Mar 07 16:01:35 I'm writing a new ridley/2.10 blog entry as we speak Mar 07 16:01:38 was about t ime again Mar 07 16:01:38 * mclasen looks at the agenda Mar 07 16:02:04 ok, one thing jrb asked me to bring up is schedule Mar 07 16:02:27 ah yes Mar 07 16:02:30 we have been kindof vague about it, so far, saying to target April/May for a 2.10 release Mar 07 16:02:30 we are in march already Mar 07 16:02:39 right Mar 07 16:02:51 and I guess we really want libglade in too ... Mar 07 16:02:56 so, maybe we should go over the major outstanding things that everybody wants to get in Mar 07 16:03:30 so, now that fc5 is basically done Mar 07 16:03:55 I will return to recent-files, which have been unfairly neglected for a long time Mar 07 16:04:05 great Mar 07 16:04:21 --- ebassi|dinner is now known as ebassi Mar 07 16:04:31 I hope that we should be able to use the big amount of work done by ebassi, and get that finished in 2 weeks Mar 07 16:04:32 mclasen: :-) Mar 07 16:04:54 after that, I need to return to introspection Mar 07 16:05:07 that is basically my plan for the remaining time until gtk 2.10 Mar 07 16:05:14 --> jdahlin (~jdahlin@200-171-140-32.dsl.telesp.net.br) has joined #gtk-devel Mar 07 16:05:51 what are the other big pieces for 2.10 ? Mar 07 16:06:08 tooltips Mar 07 16:06:11 working on that Mar 07 16:06:21 hope to get a new/updated proposal out this week Mar 07 16:06:31 async filechooser, and the review is about done iirc Mar 07 16:06:40 --> gjc (~gjc@bl6-38-113.dsl.telepac.pt) has joined #gtk-devel Mar 07 16:06:46 printing, too Mar 07 16:06:54 async filechooser will need some more rounds, or how does it look ? Mar 07 16:07:03 yeah, if we can mention printing we can also mention libglade ;) Mar 07 16:07:07 some tree view thigns Mar 07 16:07:09 gtkbuilder if not already mentioned Mar 07 16:07:11 eh Mar 07 16:07:20 and if I have time left, I want to get multi row DnD in Mar 07 16:07:32 mclasen: small API change, some polishing i think Mar 07 16:07:41 mclasen: i haven't seen the full review yet Mar 07 16:07:42 I'm not sure if printing will be sufficently done by april, will it, jrb ? Mar 07 16:08:12 mclasen: I don't know. That's up to Alex. I was hoping he'd be able to present something today Mar 07 16:08:32 well, he sent the mail on Friday Mar 07 16:08:32 they seem to be making a lot of progress though Mar 07 16:08:48 as for canvas, I guess we need to look at libccc Mar 07 16:08:53 probably 2.12 candidate though Mar 07 16:08:57 yes, definitively Mar 07 16:09:06 canvas seems a longer shot to me Mar 07 16:09:14 but the canvas also doesn't do as badly externally Mar 07 16:09:14 jdahlin: libglade will be renamed gtkbuilder? Mar 07 16:09:34 jrb: true, but would be really nice to have in gtk+, together with printing Mar 07 16:09:46 kris: maybe, the name libglade is just rather unlikely Mar 07 16:09:53 jdahlin: yeah true Mar 07 16:09:53 --> AME (~chatzilla@portal.extratech.com) has joined #gtk-devel Mar 07 16:10:21 so, how do we schedule this to ensure that we are in good shape when May comes ? Mar 07 16:10:26 +1 for functional names Mar 07 16:10:49 mclasen: i guess all big stuff should be in before April ~15 ... Mar 07 16:10:58 is it realistic to target end of March for having all big pieces in basically committable shape ? Mar 07 16:11:01 then allow one month for polishing bug fixing Mar 07 16:11:04 --> federico (~federico@dsl-201-128-39-153.prod-infinitum.com.mx) has joined #gtk-devel Mar 07 16:11:14 and release just before the end of May? Mar 07 16:12:03 I'd prefer a little later, otherwise the gtkbuilder work will be a little tight Mar 07 16:12:40 * kris wouldn't mind slipping to the first half of June Mar 07 16:12:47 mclasen: do you want introspection in? I can help, but don't think API freeze on April 15 would be feasible.. Mar 07 16:12:59 why not tie the schedule to gnome 2.16? Mar 07 16:13:20 I would rather not tie the schedule to any gnome release Mar 07 16:13:33 because gnome devs always complain "gtk+ not stable" :P Mar 07 16:13:36 jdahlin: we need a lead time for 2.16 Mar 07 16:13:43 jdahlin: but june as a target isn't too bad Mar 07 16:14:00 IIRC we tied to the gnome schedule with 2.4 Mar 07 16:14:04 gjc: I want, but we'll have to see how easy finishing it will be (some ugly things are still missing, like a header parser...) Mar 07 16:14:08 and I disappeared during the 2.4 cycle (sorry for that) Mar 07 16:14:12 jrb: when is 2.16 going to freeze? Mar 07 16:14:19 (but I still think 2.4 was kind of rushed because of the gnome released) Mar 07 16:14:24 late july? Mar 07 16:14:28 <-- AME has quit (Chatzilla 0.9.67+ [SeaMonkey 1.0/2006013012]) Mar 07 16:14:42 basically, we need a stable .0 release out before gnome freezes Mar 07 16:14:48 jdahlin: I'd really like to target May for a release. If we slip a little bit, and if gtkbuilder is a bit later than the other big pieces, so be it Mar 07 16:14:53 but we should try Mar 07 16:15:33 let's just target end of May then, with a possible slip to first half of June? Mar 07 16:15:34 mclasen: the api can get in earlier, just not widget support for everything Mar 07 16:16:05 jdahlin: thats fine Mar 07 16:16:43 great Mar 07 16:16:46 lets move along the agenda (I'm a bit time-challenged today) Mar 07 16:16:54 as always ;) Mar 07 16:17:03 kris: wait till you have a family... Mar 07 16:17:15 just a last note, adding interfaces to objects is probably breaking the api Mar 07 16:17:29 mclasen: ;) Mar 07 16:17:48 jdahlin: you need to make sure you handle the case of missing implementations Mar 07 16:17:59 --> muntyan (~muntyan@pool-71-113-224-59.herntx.dsl-w.verizon.net) has joined #gtk-devel Mar 07 16:18:26 mclasen: true, good workaround Mar 07 16:19:23 ok, we already talked quickly about the async filechooser status. Do you have anything to add, federico ? Mar 07 16:19:26 kris: I just mailed you what I have so far to your @imendio address. this is taking way longer than I expected. Mar 07 16:20:17 great, thanks Mar 07 16:20:48 ok, notebook signal problems were already discusses, we can skip that Mar 07 16:20:49 mclasen: each of the big patched files is taking me half a day to check :( I'm happy that it seems to work, modulo a few bugs. Mar 07 16:21:06 I plan to do fixes on the branch before we merge, but I guess that won't surprise anyone Mar 07 16:21:30 federico: ok, then I guess we'll just have to wait until your review gets finished Mar 07 16:22:03 my personal agenda item is bug 33360, which is about GtkAboutDialog Mar 07 16:22:20 as for -Wl,--export-dynamic, the easy and good solution now is simply to remove gmoedule-2.0.pc and inform people to use gmodules-no-export-2.0.pc by default (in NEWS probably). that should fix the issue good enough. rationale was presented on gtk-devel-list today. Mar 07 16:22:39 mclasen: you are missing a "3" Mar 07 16:23:02 kris: too many 3's in bugs lately (I managed to file bug 333333...) Mar 07 16:23:08 haha Mar 07 16:23:40 rambokid: that has the potential to break the build of a lot of things though, doesn't it ? Mar 07 16:24:22 if they have gmodule-2.0 in their list of required pc files... Mar 07 16:25:51 what about having gmodule-2.0 and gmodule-export-2.0 ? Mar 07 16:26:08 yes, that's what i was thinking too Mar 07 16:26:11 jdahlin: currently we have gmodule-2.0, gmodule-no-export-2.0 and gmodule-export-2.0 Mar 07 16:26:13 mclasen: that is the whole point, really. because the third party packages need to switch to gmodules-no-export-2.0.pc. if you'd want a more subtle change, you could take the flag out of gmodule-2.0.pc and "just" deprecate it. that will break very few builds, but the few builds that break will be kinda hard to figure for some maintainers. Mar 07 16:26:56 rambokid: I guess I'm just as conservative as owen... Would be nice if pkgconfig had a means of deprecating .pc filess Mar 07 16:27:18 mclasen: so it's break all gmodule users in obvious ways (and have them read NEWS as a result), or break a few in a hard-to-figure manner. i prefer the first option because allthough some people won't like it, they won't have to spend endless hours debugging something stupid and hard to figure. Mar 07 16:27:56 it's not that hard to figure Mar 07 16:27:57 mclasen: yeah, like have it spew a message if you usd a .pc file... kinda like the linking warning you get for mktemp() nowadays ;) Mar 07 16:28:29 what about users building soft released a year ago, with dead maintainer? Mar 07 16:28:50 if applications have g_warning("plugin import failed %s", g_module_error()) when importing plugins, it will be quite obvious what's going on Mar 07 16:29:30 gjc: it depends on where it breaks. if you depend on -Wl,--export-dynamic somewhere and it's suddenly gone, the exact error you're getting depends much on your code and build setup (could cause errors only during runtime even) Mar 07 16:29:32 <-- jrb has quit (Read error: 104 (Connection reset by peer)) Mar 07 16:30:39 gjc: so in general, it's very hard to figure. removing just gmodule-2.0.pc is very easy to figure though, and definitely breaks during the build process (at configure time even) Mar 07 16:31:10 and doing that in 2.10.2 sounds ok to everybody ? Mar 07 16:31:37 i guess this is only a problem to distributions.. Mar 07 16:31:45 no strong opinion here,. Mar 07 16:32:05 --- jdahlin is now known as jd-afk Mar 07 16:32:21 mclasen: didn't you mean 2.10.1? that is the next release, right? Mar 07 16:32:32 rambokid: 2.10.1 was released today Mar 07 16:32:56 mclasen: i'd really like to get owen's opinion on whether it's ok to do that during a stable branch... Mar 07 16:33:24 mclasen: wow, you already threw it out? (haven't fixed gtk+-2.8 yet even) Mar 07 16:33:51 rambokid: I'll send mail to owen and cc you Mar 07 16:34:00 mclasen: i cc:-ed him already. Mar 07 16:34:04 rambokid: cool Mar 07 16:34:21 kris: mailed you again - I just finished with gtkfilesystemmodel Mar 07 16:34:24 ok, I'm on the way out Mar 07 16:34:30 kris: damn, dude, you are smart. That file scares me to death. Mar 07 16:34:40 feel free to continue without me, sorry for being so short on time... Mar 07 16:34:44 mclasen: you'll put up the log, right? Mar 07 16:34:52 I'll make an honest attempt Mar 07 16:34:59 good. Mar 07 16:35:11 i can put the log online Mar 07 16:35:13 --- You are now known as mclasen_away Mar 07 16:35:18 federico: rewriting the _path_do function there was quite an experience Mar 07 16:36:16 kris: yeah, it's really hard to understand what it is supposed to do Mar 07 16:36:26 and it kills me that right now we don't really use it :) Mar 07 16:36:46 yeah Mar 07 16:36:51 we do at only 2 places or something Mar 07 16:36:53 ok, I have to run to lunch Mar 07 16:36:56 it was one of the last things I fixed Mar 07 16:36:59 have a nice meal Mar 07 16:37:05 kris: I mean, we don't use the multi-level tree right now Mar 07 16:37:13 federico: true Mar 07 16:37:21 but I want to replace ACTION_SELECT_FOLDER with a folder tree Mar 07 16:37:25 I did bring up a file chooser with the multi-level tree once Mar 07 16:37:28 to remove the ambiguities in "which folder am I selecting?" Mar 07 16:37:33 oh, nice Mar 07 16:38:57 ok, bbl Mar 07 16:39:17 <-- johannes_ has quit (... und tschüß) Mar 07 16:39:17 kris: I hope I can be done with at least everything but gtkfilechooserdefault.c today Mar 07 16:39:57 federico: would rock Mar 07 16:40:07 ok, I guess we discussed everything but bug 333360 now Mar 07 16:40:09 (that one also scares me) :) Mar 07 16:40:23 i guess we can move 333360 to next week and close this meething? Mar 07 16:40:25 meeting even Mar 07 16:40:48 --> ijuma_home (~ijuma_hom@ismael.plus.com) has joined #gtk-devel Mar 07 16:40:54 --- federico is now known as f_lunhc Mar 07 16:40:59 --- f_lunhc is now known as f_lunch Mar 07 16:42:20 Bug #333360, reported by dsd@gentoo.org: gtk+/2.8.x (UNCONFIRMED): GtkAboutDialog deletes upon delete_event but hides upon close button click - http://bugzilla.gnome.org/show_bug.cgi?id=333360 Mar 07 16:42:39 now you can close the meeting Mar 07 16:45:11 ok then Mar 07 16:45:16 we're done for today Mar 07 16:46:31 kris: seems so :-)