Meeting on in http://yarrr.gnome.org/wiki/index.php/Gtk/meeting-7-12-05 Meeting started July 12 2005 16:00 EDT In attendence: Federico Mena Quintero (FedericoMenaQuintero), Jonathan Blandford (JonathanBlandford), Luis Villa (LuisVilla), Owen Taylor (otaylor), Alexander Larsson (alex), Matthias Clasen (mclasen), Billy Biggs (vektor), Maciej Katafiasz (mathrick), Johan Dahlin (jdahlin), Sebastian Bacher (seb128) Agenda ------ * 2.8 status * Pango/Cairo update (otaylor) * GObject thread safety/atomic refcounting * 2.8 bug list Summary ------- * owen to release pango * matthias to do another set of 2.7.x releases by the end of the week * matthias to "test commit" atomic refcounting * alex to file a "fuzzy expander" bug [DONE] Log --- 15:29 FedericoMenaQuintero dum de dum 15:31 FedericoMenaQuintero tra la la (bkor) 15:33 JonathanBlandford Federico, use http://yarrr.gnome.org/wiki/Gtk/meeting-7-12-05 15:34 JonathanBlandford FedericoMenaQuintero: don't use the :8080 page 15:34 FedericoMenaQuintero am I here now? (LuisVilla) 15:34 JonathanBlandford FedericoMenaQuintero: yes. 15:36 LuisVilla is agenda above or below? 15:36 otaylor I think the agenda above is supposed to be the right one 15:37 * jdahlin notes that yarrr looks pretty good 15:38 LuisVilla Oops. 15:38 LuisVilla should have used cancel there. 15:38 mclasen does bugzilla work for anybody else ? 15:38 JonathanBlandford mclasen: WORKSFORM 15:40 LuisVilla mclasen: broken here, and AFAICT the bugzilla user on window doesn't have permission to read the bugzilla.gnome.org httpd logs :/ 15:40 LuisVilla jrb: you've tried a query? 15:40 otaylor Apparently show_bug.cgi works but bug_list.cgi doesn't 15:40 alex ugh. constantly reloading the wiki is pretty sucky. Why not use the comments for it? 15:41 mclasen so we probably won't discuss specific bugs again... 15:41 otaylor LuisVilla: You aren't in gnomeweb? 15:41 LuisVilla might be 15:41 LuisVilla I think I know what the problem is 15:42 LuisVilla though I have no idea what might have caused it 15:42 LuisVilla gimme a sec 15:42 otaylor LuisVilla: Nothing useufl in the logs anyways 15:42 LuisVilla OK, fixed 15:42 LuisVilla the permissions got changed somehow 15:42 LuisVilla which sort of scares me (alex, JonathanBlandford) 15:44 otaylor LuisVilla: Quite likely someone typed 'cvs up' as bugzilla rather than runing the update script 15:44 otaylor Or at least, that used ot be the reason for that happening 4 years ago 15:45 LuisVilla I thought we fixed that 15:45 LuisVilla anyway 15:45 LuisVilla poking around the bug team for that 15:45 FedericoMenaQuintero do we need gdk_pixbuf_set_as_cairo_surface(pixbuf, cr) or something like that? 15:45 LuisVilla we have a guilty party 15:45 otaylor Federico: gdk_cairo_set_source_pixbuf() 15:46 alex otaylor: maybe FedericoMenaQuintero means to draw on a pixbuf 15:47 otaylor alex: Ah. That's harder ... really needs libpixman support for pixbuf formats 15:47 otaylor alex: It would presumably be gdk_cairo_create_for_pixbuf() if we implemented it 15:49 mclasen that would be very nice, we would finally have a drawing api for pixbufs (otaylor, alex, JonathanBlandford) 15:51 FedericoMenaQuintero yeah 15:52 otaylor mclasen: Of course, cairo takes away a lot of the need to draw to a pixbuf 15:53 otaylor Check the green checkmark on this sentence if you have a green checkmark (FedericoMenaQuintero, LuisVilla, otaylor, alex) 15:54 otaylor I actually wanted to write "Check the green checkmark on this sentence if you don't have a green checkmark", but that wouldn't work very well... 15:55 vektor Presumably I shouldn't have to keep hitting reload to see new messages, no? 15:56 otaylor vektor: No, you shouldn't. (The page should be obnoxiously reloading) 15:56 otaylor vektor: I wonder if that could be a proxy issue 15:56 vektor Seems to be failing for some reason then. :) 15:56 mathrick FWIW, I am behind proxy 15:56 mathrick rather poorly configured one, too 15:57 JonathanBlandford testing (LuisVilla) 15:57 mclasen I don't have a green checkmark, but maybe we should st art the meeting by having jrb give us directions on where to report our feedback on yarrr 15:58 mclasen jrb ? 15:58 mclasen ah, you are JonathanBlandford here 15:58 mclasen does nick highlighting work here ? 15:58 JonathanBlandford http://yarrr.gnome.org/wiki/Yarrr/Feedback 15:58 * mathrick sez WikiLogins are not good nicks for chat 15:59 JonathanBlandford mathrick: agreed. 16:00 mclasen I hope that we have most people here now 16:00 alex LuisVilla: can you close or cancel the haiku comment? 16:01 mclasen do we have other agenda items in addition to the ones listed above ? 16:01 JonathanBlandford (it was a lovely haiku, BTW) 16:01 jdahlin mathrick: nah, irc nicks are just confusing 16:01 jdahlin mathrick: nah, irc nicks are just confusing 16:01 mathrick jdahlin: ButTheyDontTakeHalfTheScreen 16:02 LuisVilla hrm 16:02 JonathanBlandford mathrick: only authors can close comments 16:03 JonathanBlandford mathrick: so you need to close it now. (-; 16:03 LuisVilla I have a close button but can't seem to actually close it :) 16:03 LuisVilla oh 16:03 mathrick fuck 16:03 LuisVilla all authors need to agree to cancel it? 16:03 otaylor LuisVilla: You are crossed off. I think mathrick has to close it as well 16:03 LuisVilla interesting 16:03 mclasen if there are no other items, then we should maybe start with the status of Pango/Cairo 16:03 mathrick JonathanBlandford doesn't seem to tab-complete here 16:03 otaylor mathrick: case sensitive 16:03 mathrick otaylor: ah 16:04 JonathanBlandford mathrick: cancel! 16:04 JonathanBlandford kaboom 16:04 mathrick JonathanBlandford: done 16:04 mclasen otaylor: will you do a new Pango release soon ? 16:06 mathrick JonathanBlandford: feedback page has gone blank, after several clashing revisions 16:07 LuisVilla reload 16:07 mclasen owen: what is the state of cairo ? can we expect another release soon ? I think there were some small api additions/changes, right ? 16:07 otaylor mclasen: Yes. It's gotten sort of hung up on cairo changes, but definitely by the end o fthe week 16:08 otaylor the plan is cairo-0.6.0 and pango-1.9.1 pretty much together 16:08 otaylor mclasen: There was the CAIRO_FORMAT_* vs. CAIRO_CONTENT_* incompatible API change in CVS recently 16:08 mclasen otaylor: so what are the main pango 1.10 blockers ? 16:08 otaylor The Pango release will involve some (baiscally compatible) addiitons 16:09 * mclasen should check the "1.10 Freeze" milestone 16:09 otaylor mclasen: - Propagation of XSETTING options (what I'm working on now) 16:09 otaylor - The rotated font caching problem 16:10 * mclasen notices 1.10 Freeze is virtually empty... 16:10 otaylor Then there are various "1.8.2" blockers that I'm trying to get to ... bug fixes with patches I've been trying to get to for a while, and also FreeType-2.2 compat stuff 16:10 mclasen otaylor: did you look at the printing issues yet ? 16:11 otaylor mclasen: No, I haven't 16:12 otaylor it's probably something simple... there have been basically no changes in the code paths involved there, so something probably got accidentally messed up 16:13 mclasen ok, maybe me or J5 can look at it 16:14 otaylor mclasen: It's probably just a few minutes for me, if I can find the time to look at it. But if you want to take a look, feel free. 16:14 mclasen otaylor: ok 16:15 mclasen so, the next topic is the gobject thread safety/atomic refcounting bugs 16:16 mclasen many people have expressed strong interest in getting at least the atomic refcounting patch in 2.8 16:16 mclasen and we have a new patch from wim, which incorporates the result of the timj<->wimt discussion at guadec 16:17 mclasen but rambokid told me today that he is too busy with university to look at the patches for at least two more weeks 16:17 alex can anyone else do it? 16:17 mathrick mclasen: s/wimt/wtay/ 16:18 mclasen everybody else can look at the patches and comment on them, sure 16:18 alex it would be very nice to have, more and more gnome apps are starting to interact with threads 16:18 mclasen but it fear that rambokid will not be happy if we commit them without his explicit approval 16:19 alex this is bug 166020 btw 16:20 mclasen it would be nice if we could embed a bug list in the wiki, next to the 16:20 JonathanBlandford mclasen: yeah. That would be cool. 16:20 LuisVilla federico and I talked about that a bit 16:21 LuisVilla last night 16:21 LuisVilla I'd like to be able to do some more wiki-ish things with bugzilla 16:21 * mclasen noticed he shot himself in the foot by distracting the discussion from gobject thread safety to bugzilla/wiki integration (LuisVilla) 16:23 LuisVilla (even more sidetrack: I was trying to 'check' mclasen's last comment, but the icon changed to a X- did I just check him or cross him?) 16:23 alex LuisVilla: checked, click the cross to remove your check 16:24 mclasen would it be workable to commit the atomic refcounting changes now, to get them tested in 2.7.x, and ask timj to review them before 2.8 final ? (JonathanBlandford) 16:24 alex mclasen: he does seem to agree that we want it 16:25 alex mclasen: and it feels safer to do that early and revert later than the other way around 16:26 mclasen ok, lets try that route 16:26 JonathanBlandford mclasen: as an aside, your color changes never made it in, right? 16:26 mclasen JonathanBlandford: no, instead I got cursor changes in 16:26 mclasen JonathanBlandford: which seemed less controversial 16:26 JonathanBlandford mclasen: ah, okay 16:26 JonathanBlandford mclasen: yeah. 16:26 JonathanBlandford though less useful 16:27 alex what color changes? 16:27 mclasen hey, they are maybe useless, but now you can switch to cool cursor themes ! 16:27 mclasen alex: symbolic colors, recolorable themes, color tab in appearance capplet... 16:27 mathrick mclasen: wasn't that possible before? 16:28 mclasen mathrick: not without restart/re-login 16:28 mclasen mathrick: we didn't have an xsetting for cursor theme 16:28 mclasen just an envvar 16:28 JonathanBlandford mclasen: didn't mean to criticize -- both are pretty cool. 16:28 mathrick mclasen: ah, and now it's instant effect? 16:28 mathrick 00t 16:28 mathrick w00t even 16:29 mathrick I hate the text-eating latency of yarrr 16:30 mclasen ok, so coming back to the general 2.8 status topic 16:30 LuisVilla (I added a bug query link to the agenda, if you still want that, mclasen) 16:30 mclasen LuisVilla: thats not quite what I imagined. More like an embedded bug list 16:31 mclasen so that I can read bug summaries without clicking away from the discussion 16:31 LuisVilla oh, yeah 16:31 LuisVilla I didn't mean that it fulfilled that role 16:31 mclasen 310133 is on that list, btw 16:31 LuisVilla just that well before the meeting started you'd said you were trying to get a bug list, when bugzilla was broken 16:32 mclasen ah, ok 16:32 mclasen I committed to small GOption api additions today, and I don't have any other api affecting bugs planned for 2.8 anymore 16:32 LuisVilla <- so embarassed he closed his browser tab 16:33 mclasen So I think the next 2.7.x releases will be "officially" api frozen, unless somebody comes up with an urgent api issue today... 16:34 mclasen owen: did you get a chance to look at plug/socket/win32 ? 16:34 mclasen err, otaylor 16:36 otaylor mclasen: I'm about halfway through 16:36 otaylor It's a 60 page patch, printed out 16:36 alex woa! (LuisVilla) 16:37 otaylor alex: Well, it's a good chunk code reorg, but still not easy to review 16:37 alex did anyone report a bug about the expanded treeview expander being fuzzy? 16:38 mclasen it has been discussed recently 16:38 otaylor alex: It's not in all themes 16:38 mclasen probably depends on the expander size 16:38 otaylor alex: But I don't think there is a bug report on it filed 16:38 alex otaylor: should i file one? 16:39 mclasen alex: yes 16:42 mclasen today I added a demo for iconview editing/dnd to gtk-demo. 16:42 mclasen if anybody else has a nice demo for a new 2.8 (or undemoed 2.6) feature, please add them 16:43 mclasen I think it would be nice to have all important new features demoed in gtk-demo 16:43 FedericoMenaQuintero hmm, why is 166601 milestoned for 2.8? 16:44 mathrick mclasen: it'd be also good to review gtk examples, the code there is getting very bitrotten, and distros still ship it 16:45 mclasen FedericoMenaQuintero: I believe I put it there because it had a patch which I wanted to not forget 16:45 mclasen FedericoMenaQuintero: but now the patch is committed, and the issue is still there 16:45 mclasen FedericoMenaQuintero: so the bug should probably go back to "Medium Fix" 16:46 mclasen mathrick: do you think about the tutorial ? 16:46 JonathanBlandford GNOME bug 166601 16:46 mathrick mclasen: probably, I think that's where the code is scrapped from 16:46 mclasen mathrick: ah, yes, those examples 16:47 mclasen mathrick: I think the only hope for an uptodate tutorial 16:47 mclasen is to integrate it into the api docs, as we did for the gobject tutorial 16:48 mclasen we have basically ignored the tutorial since 2.0, since tony gale kind of owns it 16:48 mathrick mclasen: yes, but at least #define GTK_ENABLE_BROKEN code should die :\ 16:50 mclasen yes, but its a lot of work to clean up/update the tutorial 16:50 mclasen anything else to discuss ? (I'm heading out in 5 minutes) 16:52 mclasen federico: did you look further into overwrite confirmation ? 16:52 mclasen err, FedericoMenaQuintero 16:53 * alex just filed bug 310173 16:54 alex (bug 310172 is for the fuzzy expander) 16:55 LuisVilla one last thing 16:55 LuisVilla 2.8 for 2.12 16:55 alex ah, the big question... 16:55 mclasen LuisVilla: has that been agreed on now ? 16:55 LuisVilla I'm fairly satisfied ATM; I'm using it and not seeing anything that is particularly scary and/or that isn't getting fixed quickly 12:56 LuisVilla no, it hasn't, but I can presumably push it that way 12:56 seb128 same here 12:56 alex It seems that we have decided to at least not depend on gtk 2.8 only features atm 12:56 LuisVilla I assume no one on the gtk team objects? 12:56 otaylor LuisVilla: Nothing that I haven't said before 12:56 mclasen LuisVilla: we (at least I) are working hard on making that possible, so no 12:56 JonathanBlandford LuisVilla: I've got conditional dependencies in evince on 2.8, but I'd much rather ship with 2.8 12:57 seb128 (some) distros would like to get an officiel rt call to know what to package :) 12:57 LuisVilla heh 12:57 LuisVilla that distro is young and hasn't learn to take decisions by the balls yet ;) 12:57 mclasen release team seems barely existant at this point... 12:57 seb128 ah ah 12:58 LuisVilla mclasen: different discussion, but yeah, it does seem fairly poor form for a team whose lead is AWOL to be criticizing the gtk team :/ 12:58 otaylor seb128: Once GTK+-2.8 is out, the official GTK+ position will be that you should package GTK+-2.8; we won't continue a GTK+-2.6 maintainence branch past then 12:58 mclasen I plan to do a final 2.6.x release at the same time I do 2.8.0, though 12:59 seb128 otaylor: yeah, but if we start using 2.7 now and for whatever reason the rt decide that GNOME 2.12 ships with 2.6.n we are screwed since downgrading is not trivial for a distro 12:59 * mclasen never new the release team had a lead... 13:00 LuisVilla jeff 13:00 LuisVilla no further comment 13:00 LuisVilla :) 13:00 seb128 ah ah 13:00 LuisVilla but yeah, the rest of the team has not exactly picked up the pieces 13:00 LuisVilla anywya 13:00 LuisVilla different discussion 13:01 mclasen anyway, lets close this meeting before we open that discussion (alex, JonathanBlandford) 13:01 * mclasen wanders away 13:01 LuisVilla night 13:01 mclasen night 13:02 seb128 'night 13:04 JonathanBlandford (as a reminder, comments go here: http://yarrr.gnome.org/wiki/Yarrr/Feedback ) Meeting ended around 17:00