=== asac_ is now known as asac [05:01] Odd: I only see two songs in my music player app. [05:01] I assume /home/ume/media/sounds is where it goes? [05:01] I made that a symlink to my host OS's music. [05:09] DanaG: Do you have subdirectories in your host directory? If so, are the two songs shown in the root directory? [05:10] From testing on my install, it seems that /home/ume/media/sounds is not checked for directories [05:11] Yeah. But there are actually 4 files in the top-level dir. [05:12] Correction: 3 files. [05:12] 2 mp3 and one mod. I guess that's why, for that dir. [05:12] Very odd. I showed about 20 when I dropped them in the top-level locally. Is one of them a different codec, or not tagged cleanly? [05:12] Ah, yes. It's the .mod that doesn't show? [05:12] Yeah, that makes sense. [05:13] I have them in dirs, and that's how I organize things. In fact, that's why I have an iAudio6, instead of countless alternatives. [05:13] Oddly enough, the video player shows nothing, despite having several videos at the top level. [05:14] That makes sense: especially for devices with >10GB, organisation makes sense. Could you file a bug about that? [05:14] At the top level of media/videos ? [05:16] Right now I'm in a power outage (on a laptop, and cable modem and router on a UPS), so now's not a great time -- but where should I go file it? [05:16] Yeah, I have several files there/ Let me check the codecs. [05:16] Note: would having the thing it's symlinked to be RO mounted cause that breakage? [05:16] http://bugs.launchpad.net/ubuntu/+source/moblin-media/ should show you the current list. If it's not there, there's a link to file a bug. [05:17] does that one package apply to both the audio and video players? [05:17] I don't think blocking RO mounts makes sense. Imagine the case where someone has a USB optical drive, and wants to play videos from there. [05:17] * persia double-checks [05:18] One video is MPEG-1 video, MP1-layer2 audio (yep, obscure. It's an old video.) [05:18] DanaG: Music, Videos, and Photos. [05:18] Another is XVID MPEG-4 video, MP3 audio. The last is DivX5 video, MP3 audio. [05:19] Oh yeah, and to build the virtualbox additions, I had to install lpiacompat stuff, since lpia stuff gave "disagrees on version of symbol" (don't remember the symbol name). [06:40] Odd... wallpapers don't give scaling settings. [06:40] And my VM seems to use the wrong resolution; what resolution is supposed to be default? [06:41] DanaG: Most testing was done at 1024x600, and a fair amount at 800x480. I don't think everything works for 800x480. [06:42] Why do so many things use 1024x600? That's an odd aspect ratio. 1024x640 is 16:10. [06:43] DanaG: I suspect it's because the glass was cheap. [06:43] Oh, and how do you use a bluetooth headset with the thing? [06:43] (I've handed the VM my USB BT adapter) [06:43] Personally, I'd rather have 1024x768, but maybe that's just me. [06:44] Does the bluetooth icon show in the taskbar? [06:44] Yup. And I've even paired the headset. [06:44] oddly enough, it shows up twice. It offers both Headset and Handsfree profiles. [06:44] That means you've got speakers & a mic. [06:45] It doesn't show up in 'asoundconf list' or 'aplay -l' or 'aplay -L' [06:49] I'm not sure. I've never gotten bluetooth audio to work (although I haven't tried very hard with Ubuntu Mobile). [06:49] I seem to remember there being some alsa module that needed loading, but I'm not sure if that is still true. [06:49] You can supposedly do it with .asoundrc nowadays. [06:49] However, oddly enough, I can't even RUN the "audio service" on the Intrepid ost. [09:14] oh, aldi has netbooks today here :) [09:15] and where is here ? [09:15] germany [09:16] ah [09:16] i was hoping the netherlands :P [09:16] aldi north ... (our aldi is split into two companies) [09:40] hi [09:40] i'm running mccaslin-lpia-hardy on a Q1... would like to calibrate the screen, anyone knows how to do that? [09:41] philn: There ought be a calibration tool in the Preferences menu. === asac__ is now known as asac [09:43] where's that menu? i'm probably missing a package or fset [09:46] philn: At the top of the screen, to the right of the home button, there ought be the word "All" and an arrow. [09:47] Selecting "All" allows you to choose "Preferences", which shows some icons that are not included in "All", one of which should be the screen calibration tool. [09:48] * ogra thinks "all" should be renamed to "applications" or something if it doesnt really include "all" :) [09:49] lowers the confusion level :) [09:49] i only see "all".. nothing related to preferences, either in that drop-down menu or in the launchers list [09:51] philn: Which image is this? one from ubuntu or self built? [09:51] i used an image created with image creator [09:51] self built [09:52] philn: Do you have particular changes you care about in this image? [09:52] philn: We use the Q1Full fset when building MIC images [09:53] "crownbeach-full-mobile-stack" [09:53] yes i care a bit ;) i compiled deps for elisa on it [09:54] philn: Here's my tip: use a ppa, or pbuilder, or build packages natively instead [09:55] well it's for some kind of demo, i'll do it better next time.. for now i just want to calibrate the touchscreen [09:55] philn: Sure, I just wanted to save you the huge timesink that building images is [09:55] and elisa 0.5 is not (ready to be) packaged yet anyway.. [09:56] philn: Did you install the ubuntu-mobile meta-package? [09:56] yes [09:57] You have moblin-applets? [09:57] yes [09:57] in fact i could launch the calibration tool from terminal [09:58] but doesn't seem to work, i can't tap the damn blinking target ;) [09:58] philn: image-creator --platform-name=menlow-lpia-ume --fset crownbeach-full-mobile-stack -c list-pkgs [09:58] It's not menlow [09:58] philn: Will tell you which packages get pulled for this fset [09:59] i'm on mccaslin [09:59] You get the idea, replace platform and fset with what you care about [10:02] philn: moblin-touchscreen from terminal should do the calibration I would think [10:02] it's in moblin-applets [10:02] asac: Hey did you see my last message to bspencer and rustyl where I was saying you might send an up-to-date patch? [10:03] lool: nope [10:04] around Date: Wed, 25 Jun 2008 22:43:48 +0200 [10:04] oh [10:04] Alexander: could you push your git tree to the mbf hardy branch or to a [10:04] new branch rebased on top of master? Or anything which allows Bob to [10:04] review the final changes or xulrunner 1.0 support. [10:04] sorry, was in the middle of the msg [10:04] lool: not sure if i have the final changes ;) ... i have my changes ... we need to sync them [10:04] will get back to you later today on this. stay tuned [10:05] asac: Cool [10:05] asac: Can I grab you two secs on other xulrunner stuff? [10:05] lool: just ask ;) [10:05] asac: evolution-rss can be built against --gecko=libxul or libxul-embedding; how to pick the correct one? [10:06] asac: And the second thing is I'm stuck porting galeon to xulrunner 1.9; I've given up on it and I've asked glandium and upstream to look at my in progress patches if they'd like to complete the port, but I didn't understand how to finish it [10:07] lool: what is evolution-rss? a standalone application? [10:07] So I'm pinging you as well in case you have some secret dependance on xul 1.9 porting; I wouldn't be surprized porting apps to xul 1.9 is the drug which keeps you awake [10:07] ha [10:07] ;) [10:07] asac: It's a plugin for evolution, so I would think it should try to reuse evolution's xul if possible [10:08] lool: does evo load xul at all? [10:08] otherwise use the -embedding thing [10:08] * ogra thought evo still used gtkhtml [10:09] philn: Ah sorry it doesn't work; I saw it broken from time to time too; perhaps file a bug against moblin-applets (upstream project) or against ubuntu-mobile (upstream project)? [10:09] lool: how did galeon fail? missing symbols? [10:09] asac: Hmm indeed not sure evo uses xul at all [10:09] asac: So libxul-embedding if there's no other xul to reuse and libxul if there's one, correct? [10:09] asac: Yeah, ultimately missing symbols [10:10] But to reach this point I did things which I'm not fully understanding [10:10] All my notes are in the xulrunner 1.9 port bug reports in the Debian BTS [10:10] asac: Oh heck, I'm not sure you want to spend time on galeon [10:10] evo: gtkhtml3.14 [10:10] Everybody says it's completely useless once ported anyway [10:10] ogra: yes exactly [10:10] (from the deps) [10:10] So no xul so should use -embedding [10:30] lool: i am sorry. my connection is really bad today [10:31] lool: whats the bug id? [10:31] just want to take a quick look [10:31] but i think galeon is definitly one of the more interesting pieces of software to port [10:31] asac_: I was saying don't spend time on galeon; it's going to be useless even once ported [10:32] asac_: Many critical features don't work at all with xul 1.9 and need to be rewritten :-/ [10:32] http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=480799 [10:32] Debian bug 480799 in galeon "Don't build depend on libxul-dev" [Wishlist,Open] [10:32] asac_: There are some inprogress patches, but the bulk of the acceptable patches are in svn.debian.org/svn/pkg-galeon/unstable/galeon already [10:35] lool: if test "x$has_gtkmozembed" = "xno"; then [10:35] is that block to be ment 1.8 or 1.9? [10:35] (in your configure.in patch) [10:35] asac_: There's a later check I thikn [10:37] * lool checking out from his hardy box [10:39] lool: the svn configure patch looks ok without looking at the target source [10:39] there is no glue startup code, but i guess thats already in upstream orig? [10:39] asac_: It should be [10:39] asac_: Upstream 2.0.5 *claims* to support xulrunner 1.9 [10:39] lool: so how does it fail? [10:40] But in practice it didn't build due to API changes [10:40] at least the NSPR switch is not turned on [10:40] asac_: Well many things really [10:40] lool: ok, i'll try a more recent image.. is it ok to try one your pre-built images? [10:40] philn: definitely [10:40] philn: I mean it can't hurt [10:41] lool: if you give me the orig for the latest svn i can run it in the background and see how bad if fails ;) [10:41] lool: there? http://cdimage.ubuntu.com/mobile/releases/hardy/ [10:41] asac_: Well I've sent a combined patch of all ugly things I did, but I think it's highly incorrect [10:41] philn: Yes [10:42] asac_: Mike said I should drop the rpath setting [10:42] asac_: Do you think I should drop it unconditionally for xulrunner 1.0? [10:42] 1.9 [10:42] asac_: BTW I checked the fedora patches and these are a subset of mine; it seems they have an easier time building against xulrunner 1.9 :-( [10:42] lool: if we only support "-embedding", then yes. otherwise you need to drop if if we use libxul-embedding* [10:42] otherwise not [10:43] (for instance epiphany should work with libxul + rpath OR libxul-embedding [10:43] but embedding is definitly preferred [10:43] I think we only need to support embedding as galeon typically bootstraps the xul [10:43] asac_: Why does epiphany need to run with libxul? [10:43] lool: fedora did only care about getting things to compile [10:44] Yeah === asac_ is now known as asac [10:44] lool: it doesnt need to. chpe just refused to believe the -embedding is the right way so i coded both options to make him happy [10:44] asac_: Concerning that gtkmozembed conditional in configure.in, I thought it wasn't too important to touch as -lxul is probably added by the pkg-config flags earlier anyway and wasn't too useful [10:44] asac: arf ok :) [10:45] lool: -lxul must not be used in standalone glue either [10:45] the only lib you link against is -lxpcomglue [10:45] and maybe nspr if your need it [10:45] so the output of pkg-config --libs libxul-embedding [10:45] asac: What's hard in galeon is that they are compatible with last 10 years of xulruners [10:46] lool: same is true for epiphany [10:46] The configure and code are ifdefed all over the place, it's not only supporting the latest xul but everything [10:46] asac: Indeed [10:46] but nobody actually knows if the code still works [10:46] they keep the configure stuff and make their own life harder [10:46] asac: Except epiphany moved to some sensible m4 macros which do the work behind the scene; in galeon the configure.in script feels like a mess now [10:47] Right epiphany still looks litlle complex [10:47] lool: yeah ;) unfortunately all this appears to come from the same source. so other build systems have it too [10:47] some follow the .m4 approach of epiphany [10:47] others copy the in-configure code from whoknowswhere [10:48] So should I set MOZILLA_GTKMOZEMBED=$gecko just like _XPCOM? [10:48] And hence make the test succeed and hence don't append -lxul to the link flags? [10:48] lool: the idea is to use the PKG_CHECK_MODULES result for libxul-embedding(-unstable) for every CFLAGS and LDFLAGS [10:48] What I hate with the current approach is that it becomes unknown what serves which xulrunner versions [10:49] I wouldn't mind having 10 macros one for each version, but having a huge snippet handling all cases is a mess :-/ [10:49] lool: yeah. assume that all the manual LDFLAGS + CFLAGS things are ment for mozilla <=1.7 [10:49] maybe even << 1.7 [10:49] wow, that's old [10:49] aka 1.4 [10:49] asac: What about rpath? should I drop it? [10:50] lool: i think i inserted a configure test to check if -DXPCOM_GLUE is defined by CFLAGS [10:50] if you want to do it right, check for that and if that is defined, drop rpath [10:51] asac: Ok; one build issue I had needed me to set some XPCOM_GLUE_US_NSPR define manually to build [10:51] asac: i didn't quite understand how it was supposed to be set [10:51] lool: http://paste.ubuntu.com/24672/ [10:52] Excellent [10:52] lool: need to set CFLAGS to the CFLAGS with the xul flags of course [10:53] but i guess you know how COMPILE macros are used ;) [10:53] http://paste.ubuntu.com/ [10:54] thats more complete [10:54] you need a LANG_POP at the end i guess [10:54] asac: ELINK [10:54] err sorry: http://paste.ubuntu.com/24675/ [10:54] asac: Should I bother setting mozilla_home? [10:55] _GECKO_CFLAGS=`pkg-config --cflags $MOZILLA_XPCOM` [10:55] lool: you need MOZILLA_HOME for non-standalone glue cases [10:55] lool: in epiphany i added a --with-gecko-home=... flag [10:55] Yeah, that's what I copied last week [10:55] as you cannot get the MOZILLA_HOME anymore [10:55] Exactly, didn't find it in .pc files [10:56] so basically ... if GECKO >= 1.9 => --with-gecko-home is required for "normal" libxul [10:56] for 1.8 you can still use the old auto-guess heuristic [10:56] lool: but i wouldnt put too much work into the "libxul" part .... really [10:56] even glandium agrees that standalone glue is the idea ;) [10:57] lool: i'd say: if XPCOM_GLUE => drop -rpath and unset MOZILLA_HOME [10:57] otherwise try the auto guess ... if that fails => bad luck [10:58] asac: What should I use for CFLAGS in the general case for the -DXPCOM_GLUE? Should I only run this test if libxul-embedding? [10:58] lool: I'd suggest to just test for XPCOM_GLUE .... and drop -rpath + MOZILLA_HOME if its set [10:58] and dont bother about any corner cases [10:59] even 1.8 had a standalone glue iirc [10:59] but nobody used it ;) [10:59] asac: So which flags should I use for the xpcom glue test? special flags for this test, the "xpcom" ones, the "embedded" ones? [10:59] lool: so you always use your GECKO_CFLAGS [10:59] lool: use the MOZILLA_XPCOM ones [11:00] There's no such thing in galeon currently [11:00] (i guess for 1.9 they are the same) [11:00] Ok, will use _xpcom [11:01] in epiphany i defined a conditional for HAVE_XPCOM_GLUE so i could remove the -rpath thing from Makefile.am [11:01] if galeon adds the -rpath in configure you probably dont need it ... but i dont know [11:01] Ok, I planned doing an AM_CONDITIONAL as well [11:02] * asac tries to remember what he wanted to do ;) [11:02] work on mobile stuff indeed :) [11:02] haha [11:03] ogra: i felt quite mobile yesterday when i experienced a great 3g connection throughout my whole RegionalBahn Trip ;) [11:03] note: my cell-phone usually has no connection throughout most parts of the same trip ;) [11:03] regional ? [11:03] wow [11:03] yeah ... not even ICE [11:03] my ICEs all have wlan now [11:04] thats the advantage of living in the middle of the country :) [11:04] ogra: NM connects much quicker to 3g ;) [11:04] like in 1 second i get connection + IP [11:04] cool [11:04] now I know what marcel ment with "dhclient is real crap" :) [11:04] heh [11:05] well, its good on servers (if you even use dhcp at all there) [11:05] ogra: the lease feature is nice [11:05] the point is that there apear to be a bunch of sleeps that slow down things by seconds (iiu him correctly) [11:05] asac: What's that -dlopen self thing? [11:05] lool: where? [11:06] galeon_LDFLAGS = -R$(MOZILLA_HOME) -dlopen self [11:06] never seen that before [11:06] i guess this can be dropped [11:07] together with the -R [11:07] asac: BTW what do you plan to do with the xulrunner-dev xulrunner-1.9-dev pacakge name mismatch? [11:08] lool: I will swallow that and provide a compatibility package [11:08] lool: its unreasonable imo [11:08] and smells like intention [11:08] :-/ [11:08] checking whether we have a xpcom glue... yes [11:08] Using libxul-embedding-unstable version 1.9 [11:08] let me commit that now [11:09] yay [11:09] looks good [11:09] In file included from MozRegisterComponents.cpp:30: [11:09] ProgressListener.h:75: warning: 'GProgressListener' declared with greater visibility than the type of its field 'GProgressListener::' [11:09] anyone has lost his caret on gnome-terminal recently? [11:09] That's the one where it looks like nspr isn't properly used [11:09] mine is gnome here now ;) [11:09] just black [11:10] lool: does galeon still use MOZILLA_INTERNAL_API [11:10] http://people.ubuntu.com/~lool/galeon_2.0.5-1_amd64.build [11:10] maybe grep for that [11:10] it does [11:10] galeon-2.0.5/configure.in: [[#define MOZILLA_INTERNAL_API [11:10] yeah thats what has to go [11:10] galeon-2.0.5/mozilla/GaleonJS.cpp:#define MOZILLA_INTERNAL_API [11:10] galeon-2.0.5/mozilla/GaleonJS.cpp:#undef MOZILLA_INTERNAL_API [11:10] to use standalone glue [11:11] lool: drop that define ... and fix the build errors :-D [11:11] well ... at least you can send the better configure upstream ... they should switch to frozen linkage [11:11] ok, now the applet works, but only the 2 first targets works. i click on third and the applet thinks calibration is finished, asks me to save settings or not [11:12] lool: http://developer.mozilla.org/en/docs/Migrating_from_Internal_Linkage_to_Frozen_Linkage [11:13] the idea of frozen linkage is that you can use that, ship binaries to the world and they will always work in future ;) [11:13] I don't quite understand how internal_api is set/not set [11:13] I mean in galeon [11:13] lool: if you define MOZILLA_INTERNAL_API you get symbols that are hidden in libxul [11:13] There are many checks with mozilla_internal_api [11:13] lool: so you just drop that define [11:15] I mean, I'm trying to locate it [11:15] It's not set in configure, configure only checks whether it needs to use that flag to get some API [11:15] And it decides it doesn't I think [11:16] lool: galeon-2.0.5/mozilla/GaleonJS.cpp [11:16] explicitly defines it? [11:17] in #ifdef HAVE_NSISCRIPTCONTEXT_INTERNAL_API [11:17] ok and that is false? [11:17] Define if nsIScriptContext is MOZILLA_INTERNAL_API */ [11:17] #undef HAVE_NSISCRIPTCONTEXT_INTERNAL_API */ [11:17] in config.h [11:17] are you trying galeon 2.0? [11:17] 2.0.5 [11:18] mozilla/MozillaPrivate.cpp also uses #ifdef HAVE_NSSTRING_INTERNAL [11:18] which is undef [11:19] but you see the define during compile? [11:19] ah [11:19] mozilla/GaleonAboutModule.cpp does it unconditionally [11:19] * asac looking [11:20] ok its the same mess we had in epiphany [11:20] nsString.h => nsStringAPI.h [11:20] s/.*nsEscape.h.*/ [11:20] it's undefined [11:20] what is undefined? [11:21] MOZILLA_INTERNAL_API [11:21] in AboutModule? [11:21] for me its defined there [11:21] asac: No in MozRegisterComponents.cpp [11:21] where my build fails [11:21] oh ok [11:22] mozilla/GaleonAboutModule.cpp uses nsstring.h and #defines MOZILLA_INTERNAL_API but Im' not that far [11:23] I don't see any nsescape there either [11:23] There's one in mozilla/GaleonAboutModule.cpp [11:24] lool: i think the only field of ProgressListener that might have hidden type is GulCString [11:26] asac: BTW ATM I'm just running svn-buildpackage from the pkg-galeon tree [11:26] lool: without the improved configure? [11:26] asac: I committed the improved configure now [11:26] lool: how do i run svn-buildpackage? [11:27] asac: You run svn-buildpackage :) [11:27] seriously, from the galeon dir [11:27] you need to create ../tarballs [11:27] and drop the upstream tarball in there with the .orig name [11:27] k [11:27] (or a symlink with that name) [11:27] Or you forget about sbp and just copy the debian/ in an unpacked tarball : [11:27] :) [11:28] lool: i dont have the orig ;) [11:29] galeon.sf.net [11:29] http://sourceforge.net/project/downloading.php?group_id=6999&use_mirror=kent&filename=galeon-2.0.5.tar.gz&22444074 [11:29] it builts ;) [11:29] err, started [11:29] Cool [11:30] I would have switched to quilt would I have known I would have to pile to many patches [11:30] hehe [11:30] When I read 2.0.5 was supporting xulrunner 1.9 I thought it was a one or two hour job to perhaps adjust things for Debian/Ubuntu [11:30] How wrong I was [11:31] I spent most of a sunday on it and a couple of evenings since [11:31] well ... not sure why they wrote it. i looked at the sources at some point when they claimed xulrunner 1.9 and didnt understood ;) [11:31] Even fedora people had to patch, but not so much as we're doing it now [11:32] lool: rv = NS_NewGenericFactory(getter_AddRefs(componentFactory), [11:32] &(sAppComps[i])); [11:32] lool: does it work on fedora? [11:32] Ah I think they patched that [11:32] thats non-frozen linkage [11:32] http://cvs.fedoraproject.org/viewcvs/rpms/galeon/F-9/ [11:33] galeon-2.0.5-build-fix.patch and galeon-2.0.5-xulrunner.patch [11:33] - presShell->GetDocument(getter_AddRefs(doc)); [11:33] + doc = presShell->GetDocument(); [11:33] It's not the same, sorry [11:34] lool: let me look for the right pattern ;) [11:34] 05:23 < philipl> hgb: Ah. flashblock doesn't work, but adblock does. [11:34] 05:24 < philipl> extensions that need xul don't work. [11:34] that's from #galeon [11:36] lool: http://bugzilla.gnome.org/attachment.cgi?id=100976 [11:36] search for "genericfactory" [11:36] thats the inittial epiphany patch which still had that frozen linkage migration [11:37] i have the feeling that after fixing this galeon will fly ;) [11:37] hehe [11:37] let me know when you have committed that to svn so i can test :) [11:38] asac: Is that backwards compatible or do I need to enclose it in a fake test and tell upstream to add a test for it? ;) [11:38] lool: its backwards compatible [11:39] the NS_New thing is just a short hand which is not available in frozen linkage [11:39] lool: even if its not, upstream should polish this imo. we are working for 1.9 here only (same in debian) [11:40] but i think thats good to have everywhere [11:42] lool: some minor bad news. they dont have the xpcom glue startup code yet. that has to go to mozilla/mozilla-embed-shell.cpp [11:44] the snippet is in http://bugzilla.gnome.org/attachment.cgi?id=102329 in the embed/mozilla/mozilla-embed-single.cpp patch [11:44] asac: committed [11:44] the generic factory part [11:44] yeah [11:45] lool: does it build already? [11:45] I just kicked a build [11:45] kk [11:46] asac: Right that xul startup code I grepped for (remembered from mbf port), but didn't findin galeon [11:46] It fails at: [11:46] JSConsoleListener.cpp: In member function 'virtual nsresult JSConsoleListener::Observe(nsIConsoleMessage*)': [11:46] JSConsoleListener.cpp:53: error: 'class nsIConsoleMessage' has no member named 'GetMessage' [11:46] I think Fedora fixed that one though [11:47] asac: debian/patches/66_non-threadsafe-gconsolemessage-isupports.patch what doyou think of the change? [11:47] I copied it from the comments in the generated templates, but dropping the threadsafe part is scary [11:47] -NS_IMPL_THREADSAFE_ISUPPORTS1(GConsoleMessage, nsIConsoleMessage) [11:47] - [11:47] +NS_IMPL_ISUPPORTS1(GConsoleMessage, nsIConsoleMessage) [11:49] lool: GetMessage is now GetMessageMoz [11:49] Yeah [11:49] ok [11:49] I just changed that, testing new patch [11:49] lool: dont drop the THREADSAFE thing [11:49] * lool thanks his amd64 and ccache [11:49] if we use XPCOM_GLUE_WITH_NSPR [11:50] It's 3 times slower on i386 [11:50] (threadsafe requires nspr) [11:50] lool: maybe fedora doesnt enable GLUE with NSPR? [11:50] GaleonWrapper.cpp:1648: error: 'nsresult GConsoleMessage::GetMessageMoz(PRUnichar**)' cannot be overloaded [11:50] do we do that? [11:50] GaleonWrapper.cpp:1646: error: with 'virtual nsresult GConsoleMessage::GetMessageMoz(PRUnichar**)' [11:50] lool: what did you change? [11:51] So this looks like type mismatch -- and I don't understand how they could possibly get this to build [11:51] asac: I just renamed GetMessage to GetMessageMoz [11:51] aMessage->GetMessage (&message); ? [11:52] thats the line that fails for me here [11:52] which should be aMessage->GetMessageMoz [11:52]