] but i don't see a main() anywhere [14:37] uh gee, i think i know that much. more like what functions are there? [14:37] its a bonobo activation thing i guess [14:37] =p [14:37] hmm [14:37] hyperair: i think the NS_InitXPCOM thing is a good point to start [14:38] and see in which function thats called [14:38] asac_: it's test code inside m4/gecko.m4 [14:39] hmm it seems everything begins from init_libs [14:39] void init_libs (int argc, char *argv[]) [14:40] i guess i'll just toss the xpcom init stuff there and hope it works [14:41] hyperair: so you try to fix the testcode now? [14:41] hyperair: look at epiphany or yelp gecko.m4 [14:41] eh? [14:41] test code? [14:41] that might be a good start [14:41] no the test code doesn't really do much does it [14:41] 5:38 < hyperair> asac_: it's test code inside m4/gecko.m4 [14:41] i mean it doesn't fail [14:41] hmm [14:41] probably doesnt run anthing then [14:42] (pure luck) ;) [14:42] so if it clears ./configure, i don;'t really have to bother right =p [14:42] heh [14:42] well [14:42] i patched a few things [14:42] it was complaining about needing gtk2 [14:42] because some default toolkit macro thing was set to cairo-gtk2 [14:42] hyperair: well. thts just installing a -dev package ;) [14:42] so i replaced strcmp with strstr [14:42] ah [14:42] dumb [14:42] yep [14:43] hyperair: i would suggest to take a look at latest ephy or yelp gecko.m4 [14:43] maybe you can even just copy that over [14:43] might need a few adaptions in main configure.in i guess though ... but well [14:43] i'm packaging this thing, so i'd rather not end up with a humongous patch [14:43] =p [14:44] hyperair: patch properly, and send upstream [14:44] hmm [14:44] i mean sending them a "more" modern gecko.m4 shouldnt be that bad ;) [14:44] heh yeah i guess [14:44] i saw many gecko.m4 and they are all a mess ;) [14:44] there hasn't been a release in years [14:44] i wonder if i should even bother continuing to package this =\ [14:44] yeah. that means that the gecko.m4 probably comes from 2000 [14:44] ;) [14:44] or before [14:45] hmm [14:45] 2006 [14:45] probably [14:45] how do you know;)? [14:45] http://www.kaskaras.net/vazaar/index.php <-- it says 2007 [14:45] no wait [14:45] 2006 [14:45] =p [14:45] gecko.m4 gets copied around [14:45] nobody knows where it originates form i think ;) [14:46] maybe it was once in the mozilla tree ;) [14:46] heh [14:46] seriously? [14:46] why don't they ship it? [14:46] then aclocal can just add it [14:46] nowadays there is pkg-config ;) [14:46] gecko.m4 still uses PKG_CHECK_MODULES [14:47] yes. only reason why it exists is to provide backward compatibility i think [14:47] so it seems to be something built on top of pkg-config [14:47] ah damn [14:47] so i should just convince upstream to toss out his gecko.m4? [14:47] =p [14:47] hyperair: i think evolution was : gecko.m4 -> manually; then later someone added stuff using pkg-config ... and so on [14:47] and folks copied that around on the net [14:47] evolution uses xulrunner? [14:48] i wasn't aware [14:48] no ... evolution like "how things became like they are at present" ;) [14:48] darwin ;) [14:48] oh [14:48] heh [14:48] then it should be put into aclocal-archive or something [14:49] yeah [14:49] i mean if it's so common and being copied around [14:49] whatever [14:49] its definitly a pain that its floating around [14:49] yeah [14:49] definitely [14:49] but what's more frustrating is why mozilla can't just use proper SONAMEs damnit [14:49] that would make linking so much easier [14:49] rather than using the stupid glue [14:49] hyperair: previously nobody really used the glue ;) [14:50] which caused the main issue [14:50] feels like an extremely cheap hack gone wrong [14:50] now we enforce it. which is better ;) [14:50] but painful [14:50] miserable [14:50] (like -rpath is not working anymore) [14:50] there are still apps which link on libxul-dev [14:50] its more cross-platform [14:50] particularly all the java gecko stuff [14:50] so i agree to some degree with them ;) [14:51] i don't see gtk and gtkmm folks messing around with glues [14:51] i mean i don't know how they pull it off on windows, but at least they use proper SONAMEs on POSIX systems [14:53] no clue. i think windows solution is to ship your own copy or do some registry rumbling ;) [14:54] fugly shit [14:54] X_X [14:54] one thing that one has to accept is that xulrunner is not designed to be a lib, but a runtime ;) [14:54] + and sdk [14:54] if one looks from that perspective things become saner imo [14:55] the fact that there is gtkmozembed is a different story [14:55] and its degree of messyness is understood ... which is why there is work on a new, stable abi for embedding [14:55] oh there is? [14:55] awesome [14:55] http://hg.mozilla.org/incubator/embedding/ [14:56] hyperair: thats supposed to be shipped on top of xulrunner providing the stability required to solve all the issues in the world ;) [14:56] mm cool [14:56] probably will take some time [14:57] but definitly result of years of complains combined with webkit mania [14:58] they even use libtool ;) [14:59] so hopefully gtkmozembed will die soonish ;) [15:02] then a lot of programs will have to switch [15:02] and those that are abandoned... will have to be dropped out i suppose [15:05] depends. if anyone is interested it will be ported i guess [15:05] most apps shouldnt be that hard to port imo [15:28] hmm finished patching. let's see if it builds [15:30] hehe [15:30] dont give up ;) [15:30] you are really close (i have the feeling) [15:32] hmm [15:32] okay, i'm having a whole series of compilation issues [15:32] mostly from gtkmozembed_glue.cpp [15:32] am i supposed to #include it or just compile it together? [15:33] asac_: ^ [15:33] hyperair: just include it [15:34] asac_: only in one file or in multiple files? [15:34] i added it on top of all instances of #include [15:35] hyperair: thats ok [15:36] hyperair: whats the problem? [15:36] In file included from blah blah blah /usr/include/xulrunner-bla/unstable/nscore.h:117:1: Warning: NS_HIDDEN redefined [15:36] wait those are warnings [15:37] error expected = , ; asm or __attribute__ before nsISupports [15:38] implicit declaration of GRE_GetGREPathwhatever. [15:38] hmmmmmmm [15:38] hyperair: missing an include i guess [15:38] yeah [15:43] asac_: are there any needed includes before including gtkmozembed_glue.cpp? [15:46] no. === asac_ is now known as asac [15:46] include nsXPCOMGlue.h [15:47] but thats included through glue.cpp implicitly iirc [15:51] aaaaaah i think i know what's wrong now [15:51] this is a C file [15:51] yeah ;) [15:51] not a C++ file [15:51] thats a good pit-fall ;) [15:51] meh [15:51] hyperair: the idea is to rename it [15:51] so can i still #include it? [15:52] uh rename what? [15:52] i mean i can't use C++ code inside a C file right [15:52] the c file to cpp ;) [15:52] * hyperair gapes [15:52] you can't be serious [15:52] hyperair: no. but you can use C++ for that individual file usually [15:53] renaming a file in a patch.... makes the patch twice the size of the original patch [15:53] of the original file* [15:53] you can also add a tiny .cpp file that only does the bootup [15:53] hmm [15:53] good idea [15:53] e.g. with function startup_standalone_glue [15:53] or something [15:53] what about the other .c files? [15:53] hyperair: you only need the _glue.cpp once [15:53] and only once [15:54] in only one of the files? [15:54] yes [15:57] i'll have to put extern "C" right? [16:01] yeah [16:01] #ifdef __cplusplus [16:01] extern "C" { [16:01] #endif /* __cplusplus */ [16:02] #ifdef __cplusplus [16:02] } [16:02] #endif /* __cplusplus */ [16:04] Hi, I need help [16:05] I'm a member of Ubuntu Asturian Translators, and I need help about Firefox [16:09] asturubuntu: are you the one whose email i answered a minute ago? [16:09] Marcos Alvarez ...? [16:10] no.... but I know Marcos Alvarez. [16:10] he's a member of Ubuntu Asturian Translators too. [16:10] asturubuntu: ah. i updated the bug and marked it as pet-bug (if its about the same issue) [16:10] bug 309312 [16:10] Launchpad bug 309312 in langpack-o-matic "make po2xpi aware of per-release whitelists (Was: Please update asturian translation)" [Undecided,In progress] https://launchpad.net/bugs/309312 [16:11] so will get fixed soon i guess [16:11] That's me https://launchpad.net/~malditoastur [16:11] and this is Marcos https://launchpad.net/~marcos.alvarez.costales [16:12] asturubuntu: did you test your translations yet? [16:12] https://wiki.ubuntu.com/MozillaTeam/Testing_your_translation [16:13] thanks for the link [16:13] yes. [16:14] asturubuntu: you used it already? [16:14] We have tested it [16:14] with those instructions? nice [16:14] yes. [16:15] Wait a moment (I need to think in english, arf, arf.... I want to say many things but you are more quickly than me... :-$) [16:16] Actually, we have Firefox in Asturian language. To do that, we have to place some files into /usr/lib/firefox-3.0.5/chrome [16:16] asturubuntu: go ahead ... write down what you want to ask and so on ;) [16:16] i have to make a coffee anyway ;) [16:16] i will answer your questions if you number them when back ;) [16:17] ok [16:17] arf, arf... arf.. :P [16:18] long time ago, we did the translations in Launchpad. (Firefox and Xulrunner templates) [16:19] you can see it here: https://translations.launchpad.net/ubuntu/hardy/+lang/ast [16:20] When Intrepid was released, we thought that we can use Firefox in our language. But it was no possible. [16:22] Then, Marcos Alvarez, tried doing https://wiki.ubuntu.com/MozillaTeam/Testing_your_translation , we have get a .jar! files. [16:25] You can find it here: http://code.google.com/p/softasturubuntutranslations/ If you go to this site, and download the compressed file named intrepid_tornes_ast_Firefox_3.0.x_v.0.2.zip , you can see the files that we must put into /usr/lib/firefox-3.0.5/chrome and /usr/lib/xulrunner-1.9.0.5/chrome to get Firefox in Asturian language. [16:26] We are happy, but we have to do this routine everytime that Firefox updates [16:26] asturubuntu: right [16:27] The question is: [16:27] you also have to run the second script ... which is a .xpi [16:27] that can be installed by all users then [16:27] asturubuntu: see point 6. of the instructions [16:27] that produces a .xpi [16:29] I don't know how to do that. (I need to talk with Marcos)... wait a moment [16:31] no problem. i have a meeting now ... include my nick if your question and i will glance over to answer [16:33] hi [16:33] I'm talking with marcos right know (he's in his job) [16:33] Hi asturubuntu [16:33] Hi asac [16:34] and now is here! :D [16:34] hi marcos! [16:34] I writed asac 1/2 hour ago [16:34] :P [16:34] I don't understand very well what asac is telling me... [16:34] what's? [16:35] about the 6th point in https://wiki.ubuntu.com/MozillaTeam/Testing_your_translation [16:35] i will go to red [16:35] read [16:36] marcos_ast: asturubuntu explained to me that you cop ythe jar! files to chrome/ ... with step 6 you should be able to produce proper .xpi's out of them [16:36] which would allow you to ship stuff from websites or anywhere [16:37] well, 2 persons work in this, I and Mikel [16:37] I learn create the xpi from US language [16:37] Do you need the step 6, asac? [16:38] do you like that I check it? [16:38] (Now i'm working, but in 4 hours, I can try it) [16:39] asac: IT BUILT OH MY GOD IT BUILT [16:39] * hyperair dances around the room [16:40] oh fsck it doesn't run [16:40] hehe [16:40] hyperair: now check that it doesnt crash [16:40] * hyperair deflates [16:40] it crashes [16:41] T_T what could be wrong [16:41] marcos_ast: step 6 is important as it does everything automatically for you [16:41] i dont need it to be tested [16:41] just wanted to point out that its easy for you to distribute the .xpi to asturians until its in officially [16:41] ok [16:41] ;) [16:42] then you're working in the update [16:42] I can see later [16:42] do you need something? :) [16:42] for it? [16:42] https://bugs.launchpad.net/bugs/309312 [16:42] Launchpad bug 309312 in langpack-o-matic "make po2xpi aware of per-release whitelists (Was: Please update asturian translation)" [Undecided,In progress] [16:56] ? [16:59] asac [16:59] we had a problem with the xpi in the past [16:59] hyperair: how does it crash? [16:59] because it's a bug in po2xpi [16:59] that not create fine locale/ast-ES/global/netError.dtd [16:59]