=== asac_ is now known as asac [01:13] hm, i've lost subpixel with the new cairo [09:19] Hi ! [09:34] hi XioNoX [09:35] asac, I've just add the license block [09:35] good ... everywhere? [09:35] e.g. on all your branches? [09:37] not the 3rd [09:41] fta: oh. dholbach voted for you ;) ... nice [09:41] XioNoX: ;) [09:41] all good are three ;) [09:41] ok, nice [10:01] asac, What is the "awsome" think you told me yesterday ? [10:02] XioNoX: hehe ;) [10:04] XioNoX: what i want is a the application pane ... but just filtered by known plugin mime-types [10:04] do you think I'll have time ? [10:05] and what is the application pane ? [10:07] XioNoX: preferences -> applications [10:07] XioNoX: thats basically what we want to manage plugin mime-types [10:08] so I work in firefox code ? [10:08] XioNoX: no [10:08] ubufox ? [10:08] XioNoX: yes. for now [10:09] XioNoX: i think all what we do for the altplugins things is kind of a prototype to show upstream that managing multiple alternatives can be nicely done [10:09] ok [10:10] XioNoX: what we want is a new menu item in the right row: "Use other plugin ..." [10:10] XioNoX: you think you can add that? [10:10] try to overlay it i'd say [10:12] I don't understand :) [10:12] XioNoX: have you found that pane? [10:13] yes [10:13] on the left there are mime-types (e.g. shockwave flash) [10:13] on the right there are options [10:13] err "Action" [10:13] XioNoX: got that? [10:14] yep [10:14] XioNoX: ok ... in action there is a drop down menu? [10:14] yes [10:14] for now we want "Use other plugin ..." there [10:15] which triggers our plugin finder wizard with the proper mime-type! [10:16] it is this part that I don't understand :D [10:17] a new "action" element called "Use other plugin" ? [10:17] yes [10:17] haaa [10:17] but nobody goes in this tab [10:18] XioNoX: true [10:18] XioNoX: however, [10:19] the idea is to use the same UI in the addons manager ... just that it doesnt use the gnome-mime database, but the mime-types currently provided by plugins [10:19] XioNoX: if you think thats too much we can also improve the "current" plugin finder wizard [10:21] I think it is a lot of work for a little thing, but I can be wrong [10:22] what do yo have in mind by "improve the "current" plugin finder wizard" ? [10:24] XioNoX: only display plugins that are not currently used [10:27] XioNoX: we also need to take care that if you install gnash after flahsplugin-nonfree that you will use gnash [10:27] if you then install flashplugin-nonfree again, you want that to be used [10:27] "install" select in alternative plugin finder [10:27] "install" == select in alternative plugin finder [10:28] when you install flahsplugin-nonfree from repos, gnash is removed, etc... [10:28] XioNoX: no its not [10:28] XioNoX: currently we maintain alternatives ... but that is not really required imo [10:28] well [10:29] for global alternatives might be ok. but what we probably want to do here is to put the a link into the profile directory [10:29] so the user can switch between installed alternatives [10:30] but in the end we lack mozilla features in order to do it in a perfect fashion [10:30] i agree [10:31] e.g. mozilla doesnt allow you to select individual plugins by individual mime-types [10:31] ok [10:31] so we need to be innovative ;) [10:32] so first, filter plugins in the richlist ? [10:32] to show only installed plugins [10:33] err, non installd plugins [10:33] XioNoX: we have a rich list? [10:34] XioNoX: if you want to go for the application panel, yes. [10:34] not rich [10:34] radiolist [10:34] yeah [10:39] I get a new version of ubufox ? [10:39] or I work on an existing one ? [10:40] XioNoX: i think it belongs in the altplugins branch [10:40] ok [10:41] XioNoX: what you can do is to check whether /usr/lib/ ... exists [10:41] that is a bad hack ... but it probably gives you a quite accurate idea whether the package is installed or not [10:41] but this works only for packages :( [10:41] not for user-scale install [10:45] radiolist, yes. [10:46] XioNoX: well. for user based install we could try what i sadi ... e.g. create the proper link [10:53] it is a bad hack, no ? [10:53] it is like the plugin is installed twice, no ? [10:55] no ... i think profile plugins will win [10:55] but not sure [10:56] if it is like extentions, profile win [10:58] not 100% sure ... i think plugins are rather stupid [10:58] whatever is scanned first and matches mime-type wins [10:58] so profile probably will win [10:58] and we can't use the about:plugin stuff ? [10:59] fopr what? [10:59] to know what plugin is used [10:59] for wich mimetype [10:59] XioNoX: you can iterate all plugins [11:00] XioNoX: however, its hard to match what package each plugin is mapped to [11:00] and about:plugins don't make différence between pakages and profD plugins [11:00] XioNoX: well ... we can find the install path [11:01] pluginreg.dat [11:01] because it will be the same thing if we make links in te user dir [11:01] in the profile [11:01] not sure if it follows links [11:01] if it does it becomes even trickier [11:02] using pluginreg should be easier, no ? [11:03] XioNoX: well ... that info you can get through the navigator object [11:03] in javascript [11:03] ? [11:04] there are mimetype, plugin name, and file location [11:11] yes [11:13] XioNoX: ok. how about this: [11:13] we add the ability to set a pref modules.plugins.mimetypes.application/pdf = PATHTOPLUGOm [11:13] to firefox core [11:14] we add the ability to set a pref modules.plugins.mimetypes.application/pdf = PATH/TO/PLUGIN [11:14] e.g. pref("modules.plugins.mimetypes.application/shockwave-flash", PATH/TO/PLUGIN) [11:15] XioNoX: i guess that would be helpful right? [11:16] it wil be the same as the one present in the pluginreg.dat, no ? [11:16] and who will set this pref ? [11:16] XioNoX: our wizard will set that pref [11:16] XioNoX: (or the user manually if he wants) [11:17] XioNoX: and yes: the path would be the one in pluginreg.dat [11:17] so why using an intermediate pref ? [11:17] XioNoX: why? because the pluginreg.dat format doesnt know anything about preference [11:17] why we don't just ask pluginreg.dat when needed ? [11:18] and there is no write-access interface available anyway [11:18] XioNoX: pluginreg.dat has no information about preference when there are multiple plugins for the same mimie-type [11:18] but we don't have to modify it, just get values [11:18] XioNoX: which value? [11:18] there is no mean to express user-preference in pluginreg.dat [11:19] thuse we would need to extend that format ... which i dont want -> thus the pref [11:19] ha ok [11:19] and the reason why i dont want to extend that format is that there is no API to manage plugin registry at all [11:19] so we would have to invent even more [11:20] you want that if there are many plugin for the same mimetype installed [11:20] the user can chose [11:20] yes [11:20] right [11:20] the final idea is still to have the drop down box from prefernec -> applications where the user can switch easily when there are already multiple installed [11:20] and not if there are 1 installed and you install an alternative it remove the old one [11:20] or install new ones thjrough "get other plugins ..." [11:20] XioNoX: right [11:21] XioNoX: doing that will cover the user install case i hope [11:21] XioNoX: only thing that is tricky is to figure which path the "just-installed" plugin has [11:21] for packages that should be more or less easy [11:21] for .xpi installs that can be harder, but we will figure something i am sure ;) [11:22] XioNoX: i can write the firefox core changes to honour such preferences [11:22] you can write the logic that allows users to switch to other plugins ;) [11:23] this seam to be a big project [11:23] and there are some hope that this will be included upstream ? [11:23] XioNoX: its not a big project [11:24] XioNoX: the two parts are managable [11:24] ok [11:24] XioNoX: if you use the current plugin finder wizard at least [11:24] implementing it the applications panel way is more work [11:24] and is what i suggested in the upstrewam bug [11:24] ok [11:25] so whether this can land, depends on how well we make it work for us [11:25] imo [11:25] once the people see that it works, there is no real way to refuse to have the proper solution imo [11:26] XioNoX: i think the clean solution would be to split the features apart: [11:26] 1. choose which of the installed plugins to use [11:26] 2. install more plugins -> which would end in a dialog for 1. [11:28] makes sense XioNoX ? [11:28] when we clic on "Use other plugin" ? [11:28] or were ? [11:28] fta: there? what new liferea is that? [11:29] asac, going to eat [11:29] XioNoX: ok [11:29] see you [11:32] * gnomefreak hopes new gnome gives me back system bell [11:45] gnomefreak: unlikely ;) [12:19] back [12:21] XioNoX: ok. i think i have a first patch for the thing [12:21] already ? [12:21] XioNoX: however, you probably need to spin your own xulrunner [12:21] XioNoX: for the "select plugin by mime-type" ...yes. [12:21] err [12:21] for the preference things [12:21] spin your own xulrunner ? [12:22] XioNoX: let me push it ;) [12:22] ok [12:24] XioNoX: mkdir myxul; cd myxul; bzr branch lp:~mozillateam/xulrunner/xulrunner-1.9.head; cd ..; mkdir tarballs; cd tarballs; wget http://launchpadlibrarian.net/16695860/xulrunner-1.9_1.9.0.2%7Ecvs20080807t1327%2Bnobinonly.orig.tar.gz; cd ../xulrunner-1*/; sudo apt-get build-dep xulrunner-1.9; sudo apt-get install bzr-builddeb; sudo apt-get install built-essentials; edit debian/patches/series (enable the last patch here); bzr bd --merge --dont- [12:24] err [12:24] XioNoX: mkdir myxul; cd myxul; bzr branch lp:~mozillateam/xulrunner/xulrunner-1.9.head; mkdir tarballs; cd tarballs; wget http://launchpadlibrarian.net/16695860/xulrunner-1.9_1.9.0.2%7Ecvs20080807t1327%2Bnobinonly.orig.tar.gz; cd ../xulrunner-1*/; sudo apt-get build-dep xulrunner-1.9; sudo apt-get install bzr-builddeb; sudo apt-get install built-essentials; edit debian/patches/series (enable the last patch here); bzr bd --merge --dont-purge - [12:24] thats ;) [12:24] and what does it wo ? [12:25] it builts the latest xulrunner-1.9 from our packaging branch ;) [12:25] to do that it branches the packaging branch, gets the tarball [12:25] installs the build-dependencies [12:25] and bzr-builddeb [12:25] and then you edit debian/patches/series to enable the my patch [12:25] and build using the "bzr bd ..." command [12:25] ;) [12:25] all that is done in a directory called myxul ;) [12:28] XioNoX: that isnt tested yet ... at best start with UI for selecting a plugin ;) [12:28] applying that to the backend will be as easy as setting a pref ;) [12:29] ok [12:29] myxul ahve to be in the bzr working folder (ubufox plugin) or not [12:29] ? [12:29] XioNoX: no ... my description was based on that you dont have anything [12:29] ;) [12:30] if you already have the xulrunner-1.9.head branch on your disk you can start with getting the tarball [12:30] ha ok [12:30] thx [12:30] XioNoX: anyway. chances are 50% that this works out of box. so better start implementing the "select plugin for mimetype" dialog [12:31] XioNoX: http://developer.mozilla.org/en/docs/DOM:window.navigator.plugins [12:31] thats how you can access installed plugins (i hope) [12:31] well ... i should work (looked at the doc) [12:31] great [12:32] first i'm trying the xulrunner commands [12:32] so what we want is a dialog that allows you to switch to another installed plugin that serves the mime-type [12:33] XioNoX: sure. the built will take 20-45 minutes anyway ;)= [12:34] E: Les dépendances de compilation pour xulrunner-1.9 ne peuvent pas être satisfaites. [12:34] yeah, french :) [12:35] can't satisfy build deps for xulrunner... [12:35] XioNoX: you forgot to run the apt-get commands [12:35] at least the build-dep one [12:35] ;) [12:35] sudo apt-get build-dep xulrunner-1.9 [12:35] if you did, let me know what dependency is missing [12:35] tell me this [12:35] ho [12:36] i don't know it just say that it can't [12:36] i don't know, it just say that it can't [12:37] XioNoX: you need deb-src lines in your sources.list [12:37] (apt) [12:37] I have them [12:37] well. it would work then [12:38] here is the problem [12:41] XioNoX: ? [12:41] you just need to install the missing dependencies [12:41] thats all [12:42] if build-dep is broken for you do it manually ... but it works for everyone ;) [12:46] ok, the probleme come from the package python2.5... [12:51] XioNoX: whats the problem? [12:53] I have python2.5....ubuntu5 and python2.5-dev need python2.5....ubuntu4.1 only [12:53] I've installed python2.5....ubuntu4.1 with dpkg, but now it is the pess in the pacakges [12:53] XioNoX: dontgrade ;) [12:53] downgrade i mean [12:53] is that a problem? [12:54] yes, but i can't [12:54] if i downgrade it wan't to remove the half of my system [12:55] XioNoX: where did you get the python debs from? [12:55] ok fix [12:55] good [12:55] I don't know [12:55] reps [12:55] repos [12:56] broken pakage manager have to be improved [12:57] build running [12:59] I'll see of it is true that 64bits compile faster [13:03] asac: morning! say, which system settings plugin do you use for NM in ubuntu then? [13:05] rbu: in ubuntu i havent enabled ifupdown yet. [13:05] rbu: tonight i made a break through though ;) ... so in the near future Ill enable it ;) [13:06] so hopefully we will have: --plugins=ifupdown,keyfile [13:07] asac: ifupdown would store the settings in normal debian settings files? [13:07] rbu: ifupdown is just legacy support ... i currently dont feel to implement a read/write approach. so for now its read-only [13:08] keyfile would be used for write support [13:08] i hope in this order it will automagically work ;) [13:09] asac: ah, so it would parse that /etc/network/interfaces file? [13:09] rbu: yes [13:09] rbu: when time permits i will also add support for wpa_supplicant.conf [13:09] asac: is mbiebl going with the same thing, or are you guys not talking to each other :-) [13:09] but maybe to a separate read-only plugin [13:10] rbu: mbiebl is waiting for my implementation iirc [13:10] rbu: the idea is that it gets committed upstream asap [13:10] i am overdue to deliver that [13:10] asac: lazy !$!§$% [13:10] :-) [13:10] rbu: well ... its a moving target [13:10] asac: yeah, but wpa_supplicant.conf would be cool :-) [13:10] the system-settings thing was broken for quite some time (for connections that required secret) [13:11] rbu: thats simple ... wpa_supplicant.conf alone is just to parse the setting and creating NMConnections with those settings [13:11] (not bound to a mac address) [13:12] not yet sure where to configure which setting to use. problem is that ifupdown also would need wpa_supplicant.conf imo. but thats something i have to figure out elsewhere [13:12] and a wpasupplicant top-level plugin makes sense ;) [13:13] asac: well, you're in that swamp a lot deeper than me, i've just learned about what nm-settings is today, so ... way to go for me [13:13] rbu: hehe ;) [13:14] rbu: learning by doing. i am still learning as well ;) .... but if you want to help, you could probably implement a wpa_supplicant.conf parser [13:14] (which would be independent from nm-settings) [13:14] (and it would definitly speed this feature up) [13:15] * gnomefreak ran into issues ill be back when handled [13:20] asac, compillation done [13:21] XioNoX: hehe ;) [13:21] XioNoX: ok ... install the packages and see if firefox still works ;) [13:22] XioNoX: then see if you get a crash when using any plugin [13:22] then see if you can force gnash or something even though flashplugin-nonfree would be used by default [13:22] XioNoX: but well :) ... at best just go and implement the dialog ;) ... that will be tricky enough. I will take care that the "force" works [13:38] asac, I was on tel [13:48] xulrunner works well ;) [13:54] can anyone using my ppa confirm that lcd filtering is broken (or not) with my cairo 1.7.4 ? [14:00] XioNoX: how does it work well? [14:00] XioNoX: did you try to switch plugins? [14:00] e.g. by setting manually the pref?` [14:01] fta: i can try after lunch break [14:01] (which i should really do now ;)) [14:01] we can do that ? [14:01] XioNoX: yes. thats what i added [14:01] ;) [14:01] XioNoX: read the patch for instructions [14:01] XioNoX: i documented it on top [14:01] ok [14:02] XioNoX: you can only switch between plugins that are in pluginreg.dat [14:02] ok [14:02] so maybe you need to create a link to libgnashplugin.so in your profiles plugins/ directory [14:02] (given that you use the nonfree thing by default on system) [14:38] asac, no news of pitti ? [14:39] fta: if he doesnt talk on #-devel he is probably on holiday [14:39] ok [14:45] asac, I don't see any new plugin in the applications tab [14:46] no new choice for flash [14:47] XioNoX: why would that happen [14:47] there is no code for that at all [14:47] what does your code ? [14:47] XioNoX: read the patch [14:47] there is documentation [14:48] good idea :) [15:00] http://www.glazman.org/weblog/dotclear/index.php?post/2008/07/29/Google-and-MPL [15:01] (not new, i know. i'm still catching up with my 1000+ unread articles) [15:01] asac, btw, did you see my xul1.9 issue with liferea yesterday ? [15:09] checking for XulRunner 1.9+ support... checking for XULRUNNER... yes [15:09] Fatal: XulRunner enabled, but XULRUNNER_HOME is empty! [15:09] make: *** [config.status] Error 1 [15:09] asac, ^^ that's the new liferea [15:09] asac, http://paste.ubuntu.com/38620/ [15:11] http://clarkbw.net/designs/dialog-invasion/dialog_invasion.ogg [15:33] mozilla bug 408925 [15:33] Mozilla bug 408925 in Startup and Profile System "XRE_InitEmbedding does not initialize properly profilelock and gDirServiceProvider" [Normal,Resolved: fixed] http://bugzilla.mozilla.org/show_bug.cgi?id=408925 [16:00] asac: ping [16:04] pong [16:05] fta: i did see the xul issue ... i asked you where you gto that new liferea from ;) [16:05] upstream [16:05] fta: what is XULRUNNER_HOME used for? [16:05] -rpath? [16:05] if so, just wipe it [16:05] no idea [16:05] i want this bug 203157 fixed [16:05] Launchpad bug 203157 in liferea "Liferea uses a lot of CPU at times" [Medium,Confirmed] https://launchpad.net/bugs/203157 [16:06] fta: search the Makefiles for XULRUNNER_HOME usage ... i guess its again a -rpath thing which means that they are stupid ;) [16:06] oh dear ... the rtl8187 driver just sucks [16:07] ok that doesnt make any sense [16:07] * asac reboots to latest kernel again [16:09] ./debian/patches/xulrunner-1.9: gtk_moz_embed_set_comp_path(XULRUNNER_HOME); [16:11] fta: yeah thats stupid [16:11] fta: they can use the GRE_PATH they found [16:11] or just leave that lone (if they use push_startup [16:11] ) [16:12] asac, how can I interract with the preferences dialog ? [16:12] ./src/mozilla/Makefile.am:liblihtmlx_la_CFLAGS = $(XULRUNNER_CFLAGS) $(PACKAGE_CFLAGS) -DXULRUNNER_HOME=\""$(XULRUNNER_HOME)\"" [16:13] XioNoX: why do you want that? [16:14] fta: yes, drop that define ... and #ifdef XULRUNNER_HOME the set_comp_path above [16:14] it's already in that patch [16:14] so basically, it's not used [16:15] asac, I don't really know how to start.. [16:17] I have to modify the list of options in the application tab for some plugins (like flash, etc...) [16:17] no ? [16:22] Hey asac [16:37] E: stream.c: Assertion 'pa_atomic_load(&(s)->_ref) >= 1' failed at pulse/stream.c:948, function pa_stream_writable_size(). Aborting. [16:37] Program received signal SIGABRT, Aborted. [16:37] crimsun, ^^ [16:42] XioNoX: I'd suggest to overlay the element [16:43] but I have no clue to find it in the firefox sources [16:44] XioNoX: use dom inspector [16:45] how ? [16:45] ok find [16:52] asac, could you please test cairo ? [16:58] asac, I'll continue from home [17:00] http://paste.ubuntu.com/38830/ [17:00] too bad i can't retrace post-mortem [17:02] fta: what shall i test? [17:02] lcd filter [17:02] subpixel blabla [17:02] not sure how that is testable [17:02] whats the problem that gets fixed? [17:02] e.g. how can i recognize that [17:03] just a shift for an ubuntu patch to upstream support. for me, it regressed [17:03] fta: how?` [17:03] everything looks different once restarted [17:03] libcairo - 1.5.4-0ubuntu1~fta3 ? [17:03] no [17:03] 1.7.4 [17:04] * asac restarts X [17:07] samarium (xen-i386) Building i386 build of xulrunner-1.9 1.9+nobinonly-0ubuntu2~gutsy0~jjv in ubuntu gutsy RELEASE [gnomefreak] [17:07] oh, he's not here [17:08] fta: i think everything looks better [17:08] but i am using "best shapes" in font appearence dialog [17:08] hm [17:08] so the difference isnt that bad [17:09] smoothing (LCD) always looked bad for me [17:10] does that make a difference for you? [17:12] i'm using lcd [17:21] fta: lcd regressed a bit. but i am not 100% sure. [17:21] hard to fix the difference for my eyes [17:21] i think some lines are not of the same width with 1.7.4 [17:21] while they are more accurate in 1.6.4 [17:22] but best shapes is better in any case ;) [17:22] * asac reverts to that [17:24] http://www.sofaraway.org/ubuntu/tmp/lcd.png [17:24]