/srv/irclogs.ubuntu.com/2008/09/10/#ubuntu-mozillateam.txt

ftano, ppa != dbgsym00:01
ftahttp://paste.ubuntu.com/45072/00:02
asacfta: dont you do a testcrun locally anyway? couldnt you preserve those -dbgsym for the tiome being?00:02
asacflashplayer ahoi ;)00:03
asacis that really a regression in latest firefox?00:03
ftatoday, i just did a minimal local test build00:03
asacfta: we need a name :(00:04
asacotherwise the 3.0.2 package will never enter intrepid :-D00:04
ftaa name ?00:04
asacyeah something to replace webbrowser in 3.0.head branch;)00:05
ftaoh00:05
asaci think suggestions were plain-webbrowser00:06
asacunbranded-webbrowser00:06
asacand whitebox-webbrowser (which was from me obviously ;))00:06
asacand pitti didnt mislike any of those00:06
asacplain webbrowser ... does that sound stupid?00:06
asacparanoia00:08
asac;)00:08
asacparanoia-3.0 ;)00:08
ftaunchained-browser00:08
asachehe00:08
asacnot that bad ;)00:08
asacunleashed ;)00:09
asacflying-browser ;)00:09
asacunique-webbrowser00:09
asacuni-browser ;)00:10
ftadoes it have to contain browser ?00:10
kgoetz*g* washed-fox00:10
asaceverything else would again create a mark00:10
ftaredpanda00:10
asacyeti-browser ;)00:10
asaca yeti is probably similar to an iceape ;)00:11
asacnow i wonder why we (as in debian) didnt go for yeti-browser yet-mail, yeti-suite, yeti-cal ;)00:12
ftadebrowser00:12
asacde like "the" ?00:12
asacor like .deb?00:12
ftalike deb00:12
asacactually not that a bad idea00:12
asacjust dbrowser ;)00:13
ftaalso the idea to remove/free00:13
asacwho do you see remove/free presented in "debrowser"?00:13
ftade- privatif00:14
asacok00:14
asaci think i know what you mean00:14
ftamaybe more french than us00:14
ftadis-00:14
asacbut wouldnt that more make it read like non-browser ?00:14
asaci like dbrowser00:15
ftaor just brother00:15
asacespecially how the d and b are mirrored ;)00:15
asachehe00:15
asacbother ;)00:16
ftafreendly00:16
ftaalready taken00:16
asacoh .... pitti said we need a proper noun ;)00:16
asacso bother wont fly00:16
ftarocket00:17
asacimo the name should reflect that we a) had to escape the "mark" b) couldnt use "webbrowser"00:17
ftalol00:17
kgoetznomark-browser? :P (although it might sound like sabdfl isnt allowed to use it)00:18
asaci think debrowser or dbrowser would be a good name ;)00:18
asackgoetz: ;)00:18
asacmaybe "marked-browser ";)00:19
ftaliberty00:19
ftaliberated00:19
asac*sigh* how much i hate choosing names :(00:20
kgoetzasac: its bad enough when its a new computer system, let alone something other people need to like ;/00:21
asacits like banging your head against the wall. the harder you hit the more it hurts00:21
kgoetzis there a type of fox without foxy markings?00:21
Jazzvau-browser? :)00:22
Jazzva(I think I suggested that few days ago)00:22
asackgoetz: i really dont want to create a new word that people would consider a mark when looking at it00:22
ftalooks like micro to me00:22
kgoetzasac: ah, nod.00:22
asacthe name should put an end on that thing00:22
[reed]asac: what's with the name choosing?00:23
ftaeheh00:23
[reed]why not "firefox" ?00:23
asac[reed]: because firefox is already in use ;)00:23
[reed]what's this browser then?00:23
asac[reed]: no seriously. firefox will stay firefox ;) ... its just a branding package00:23
[reed]ah00:24
ftahm, someone referred to my ppa in there :S http://blogs.adobe.com/penguin.swf/2008/08/windowless_mode_fix.html#comment-160106100:24
asacfor those downstreams that cant or dont want to follow mozilla requirements00:24
asacfurther this effort helps us to fix the bug that its too hard to rebranding firefox in ubuntu00:24
asacwhich sometimes got thrown at us (and which is justified) when we complained about how mozilla handles trademarks ;)00:25
asace.g. mozilla provides the --enable-branding switch00:25
asacwhat does debain/ubuntu provide for downstreams? not much00:25
kgoetznot sure about currently, but it used to provide a broken --enable-branding switch :P00:26
asackgoetz: thats fixed since 2.0 afaict00:26
kgoetzasac: nod.00:26
ftabad ff3.1 taking 100% cpu after visiting a <video> page, bad red panda00:27
asachehe00:28
kgoetzasac: so i can be sure i understand - theres goign to be a package with the bulk of 'firefox' in it, then a special package that holds the offical branding? (ala ubufox?)00:28
asackgoetz: well. this is all not set in stone. but for now we have a package that holds everything except the branding and then two branding packages00:28
asacfirefox-3.0-branding00:28
asacwhateverwenamit-3.0-branding00:29
asacand meta packages to track00:29
asacfirefox00:29
asacand whaeverwenamit00:29
ftayounameit-browser00:29
asackgoetz: the difficult thing is that we have two cases here00:29
asackgoetz: 1. downstreams what clean sources00:29
asackgoetz: 2. downstream wants clean binaries00:30
asacs/what/want/00:30
kgoetzfta: wenamedit-browser perhaps.00:30
Jazzvastripped-browser00:31
asacso when you reassemble a distro you can sync the binary packages and leave out the official branidng ones00:31
asacor  as you do, respin it00:31
asackgoetz: i think for you the part that we have a branding by default in source package might be helpful00:31
asackgoetz: but in general we need to make things more generic00:31
asackgoetz: so i need your input on what you want actually00:31
kgoetzasac: i can tell you about gNS specifically, and what i'm ultimately (and slowly) trying to achieve there. a version of firefox that doesnt promote non-free software (eg, recomend you install adobe/macromedia flashplayer), and contains no non-free software. the first part is what i hope to be working on, the second part is fixed by ubuntus -nobinary packages00:33
kgoetzthe promoting non-free software (and in the case of upstream source the blobs), not the artwork.00:34
asackgoetz: the first part is already there00:34
asackgoetz: well mostly00:34
asackgoetz: there also is the icecat extension which we (like in ubuntu and gNS) might look into incorporating at some point00:34
asackgoetz: anyway. the idea bout the first part is the plugin finder wizard for plugins00:35
asackgoetz: the icecat idea is quite primitive. what you really want is our plugin finder wizard with just all the non-free stuff removed from the database00:36
asackgoetz: on the extensions front we should look into contributing a filter feature for licenses to 3.100:36
asackgoetz: afaik AMO already has license information and we could leverage that somehow00:37
kgoetzasac: i've not looked at teh plugin finder wizard - is it in the main source package? (if so i can get it now and have a look)00:37
asackgoetz: its in ubufox00:37
kgoetz*grabs a copy*00:37
asackgoetz: if you have no flash installed, go to www.wetter.de00:39
asacif you click on the "install missing plugins ..." you should see three options: adobe, gnash, swfdec00:39
asacso thats the principal idea00:39
asacthe only bad about all this (regardless of the pfs or icecat approach) is that most flash sites dont allow any of those approach to be effective00:40
asacas they use the flash detection kit00:40
asacso the final piece required is a anti-flash-detection-kit code :)00:40
kgoetzhehehe.00:41
asackgoetz: i have part of that00:41
asace.g. i can detect when flash-detection-kit does a probe for the kit00:41
asacwhat i didnt manage yet is to get the right even send to the browser instance00:42
asacso it pops up the install missing plugins ... notification00:42
asacs/for the kit/for the flash player/00:42
kgoetzi assume pfs is the plugin finder?00:43
asacyes00:43
kgoetzhm. ubufox doesnt work on lpia. i guess that or it doesnt need to be specially listed in the arches="00:47
kgoetzlist00:47
asackgoetz: it should be there00:47
asacits _all00:47
asackgoetz: ^^00:47
kgoetzroger.00:48
asackgoetz: bzr branch lp:ubufox; cd ubufox; sh build.sh; firefox ubufox.xpi00:48
kgoetzi'm looking at an `apt-get source` checkout00:49
kgoetzjust fyi00:49
asackgoetz: yeah. thats outdated00:50
asackgoetz: 0.6 should have been uploaded already.00:50
* kgoetz branches bzr00:50
asacanyway. the package should be good enough to get the point00:50
kgoetzbbs00:52
asacoff00:59
asacnight00:59
kgoetzasac: will ubufox work with FF2? i'm guessing its 3 only?00:59
kgoetzlater mate01:00
Jazzvakgoetz, ubufox should work with FF2 too01:10
Jazzva(at least that's what Depends line says, I don't remember that I tested it lately in FF2)01:11
kgoetzJazzva: cheers.01:11
Jazzvanp :)01:11
kgoetz:)01:12
kgoetzfta_: \o/ nice work on motu (yeah. i'm a little behind ...)07:20
asacjtv: are all the .xpi things recovered by now?08:30
asacjtv: hi! ;)08:31
* asac reboot08:37
asachmm ... i guess that rebooting while backup is running is really a good idea ... so i'll wait a bit08:39
jtvasac: the XPI part is resolved, but now we're looking at some build problems.08:42
asacjtv: but thats not in production?08:43
jtvasac: it's not with the firefox/xulrunner stuff though08:44
asacjtv: ok. so to get it straight. the current export works and is in the "new" xpipo format?08:44
asac"new" aka new path format in the comments08:45
jtvasac: Yes08:45
asacjtv: point is that i still have to fix po2xpi and wonder if I can just grab an export and that its complete ;)08:45
jtvasac: here's a recent export you can test with: http://launchpadlibrarian.net/17397360/launchpad-export.tar.gz08:46
* asac downloads08:46
asacjtv: thats just xul/ffox?08:47
asacseems quite small08:47
jtvasac: well, it's only xulrunner...  if you see anything concrete that looks like it's missing, please let me know!08:48
asacjtv: well08:49
asacjtv: its in the product export format08:49
asacjtv: which isnt compatible with po2xpi08:49
asacjtv: cant we unify those at some point?08:49
asacjtv: e.g. directory structure an file names08:50
asacjtv: anyway. on a first glance that looks good enough to adapt the bloody C code for the po2xpi transformer ;)08:51
asacbut i would like to do a run against a full export before deploying those changes on rookery08:52
jtvasac: C code?  Really?08:52
jtvasac: what do you mean by "product" export format?08:52
jtvasac: you mean the directory layout?08:52
asacjtv: yes. the inner guts of po2xpi are C08:52
asacjtv: yes08:53
jtvasac: then that's because this *is* a product export.  :-)08:53
asacjtv: i know08:53
asacjtv: i wonder why we cannot unify both formats08:53
asacformats==dir-layouts08:53
jtvasac: I'd be interested in doing that, but I don't think I can give it a lot of priority.08:54
asacjtv: yeah. anyway, i cannot test the full po2xpi with product tarballs ... only the C transformer. but thats ok i think08:54
asacjtv: do the deltas contain the en-US.xpi's ?08:54
asacor are they still missing there?08:55
jtvasac: if they were missing before, I guess they still are.  Didn't know that was an issue.08:55
asacjtv: i didnt know either. just assumed that it was an issue ;)08:56
asacif the missing en-US.xpi were something else then it probably is not08:56
asaci will look later today08:56
asacjtv: thanks so far!08:56
* asac reboots ;)08:56
* jtv waves at asac08:56
IntuitiveNippleasac: update on LP bug #239952. The mozilla bug #444440 patch (v3) doesn't solve the issue, in fact, it makes it more confusing.11:43
ubottuLaunchpad bug 239952 in xulrunner-1.9 "firefox - the associated helper application does not exist" [Medium,Confirmed] https://launchpad.net/bugs/23995211:43
ubottuMozilla bug 444440 in File Handling "Unexpected application launched when $HOME/.mailcap contains an entry for the handled mime type" [Normal,Assigned] http://bugzilla.mozilla.org/show_bug.cgi?id=44444011:43
asacIntuitiveNipple: opk thanks11:44
IntuitiveNippleI'll rework it so it can deal with both issues and post the update to the mozilla bug with the reasons for combining11:45
asacIntuitiveNipple: hmm11:46
asacIntuitiveNipple: ok.11:46
asacIntuitiveNipple: let me know when you have upped the tested patch. please try to reproduce and verify that the initial issue is still fixed11:46
IntuitiveNippleindeed :)11:47
asacIntuitiveNipple: oh. could you attach a patch on top of the v3 patch?11:47
asacand a combined if you really want11:47
IntuitiveNippleWell the v3 patch is bad so best to rework it, it hasn't been accepted yet11:47
IntuitiveNippleI'm about to post a regression update to 444440 so they know11:47
asacIntuitiveNipple: problem is that i am also responsible for security uploads in debian and the debian xulrunner maintainer who did that patch doesnt use a patch system11:47
asacso if the patch that he applied lands in a modified way i will have issues doing the security update11:48
asacIntuitiveNipple: but well. in the end i dont mind11:48
asaci already stepped down from xulrunner security releases because the maintainer doesnt have a path system anymore11:49
asacso go ahead11:49
IntuitiveNippleahhh11:49
IntuitiveNippleokay, well I'll do it both ways then :)11:50
IntuitiveNippleget it working properly first, based on that patch, then produce a clean patch for upstream and a patch-on-patch for debian/ubuntu ?11:50
IntuitiveNippleIf the canonical sysadmins would get off their fannies and simply correct the Content-Type settings in apache/zope, we'd be happy!11:51
pmatulisasac: weird stuff, after reproducing the "bug" 6 consecutive times, another 4 consecutive times do *not* exhibit it12:10
IntuitiveNippleasac: the solution to the LP bug is simply to 'remember' the helper chosen by the user, and use it in nsMIMEInfoBase/nsMIMEInfoUnix. From your knowledge of the code, do you foresee any issues if I implement it so that the 'chosen' helper is passed through to them, rather than them figuring out which to use independently? (thinking, make the system-default case the same as the user-chosen rather than rely on flags and duplicating t12:11
IntuitiveNipplehe same discovery technique as the helper-chooser dialog)12:11
asacIntuitiveNipple: cant tell without actually looking at the patch12:15
asacIntuitiveNipple: sounds like its reasonable, but if it changes order/preference it might cause other issues12:16
IntuitiveNippleokay... don't worry about it... was just trying to avoid myself doing something counter to the intention of the original code :)12:16
asacIntuitiveNipple: well. the original code is buggy12:16
IntuitiveNippleIt does seem daft though, asking the user to confirm the helper and then, later when the download is complete, using a totally different method to determine the helper :)12:17
asacso most likely a cleanup is needed. however, i cant forsee what upstream thinks. they sometimes have wierd attitudes12:17
asacIntuitiveNipple: if thats the case then it needs to be fixed12:17
IntuitiveNippleLet me be clear then - that v3 patch is in debian already (and presumably going to land in ubuntu) ?12:17
asacIntuitiveNipple: as long as the initial helper selection still uses the mailcap/gnomevfs/user-pref way it should be fine12:17
asacIntuitiveNipple: no12:17
asacIntuitiveNipple: nothing that is in debian gets to ubuntu usually12:18
asacIntuitiveNipple: because the debian maintainer does a good job of hiding his patches and i dont bother to run after him12:18
IntuitiveNippleasac: agreed, my idea is to just propagate the initial choice (whether it be 'user' or 'default').12:18
IntuitiveNippleI see... okay, that makes sense then12:18
asacIntuitiveNipple: so whatever you develop i will look at and if thats good get that into upstream12:18
asacand into ubuntu12:19
IntuitiveNippleyeah.. so I'll work from upstream and worry about that v3 patch later then12:19
asacIntuitiveNipple: dont worry about that patch. if it also fixes the issue v3 addresses its fine12:19
IntuitiveNippleThanks... I'l let you know when the new patch is done.12:19
asacbut not needed12:19
asacthe issue you are after is more important12:19
asacIntuitiveNipple: you could look at debian changelog12:20
asacmaybe he already has a patch for that12:20
asacwhich we didnt find in bugzilla12:20
IntuitiveNippleasac: right, will do12:20
asaci somehow doubt that he didnt see this bug12:20
asacwhile he worked on the other12:20
* IntuitiveNipple thinks to himself, why would someone not use a patch system!?12:20
asacIntuitiveNipple: because he wants to stop ubuntu leeching on him12:20
IntuitiveNipple(rhetorical question, no answer required!)12:20
asaci dont bother ...e xcept that he kills the ability for me to provide security support in debian12:21
asacwhich de-facto means that debian will have no security support12:21
IntuitiveNipplekinda defeats the purpose of open-source really, doesn't it?12:21
asacIntuitiveNipple: well. his pov is probably that ubuntu does the same :)12:21
asacfor all the other packages12:21
asacIntuitiveNipple: his official argument is that he can better do it in git12:21
asacwhich is a private git unfortunately12:22
asacIntuitiveNipple: and even then its completely unfeasible as long as you dont maintain topic-branches per patch12:22
asacmozilla accumulates a bunch of patches that interleave12:22
asacand figuring out later where you did a merge error is cumbersome12:22
asacbut well12:22
asaci accepted that its that way ;)12:22
IntuitiveNippleWell, yes, I agree with the git statement (I important all source packages into git locally) *but* I use it to create the debian patches, by working in a test branch and then simply creating the debian patch with git-diff master >debian/patches/lpxxxxx.patch :)12:23
IntuitiveNipples/I important/I import/12:23
asacIntuitiveNipple: git is ok if the package will have no long-lasting patches12:23
asace.g. you can develop in git, even commit there12:23
asacas long as  either thats just a few patches or you submit it upstream and can backout on next sync12:23
asacbut thats usualyl not the case. especially when the security-backports happen12:24
IntuitiveNippleI do all patching outside of master, as debian/patches/* - git actually makes the process easier!12:24
asacIntuitiveNipple: thats good ;)12:24
IntuitiveNipplealso means I can cherry-pick commits or individual files from commits, really easily12:24
asacIntuitiveNipple: i use topic branches for "large" feature patches12:25
asacotherwise i just use quilt12:25
IntuitiveNippleI generally do a git-reset --soft HEAD^; edit; git-commit -a -C ORIG_HEAD; build_test process and when it works I do git-checkout debian-package (the branch), git-diff master..lpXXXXXX >debian/patches/lpXXXXXX.patch ...12:26
IntuitiveNippleand then update changelog and the patch control file, whatever system it uses, and then commit to the debian-package branch12:27
IntuitiveNipplemakes debugging a wide variety of packages much easier, especially when it is over an extended time-period and I'm liable to forget what was going on12:28
IntuitiveNippleright, I'll stop bothering you until the patch is ready! I'm back off to slave over the soldering iron :)12:30
hateballHello... I can see that status is undecided for https://bugs.launchpad.net/ubuntu/+source/firefox/+bug/230011 but is there any chance a fix will be out soon?13:44
ubottuLaunchpad bug 230011 in firefox "[HARDY] Firefox print failure, default papersize erroneously Letter in German install" [Undecided,New]13:44
hateballActually... it's all Non-US installs13:45
asacfta_: Jazzva: there?15:39
Jazzvaasac, yep15:40
asacfta_: Jazzva: after name discussion and a few more roundtrips we came to abrowser :)15:41
Jazzvagood enough :)15:41
asacfta_: Jazzva: unless someone has any real objections i would take it ;)15:41
asacmost likely we wont find anything better :(15:42
Jazzvawell... maybe a-browser would be more readable15:42
Jazzvathough, i'm still voting for u-browser :)15:42
asacJazzva: this was turned down because of ubuntu15:42
Jazzvaah.. ok15:43
asacdbrowser was turned down because of debian15:43
asaceven though ubrowser could mean unbranded browser and dbrowser "debranded"15:43
asacbut well15:43
Jazzvastripped-browser :)15:43
asacok. so no real objection from you ;)15:43
Jazzvanope :)15:43
asackgoetz: ^^15:44
Jazzvabtw, my provider decided to double d/u speed today :). 2048/256 at the moment :)15:44
Jazzva== faster pushes to bzr ;)15:45
Kamping_Kaiserasac, i cant think of a better name, if that helps ;) (so, no objection from me)15:46
* Kamping_Kaiser -> bed15:53
asacKamping_Kaiser: night16:02
asacthanks16:02
Jazzvaasac, so it's official name?16:02
asacJazzva: most likely. I am waiting for fta as well16:03
Jazzvamhm... i'll twitter it ;)16:03
asacJazzva: if you blog or do anything about it, better wait till its final16:06
asacand get the facts right ;)16:06
Jazzvasure :)16:06
asacdont make it appear that we "rename" firefox16:06
Jazzvathat's why I asked...16:06
asacits just that we provide a branding package16:07
asac@time16:26
ubottuCurrent time in Etc/UTC: September 10 2008, 15:26:14 - Next meeting: Ubuntu Mobile Team in 20 hours 33 minutes16:26
armin76asac: tested tracemonkey?16:27
asacno not yet16:27
asac@time16:28
ubottuCurrent time in Etc/UTC: September 10 2008, 15:28:09 - Next meeting: Ubuntu Mobile Team in 20 hours 31 minutes16:28
asacsorry. i am in meeting and unsure whether i am disconnected ;)16:28
asacso i use this as a ping :)16:28
* armin76 kicks asac 16:29
* armin76 kicks asac 16:29
asacarmin76: you could also be my pingee16:29
asac;)16:29
* armin76 kicks asac 16:29
asacarmin76: ping16:29
asachehe16:29
armin76bumb16:29
fta2asac, ok for me too16:37
asaccool16:37
asacthen we all the votes we need - finally :/16:38
armin76asac: bumb16:59
asacfta2: maybe https://bugs.edge.launchpad.net/ubuntu/+source/fontconfig/+bug/243130 is your lcd issue?17:16
ubottuLaunchpad bug 243130 in fontconfig ""/etc/fonts/conf.d/53-monospace-lcd-filter.conf", line 17: invalid constant used : lcdfilterlegacy" [Low,Confirmed]17:16
asacbut most likely its not17:16
fta2asac, it's kind of old17:28
fta2no it's not the one. the current problem is within cairo, lcd filter regressed upstream compared to our patch17:32
asacok ... off travelling  bbiw17:33
Jazzvafta_, one telecommunication question... any good advice how to remember/learn all those error probabilities? :)21:47
=== fta_ is now known as fta
ftaJazzva, hm, what ?23:16
Jazzvafta, for example probability of wrongly transmitted bit, or energy of noise, or ... something like that... :)23:55

Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!