[02:30] crimsun, bug 311334 [02:30] Launchpad bug 311334 in pulseaudio "Audio cuts out when playing ogg natively in Shiretoko" [Undecided,New] https://launchpad.net/bugs/311334 [04:03] wrong package, but ok [04:03] (fixed) === jetsaredim1 is now known as jetsaredim === asac_ is now known as asac [12:23] * asac back [12:24] white: thanks [12:24] let me know ... i will start to work on icedove 2.....19 today [12:24] happy new year everyone [12:31] hi [12:32] what is icedove 2? [12:32] 2.0.0.19 ;) [12:32] fta: ^^ [12:32] hi [12:33] tb2 ? [12:33] yeah [12:33] white is from debian security team and also seems to be iterested in helpign out a bit more [12:33] the package looks like my tb3 [12:33] on tbird 3 [12:33] fta: no icedove package looks like the tbird 2 packages we have [12:34] i mean the icedove 2 ;) [12:34] really? [12:34] fta: of course ... i am the icedove maintainer ;) [12:34] fta: icedove 3 doesnt exist yet ... but white wnated to look at them and i told him to base this on tbird 2 [12:34] err 3 [12:36] asac, http://paste.ubuntu.com/98974/ [12:36] looks like icedove 3 to me [12:36] fta: nice ... seems he already diud it then [12:36] fta: yeah [12:37] i never questioned that ;) [12:37] i guess the branding dir is a bit fucked still ... but that shouldnt be hard to fix [12:37] so a mozilla-devscripts file and it rocks [12:38] we could move it there ... but i think it makes most sense to just use tbird 3 moz-devscript stuff [12:38] and patch the branding and so in debian/rules [12:38] if you want to add a icedove.mk ... why not ;)= [12:38] though i thought we wanted to stop putting more orig logic into -devscripts [12:38] remember the project file could live in the package now [12:39] hi [12:39] fta: right ... thats what i would want in this case [12:39] will someone moderate my email on the ML? [12:39] thanks in advance! [12:39] i mean, use a project file instead of the new stuff in the diff [12:39] BUGabundo: i would if i could [12:39] remember [12:39] ;) [12:39] the pass [12:39] usually gnomefreak does that [12:39] ehehe [12:39] fta: sure. can you guide white how to do that? [12:42] yep [12:48] asac, but that mean m-d in debian... [12:57] fta: thats ok [12:57] nobody raised any real concern and even if ... ;) [12:57] who cares? [12:57] :-Ü [13:05] ok 64-bit native plugin backed out :) [13:05] why? [13:05] because we dont have a stable URL ;) [13:06] already discussed that right? [13:06] oh [13:06] also kees dumped nspluginwrapper support [13:06] which is supposed to be optional [13:07] fta: native will come back when adobe has final ... so we can download it from partner [13:07] also latest nspluginwrapper is quite perfect [13:07] shouldnt make much a difference anymore ... except that ffox doesnt crash [13:09] fta: give it a try on 32-bit [13:09] for me it just rocks ;) [13:34] is there a cdbs hook for lintian? [13:35] i mean, an existing .mk [13:35] fta: in which way? [13:35] so that it gets run automatically? [13:35] just build-dep on lintian, include a file and done [13:35] yep [13:35] fta: purpose: run auto? [13:35] ah ok [13:36] i dont think so ... i think the current hook is: "use debuild" [13:36] not dpkg-buildpackage [13:37] fta: what was the prob with mozilla bug 470379 ? [13:37] Mozilla bug 470379 in Add-ons Manager "Remove code to upgrade from 1.0 profiles" [Normal,Resolved: fixed] http://bugzilla.mozilla.org/show_bug.cgi?id=470379 [13:38] breaks our gre patch? [13:38] Jazzva: there? [13:38] ;) [13:38] well, no real problem, a chunk of code disappeared [13:39] so part of the patch is gone now [13:39] fta: did our code touch V10 function? [13:40] _upgradeFromV10 i mean [13:41] asac, http://bazaar.launchpad.net/~mozillateam/xulrunner/xulrunner-1.9.2.head/revision/396 [13:41] asac: yep [13:43] Jazzva: hi ... i uploaded nspluginwrapper ,) [13:43] asac: thanks :) [13:43] Jazzva: we have one rather important feature we need for nspluginwrapper: auto-update of wrappers [13:44] my current idea is that nspluginwrapper has a registry dir like: /var/lib/nspluginwrapper/wrapper.d/ [13:44] which nspluginwrapper looks at during postinst and recreates the wrappers [13:45] i hope we can just put the wrapper.so's in there and nspluginwrapper is smart enough to find the "source" when using the -u flag [13:45] we can try that :) [13:45] like nspluginwrapper -u /var/lib/flashplugin-nonfree/npwrapper.libflashplayer.so [13:45] err [13:46] but of course a different path [13:46] nspluginwrapper -u /var/lib/nspluginwrapper/wrapper.d/npwrapper.libflashplayer.so [13:46] if we cannput put the wrappers there directly we can have text files in there with lines like: [13:46] /var/lib/flashplugin-nonfree/npwrapper.libflashplayer.so /usr/lib/flashplugin-nonfree/libflashplugin.so [13:47] but i hope that nspluginwrapper -u just works [13:47] (though it seems to not do anything - at least when i dont change the /usr/lib/flashplugin-nonfree/libflashplugin.so ...) [13:48] nspluginwrapper -v -n -u /var/lib/flashplugin-nonfree/npwrapper.libflashplayer.so [13:48] Update plugin /var/lib/flashplugin-nonfree/npwrapper.libflashplayer.so [13:48] nspluginwrapper -v -n -u /var/lib/flashplugin-nonfree/npwrapper.libflashplayer.so [13:48] Update plugin /var/lib/flashplugin-nonfree/npwrapper.libflashplayer.so [13:48] oops [13:48] ;) [13:48] it doesnt touch it ... so maybe its rewally smart and checks md5sums? [13:49] or its broken [13:49] maybe they're already updated :)? [13:50] yes thats what i mean ... but that would mean its smart and checks md5sums ... or maybe it doesnt do anything if the api is ok? [13:50] the last time I needed to recreate wrappers was when we moved from 1.1.8 to 1.1.10 [13:50] no clue [13:50] we should try [13:50] yeah [13:50] i am just curious if its smart enough or if its broken :) [13:50] let me downgrade to 1.1.8, and test [13:51] http://paste.ubuntu.com/99011/ [13:51] thats the code [13:51] yeah downgrade would help [13:54] else if (strcmp(plugin_info.ident, NPW_PLUGIN_IDENT) != 0) { [13:54] seems to do something smart at least [13:54] right, looking at that part [13:54] else if (strcmp(plugin_info.ident, NPW_PLUGIN_IDENT) != 0) { [13:55] oh so maybe a touch of the .so is enough [13:55] to trigger it [13:55] brb, lunch time... [13:55] yeah [13:56] ok so upgrading plugins would work that way [13:56] now just to check whether upgrading/downgradiung nspluginwrapper works too [14:30] asac, where does nspluginwrapper put its wrappers? [14:31] asac: npwrapper.libflashplayer.so -> /var/lib/flashplugin-nonfree//npwrapper.libflashplayer.so [14:31] Jazzva: currently its undefined [14:31] but /var/lib/flashplugin-nonfree/ is empty... [14:31] Jazzva: we can either say: they always have to be in /var/lib/nspluginwrapper/wrapper.d [14:31] or we can say that they have to put a text file in that directory [14:32] that describes where the parts are [14:32] Jazzva: well ... it shouldnt be empty [14:32] Jazzva: if you are on jaunty its because kees dropped nspluginwrapper support [14:32] completely [14:32] i uploaded a fixed version earlier today [14:32] ah... ok [14:32] maybe its already on your mirror m [14:32] that explains it :) [14:32] flashplugin-nonfree that is [14:33] yeah ;) [14:33] Jazzva: i think we should put the wrappers directly there [14:33] and make a link in /var/lib/flashplugin-nonfree/flashplugin.so ... which is then used also if you dont use wrapper [14:34] but i can fix the flashplugin-nonfree mess [14:34] will do it when renaming the package to flashplugin-installer [14:36] Jazzva: so the idea is to ship a script in /usr/sbin/update-npwrappers [14:36] that updates the wrappers .... flashplugin-nonfree will then just call that in postinst [14:36] hmm [14:37] we should still have an option for people that don't want to use npw for some reason... [14:38] can't we just call npw -u? [14:38] Jazzva: imo thats independent ... postinst of flashplugin-nonfree should still check whether nspluginwrapper is on system before doing that [14:39] urg ... http://paste.ubuntu.com/99048/ [14:39] fta: ^^ [14:39] whats up with launchpad cert? [14:40] asac: http://paste.ubuntu.com/99049/ [14:40] Jazzva: its a bit difficult [14:40] npw -u should work :) [14:40] Jazzva: yeah i think thats ok [14:40] Jazzva: i just wonder if we need more magic [14:40] i mean: postinst would have to check whether a wrapper was already created [14:41] question here is: does postinst know the name? [14:41] and path ... where to look at? [14:41] maybe nspluginwrapper should ship a ensure-npw ... which would then do the magic of [14:41] a) guessing whether its already wrapped [14:41] b) run -u or -i depending on the result [14:42] do we know whether its safe to guess the wrapper filename like: npwrapper. ? [14:43] imo we should put that logic in nspluginwrapper ... so we just have to change it at that place if that guessing changes at some point [14:43] asac, tar: icedove-branding-2.0.0.x.tar.gz: Cannot open: No such file or directory [14:43] fta: yeah but thats the followup issue [14:43] fta: the issue is that bzr branch bails out due to cert issues [14:43] asac: I'll check in the source [14:43] so i dont have a tarball [14:44] asac, is that a signed branch ? or a branch with signed commits ? [14:44] Jazzva: i think its curently safe ... what i mean is that we need a script to do the update-install stuff in nspluginwrapper ... so we dont have to put that guess-logic in each and every package [14:44] fta: i think its a ssl cert issue [14:45] because its a curl issue [14:45] bzr: ERROR: pycurl.error: (60, 'server certificate verification failed. CAfile: /etc/ssl/certs/ca-certificates.crt') [14:45] do you have that file? [14:45] i do [14:45] asac: right, we could run "nspluginwrapper -l" to check for installed wrappers, and then to run "nspluginwrapper -u" on them, if that works [14:45] fta: try to run: [14:45] i need to run, cu ~7pm [14:46] sh -c "cd $tmpdir; bzr export --format=tgz --root= icedove-branding-2.0.0.x.tar.gz lp:~mozillateam/thunderbird/icedove-branding-2.0.0.x ; tar xzf icedove-branding-2.0.0.x.tar.gz" [14:46] Jazzva: do you run jaunty? can you try the command above? [14:46] fta: ok [14:46] cu [14:46] asac: ok [14:46] bug 310675 [14:46] Launchpad bug 310675 in gnutls26 "failure to trust login.launchpad.net, while openssl does" [Undecided,New] https://launchpad.net/bugs/310675 [14:46] Jazzva: err sorry [14:46] ;) [14:46] Jazzva: rock thanks [14:46] james_w: ^^ [14:47] "--root= icedove-branding-2.0.0.x.tar.gz" ? [14:47] Jazzva: so nevermind ;) [14:47] fta: thats ok ... it always worked [14:47] ok :) [14:47] james_w: happy new year .... do you know that bzr bd _always_ recreates tar if we have .bzr-builddeb/default.conf magic? [14:47] asac, http://paste.ubuntu.com/99053/ [14:47] james_w: that caused issue for me because the orig doesnt match everywhere [14:47] asac: yes [14:47] it's a known problem [14:48] james_w: ok so its accepted as a problem. thanks [14:48] james_w: is there a bug [14:48] fta: yeah i was dumb [14:49] fta: i pasted the wrong code ;) ... the code that fails uses https://code.launchpad.net/~mozillateam [14:49] as branch [14:49] fta: thanks anyway [14:50] asac: can't find one [14:50] asac: this should update wrappers, at least it seems logical to me :)... nspluginwrapper -v -n -u `nspluginwrapper -l | grep "npwrapper"` [14:52] Jazzva: hmm ... for that we would need to teach nspluginwrappera bout our other locations [14:52] e.g. xulrunner-addons/plugins firefox-addons/plugins/ [14:53] but i think thats the right way to go [14:53] otoh what if users have their own wrapper installed? [14:53] shouldn't npw already know about it? [14:53] Jazzva: how? i think the current -l logic is quiet dump. will probably scan a fixed set of directories (e.g. profile + system dirs) [14:54] i dont think it checks whether the plugin originates from itself [14:54] ok, i'll check [14:55] Jazzva: we could whitelist a bunch of locations [14:56] Jazzva: but maybe we should really just do what you suggested [14:56] and teach npw about more locations [14:56] and later fix the "dont do that if its not generated by the packaged npw" [14:56] asac: it's in process_list, and it checks dirs returend by get_mozilla_plugin_dirs() [14:57] (in file npw-config.c) [14:57] Jazzva: yes. do you see any meta info field that points to the npw lib? [14:57] strings /var/lib/flashplugin-nonfree/npwrapper.libflashplayer.so | grep /usr/lib/nspluginwrapper/ [14:57] /usr/lib/nspluginwrapper/x86_64/linux/npwrapper.so [14:58] Jazzva: ^^ [14:58] so there seems to be some meta info available that we could use to limit output of -l [14:58] ah... I see... [14:58] of course at best fix -l itself and not write something around it [14:58] ok lets go for it [14:59] question now is whether we want to provide a convenice ensure-and-update-npw [14:59] convenience [14:59] also we might want to have something like delete-npw 2D [15:00] NPW_PluginInfo should be the struct that holds all information about wrapper... [15:00] yeah [15:01] typedef struct __attribute__((packed)) { char ident[NPW_PLUGIN_IDENT_SIZE]; char path[PATH_MAX]; time_t mtime; char target_arch[65]; char target_os[65]; [15:01] } NPW_PluginInfo; [15:01] in sysdeps.h [15:02] right... [15:03] so is "path" the path from above? [15:03] maybe a printf(..) would help [15:03] ;) [15:03] asac: we could just add our dirs in "static const char *default_dirs[]" in npw-config.c, which I think we already do in one patch [15:04] +static const char *debian_link_dirs[] = { [15:04] thats all quiet messy imo [15:04] yeah ;) [15:05] i think its really just a hack for the "instalL" [15:05] doesnt register the dirs in there properly [15:05] 000_debian_make_symlinks.diff [15:05] or? [15:06] Jazzva: so we have NSPLUGIN_DIR ... whiuch we use to installl "just" to a specific dir [15:06] maybe we can just use the current dirs ... and the NSPLUGIN_DIR (if set) to implement -l [15:07] asac: btw, path field is the path to the original plugin "printf(" Original plugin: %s\n", plugin_info.path);" [15:08] hmm [15:08] so its not in that struct [15:09] Jazzva: we have: [15:09] src/npw-config.c: && strcmp(plugin_info.path, NPW_OLD_DEFAULT_PLUGIN_PATH) != 0; // exclude ARCH npwrapper.so [15:09] src/sysdeps.h:#define NPW_OLD_DEFAULT_PLUGIN_PATH NPW_LIBDIR "/" HOST_ARCH "/" NPW_WRAPPER [15:09] src/npw-config.c: && strcmp(plugin_info.path, NPW_OLD_DEFAULT_PLUGIN_PATH) != 0; // exclude ARCH npwrapper.so [15:09] src/sysdeps.h:#define NPW_OLD_DEFAULT_PLUGIN_PATH NPW_LIBDIR "/" HOST_ARCH "/" NPW_WRAPPER [15:09] for me this seems like plugin_info.path is really supposed to be the wrapper.so path [15:09] and not the original plugin [15:09] strange [15:09] maybe there are two plugin_info ? [15:11] not sure... [15:11] Jazzva: oh ok [15:11] maybe the wrapper.so is a valid plugin and this function just ensures that it doesnt wrap itself [15:11] so yeah [15:11] so there is most likely no such info available [15:11] too bad [15:13] so we would need to extent this thing ... but well [15:13] should be possible to do in future [15:13] so nothing really blocking us imo [15:14] Jazzva: i think if we have all right dirs registered, just running [15:14] nspluginwrapper -u [15:14] (without file) should just work [15:14] well i mean its supposed to work [15:14] ;) [15:14] semantic like: upgrade all [15:15] asac: will work, if we change it :). I think I already tried running "npw -u" :) [15:15] npw -a? [15:16] asac: nspluginwrapper -v -a -u [15:17] yeah ;9 [15:17] just found it too [15:17] sudo nspluginwrapper -v -a -u [15:17] hehe [15:17] :) [15:17] i think we also need -n ? [15:17] which is done here: static int auto_update_plugins(void) on line 952 in npw-config.c [15:18] so, if we need to add more dirs, we can patch get_mozilla_plugin_dirs... [15:19] yes. so hard code or use NSPLUGIN_DIRS=dir1,dir2,dir3 ... to extend that list? [15:20] maybe hardcode and send upstream [15:21] so adding xulrunner-addons firefox-addons mozilla and /var/lib/nspluginwrapper/wrapper.d/ [15:32] asac, so are we going that way? to patch default_dirs? [15:32] Jazzva: yep ... want to do that? [15:32] ok... [15:33] do we use /usr/lib64/firefox-addons and stuff for x64 sysstems? [15:33] Jazzva: you could also adjust the nspluginwrapper dir used by the debian patch that creates symlinks [15:33] *systems [15:33] it should use the wrapper.d too i think [15:33] hmm [15:33] yeah i thkink that would be right [15:34] Jazzva: i think we dont need /usr/lib64/ for now ... just /usr/lib/ i think [15:34] but we might want to revisit that situation when we come to 64-bit native plugin and stuff [15:35] asac: also, take a look at this :) [15:35] s/situation/decision/ ;) i am stupid [15:35] asac: http://paste.ubuntu.com/99071/ hehe :) [15:36] 'n' for "native", and 'n' for "nosymlinks" [15:37] lets use "z" then ;) [15:37] for nosymlinks [15:37] and send that changed patch to debian [15:37] and to also document it in print_usage [15:37] since it's documented in manpage, as -n parameter, and it's not documented in --help [15:38] yeah [15:38] ok, I'll fix that later. I'm gonna work on school project a bit :) [15:38] so we can continue this later tonight :) [15:38] yeah and iam going http://identi.ca/notice/1684595 [15:38] cu later [15:38] have fun, cu [15:39] sport isnt much fun for me [15:39] ;) [15:39] but thanks! [15:39] u2 [15:39] heh [15:39] thanks [18:31] crimsun, it's back ? http://paste.ubuntu.com/99150/ [19:41] fta: upload.ubuntu.com ftp is down? [19:41] no idea [19:46] so the glue in debian will go to icedove-dev [19:47] as it seems [19:47] and iceape will be dropped or EOL'ed once we find that we cannot provide sec support in lenny [19:48] do you mean, no seamonkey at all in debian? [19:56] yes [19:56] nobody cares enough to prepare a build and QA that based on new tarballs [19:57] for security updates [19:57] maybe someone should inform upstream [19:58] about what? [19:58] that nobody wants to step up for iceape? [20:01] or that they managed to make seamonkey become another victim of heavy maintenance burden due to mozilla security policy? [20:01] ;) [20:01] seriously, it shows how hard it is for mozilla software to stay in a distro [20:12] Jazzva: ok so just running -u as root sounds a bit scary as it seems to update user profile wrappers too [20:13] asac: patching the source to add option --no-profile-dirs? [20:14] so, if that's set, then we can skip return of profile dir in get_mozilla_plugin_dirs :)... [20:15] #rank name inst vote old recent no-files (maintainer) [20:15] 4976 seamonkey-browser 9732 4262 4565 903 2 (Ubuntu Mozilla Team) [20:15] 6136 iceape-browser 6465 118 4236 13 2098 (Ubuntu Mozilla Team) [20:15] 33581 seamonkey-2.0-browser 42 16 19 7 0 (Unknown) [20:15] wow [20:15] no-files ok [20:15] quiet a bunch of gutsy users? [20:16] i think no-files is transitional package [20:16] 8565 seamonkey 2993 0 6 0 2987 (Ubuntu Mozilla Team) [20:16] and just iceape?` [20:16] 10448 iceape 1798 0 0 0 1798 (Ubuntu Mozilla Team) [20:17] so seamonkey is probably [20:17] ~ the same as the users thatinstalled it to use it [20:17] the rest is pulled in by rdepends? [20:17] are there rdepends? [20:17] hmm [20:18] fta: look at mozilla-noscript [20:19] that has rank 5630 and pulls in seamonkey-browser? [20:19] 5815 mozilla-noscript 7233 955 4953 801 524 (Arnaud Renevier) [20:19] yeah ... that pulls in seamonkey-browser [20:19] i have the feeling that this is an accident then [20:19] almost everyone wanted -noscript [20:19] but well ... quiet guesswork [20:20] i dropped -noscript since ff3.0, adblock is enough [20:20] just seems strange that almost every seamonkey user uses noscript ;) [20:20] fta: yes :) [20:20] fta: but it seems that we didnt transition it for firefox users [20:20] and now users that used that in the past have seamonkey-browser ;) [20:20] (just one/my theory) [20:20] 1514 thunderbird 180467 45689 122773 11956 49 (Alexander Sack) [20:20] 1667 mozilla-thunderbird 131112 617 48249 114 82132 (Alexander Sack) [20:20] 25793 thunderbird-3.0 141 16 66 59 0 (Unknown) [20:21] 141, lol [20:21] most were at uds then ;) [20:21] its b1? [20:21] why not upload to jaunty? [20:21] it's my ppa [20:22] no reason [20:22] ok will look and if i have any idea that we need something before let you know [20:23] 14973 firefox-3.1 891 146 312 433 0 (Unknown) [20:23] 35431 firefox-3.2 40 5 0 35 0 (Unknown) [20:23] +EXTRA_SYSTEM_CONFIGURE_FLAGS [20:23] why do we set that to NULL? [20:23] what harmful stuff is in there by default? [20:24] (xul ...head) [20:24]