Meeting on irc.gnome.org:#gtk-devel Meeting started December 13 2004 16:02 EST (21:02 UTC) In attendence: Manish Singh (yosh), Owen Taylor (owen), Matthias Clasen (mclasen), Anders Carlsson (andersca), Robert Ögren (roboros), Tor Lillqvist (tml), Douglas Pollock (desudation), Billy Biggs (vektor) hey. I have to break off at 5 sharp today, so I want to start on time I had posted a number of agenda items yesterday, do we need to discuss other things ? if not, lets start with the 2.6 blocker list. the one bug I am currently working on is 152483 Improper alignment of icons in GtkIconView that's great once that is committed, we only have the win32 filename encoding bug and a menu keynav bug left on the mustfix list I think tml committed the necessary changes for the filename encoding bug, or are there outstanding issues, roboros ? there is some remaining issues with the filechooser functions that return uris for abi stability either they need wrappers or the g_filename_from_uri need wrapper if I understood tml correctly does win32 generally look good, apart from that one issue ? the module loading problems have been cleared up, right ? yeah, much better after tmls commit ok, I'll ping tml about the uri issue pango's configure.in probably needs to be updated to require glib 2.5 for the gstdio stuff the menu keynav bug should not block 2.6, i think. It is unfortunate for billh that we punt it again, but he got some fixes since last week... roboros: I can do that later tonight, I need to commit a pango configure patch anyway roboros: There's another (unrelated) patch that bumps the glib required version and I gave mclasen the commit-goahead ofr that great :) Jimbob: how does the filechooserbutton-gone-combobox patch look ? Can we get that committed in the next few days ? Yeah did you experiment with ellipsising only that one row ? I need to fix up the model (it sucks pretty hard right now), but it should be done w/i a few days time. mclasen: Yeah, worked fine nice We should probably ellipsize the the bookmarks as well Well, I should :-) I'll wait for that patch then...my current thinking is to do some doc cleanups over the next few days, and start the releases around Thursday I'm trying to to get Pango ready for 1.8 on that schedule as well did we miss any important bugs which should be on the blocker list ? andersca: you had a cell view api cleanup...do you still want to get that into 2.6 ? mclasen: I'd like that, but it's not finished yet mclasen: and I think it's necessary mclasen: one alternative would be to just remove the set_ functions for now and add them in 2.7 andersca: we're talking about the setters which set cell renderer properties and also queue a resize+redraw on the cell view, right ? yeah andersca: if we can't get the cleanups done this week, making those functions private for now would be ok IMO we should probably document the need to queue a resize+redraw in the cell view docs somewhere in that case hey tml Dec 13 16:21:03 * tml says hello tml: we were talking about bug 159719 earlier any conclusions? no, basically wondering whether we should wait for further gdk/win32 filename encoding changes before doing 2.6 further gdk/win32 changes wonät be visible in headers etc, and thus don't imho need to block 2.6 ok. I could still wait for them...(my current plan is to start distchecking on Thursday) tml, no header changes for the uri encoding stuff? Dec 13 16:26:02 * rambokid whishes gtk team meetings would occour o tuesdays only. roboros: no, shouldn't be tml: now that you are here, maybe we can discuss g_rmdir() is there btw any kind of standard that would descirbe what file: uris look like, especially what encoding they use? Or any syntax besides the %XX single-byte-as-hex, to encode a unicode code point? (something like \u{xxxx} or whatever)? there are various RFCs covering uris and the file: scheme, I don't have the numbers at hand though g_rmdir was left out from gstdio by mistake, but on the other hand now that g_remove does rmdir(), too, it isn't that necessary. just for completeness it would be nice tml: file uris are horribly ill defined, especially on windows tml: On Unix, we've sort of converged on the principle of using % encoding of the bytes on disk On the principle that every file should be representable as an URI, and that figuring out display of misencoded filenames can be left up to the app just like filenames encountered on disk tml: But that doesn't really work as a principle on Windows. * tml tests what IE does it seems nobody has strong objections against g_rmdir() * tml wonders how to get IE to show file: URIs... ok, moving to the next topic...pool milestones I have moved all glib bugs to the pool milestones over the past weeks But that is somewhat boring work, and gtk+ has quite a few more bugs, so I would appreciate if people who feel responsible for some bugzilla component could help with sorting the bugs into the pool milestones * tml promises to do once we have all bugs on the pool milestones, we can start thinking about moving them back onto 2.8 milestones ie do some planning for 2.8 btw, is there some pango release also coming up, or is pango on an unrelated schedule? tml: I'm going to try to get 1.8 out this week Should we set some timeframe for getting a 2.8 plan together ? mclasen: I guess the question is how many people will be around next week .. whether we should kill that meeting and do planning after the holidays tml: what's your thinking on bug 140199? owen: I was about to propose to call it a year and do 2.8 planning in the first half of January printing is really something i haven't spent many thoughts on... tml: Well, GTK+ plan there is definitely Cairo tml: Though there's certainly a few things to be sorted out for that with respect to Win32 * mclasen sneaks out of the meeting... i guess if it makes jpe happy, something like the suggested patch could be accepted if clearly earmarked as being temporary, and not for general use? (But then jpe probably builds his own versions anyway, with whatever added features the official sources don't have?) the HDC stuff seems to be most useful for someone who either a) desperately needed to get printing working in the short term or b) was using Pango without GTK+ tml: Every extra bit in the API is something that has to be carried forward... ok, my lift is about to leave, so to recapitulate: I'll try to produce 2.6 releases by the end of this week, we won't have a meeting next week, since Chrismas gets close, and planning the 2.8 release will be the topic of the first meeting in 2005 see you all then... mclasen: ok. have a nice lift trip ;-) later mclasen seeya tml: Hmm, http://bugzilla.gnome.org/show_bug.cgi?id=161205 just came in ... is it possible that underlines are comprehensively broken in current CVS GTK+? hmm, what's a quick way to check, where should one see underline? tml: Well, menus, and there are a bunch in the 'labels' test in testgtk sigh, yes, no underlines in the "labels" test tml: OK, probably not that hard too fix ... the win32 backend should have been able to simple use the fallback implementation of draw_trapezoids in GdkDrawable tml: so probably something minor going wrong with that yeah, can't be anything major tml: I'll let you make the call on jpe's additions... though if not in in the next few days, then hopefully we'll have cairo support before the next opportunity to add them OK, let's declare this meeting done ====================== Meeting ended December 13, 17:05 EST (22:05 UTC)