Meeting on irc.gnome.org:#gtk-devel
Meeting started July 19 2005 16:01 EDT


In attendence:

Matthias Clasen (matthias)
Kristian Rietveld (kris)
Maciej Katafiasz (mathrick)
Owen Taylor (owen)
Tor Lillqvist (tml)
Jonathan Blandford (jrb)


<matthias>	ok, so we are in good old irc again
<kris>		yay irc!!
<matthias>	are there any things we need to discuss ?
<matthias>	here is my own list:
<matthias>	plug/socket 
<matthias>	status of pango
<matthias>	pango/printing
<matthias>	some bugs: 300089, 173098
<matthias>	hey owen, federico__, I just outlined my topics for today:
<matthias>	plug/socket , pango, pango/printing, bugs: 300089, 173098
<kris>		(wanted to add that i will be on holiday from Jul22 until Aug8)
<matthias>	I will be away from Jul 26 or 27 to Aug 1
<mathrick>	matthias: random mumbling, but would adding "possible 6 month release cycle" to that make sense?
<matthias>	sure, we can discuss the next release cycle, but I'm not sure discussing timeframes makes  much sense without nailing targeted features first
<matthias>	ok, if there are no further topics, maybe we can start with plug/socket. 
<matthias>	owen: is there a chance to get the plug/socket patch reviewed/committed this week ?
<owen>		"blocking on me finishing Pango stuff"
<owen>		matthias: yes
 *		tml is grateful
<owen>		matthias: I have the Pango/Cairo stuff pretty much working happily, so I should be able to turn my attention to that soon
<matthias>	owen: nice, thanks
<matthias>	owen: does that include the hinting/dpi stuff ?
<owen>		matthias: Yeah, that's basically what it is
<owen>		I have that fixed (and the 180 rotation as a side effect)
<matthias>	owen: I had a very quick look at libgnomeprint, but couldn't immediately see why printing does not work
<matthias>	maybe I'll have another go at it, or I'll leave it to you. I guess that would qualify as a blocker... 
<matthias>	but as you said, its probably easy to fix if you know what to look for
<owen>		matthias: Yeah, definitely a blocker. I do think it should be pretty fast for me do track down
<matthias>	so nothing to worry about, a just-wait-for-owen bug...
<matthias>	kris: did you have time to check out the treemodelsort testcase which I attached to bug 300089 ?
<tml>		owen: about pango, what do you think about my patch to #162681
<kris>		matthias: i was amazed how simple the testcase was (:
<kris>		matthias: i want to debug it tonight
<kris>		like in 30 mins or so
<matthias>	kris: cool, thanks
<owen>		The description of the patch looked reasonable
<matthias>	kris: the testcase was constructed by sprinkling gtk_tree_model_sort_row_changed() with printfs
<matthias>	kris: and then tweaking the testcase until I get to the critical spot in that function...
<kris>		ok
<kris>		i hope to get a patch in soon though, it will need a decent amount of testing
<owen>		tml: though the fact that it doesn't work for Luxi Sans,Arial or whatever is probably not right
<matthias>	I want to do another 2.7.x release this Friday, with the plug/socket patch. If you have a fix by then, that would be good to have enough testing
<kris>		yep
<owen>		tml: There is a certain worry about putting complicated stuff in that code path, but we probably should be caching fontsets on win32 anyways
<tml>		owen: caching in what sense?
<owen>		tml: Keep track of some number of fontsets and when requested again, just return them
<tml>		ok
<matthias>	tml: I stumbled over bug 173098 today, and wondered if it can be right that gutils.h contains a non-inlined function body on win32 ?
<tml>		matthias: yeah, it can be made inline
<matthias>	ah, I see that you commented on it...
<tml>		it can well be inside #ifdef G_CAN_INLINE even
<tml>		owen: what about that pango_layout_line_get_empty_extents() should use fontset and not font bug i entered some days ago?
<tml>		#310700, even
<owen>		tml: I have to think about that ... using a font is definitely intentional there... but the right thing to do may be to use the first found font in a fontset
<owen>		tml: The general approach in fact for load_font() for fontconfig is to do the work for load_fontset() and then just use the first found font.
<tml>		owen: but then it is definitely called with a layout with a fontdesc for "sans" ?
<tml>		oh ok. should pangowin32 do the same?
<owen>		tml: Yes, but I'm not sure that load_font("sans") should fail ... I think it should probably give you arial (say)
<tml>		should that happen (ideally) even without pango.aliases? i.e., is it wrong to depend on pango.aliases having those sans, serif and monispace lists?
<owen>		tml: There's that open bug about trying to derive those from various aspects of system config, on some systems
<owen>		tml: Building the lists into Pango, or building fallbacks into cairo probably would make the install more robust...
<matthias>	like the good old "fixed" font...
<tml>		the problem with pangowin32 is that it was largely copied from pangox without really understanding how it all hangs together ;-)
<owen>		tml: Well, also mostly copied *before* things hung together :-)
<owen>		tml: (Even now, issues like caching are still giving me some headaches)
 *	matthias detects the need for a "pango secrets unveiled" book
<matthias>	would fit nicely next to "GTK+ internals for dummies"...
<matthias>	ok, mathrick brought up the topic of the next release cycle...
<matthias>	is it too early to discuss this, or should we brainstorm a little proto-2.10-plan ?
<matthias>	seems to early...
<kris>		i think it's a bit too early ...
<kris>		i do have some little plans
<jrb>		seems a bit early
<matthias>	ok then, I don't have further items to discuss. I plan to do another round of 2.7.x releases on Friday,
<matthias>	so get your fixes in
<kris>		matthias: could you look at my really tiny patch in #309211?
<kris>		matthias: so i can commit it
<matthias>	sure
<jrb>		as a plug, people should look at gazpacho
<matthias>	jrb: is in extras ?
<jrb>		it's pretty nice right now -- feels like it's gotten further than glade already
<jrb>		no.  we should package it
<jrb>		matthias: it runs 'out of the box' though
<matthias>	jrb: gnome cvs ?
<jrb>		no
<jrb>		svn.sicem.biz
 *	matthias watches "yum install" install gazpacho...
<jrb>		 svn co svn://www.sicem.biz/gazpacho 
<matthias>	jrb: 0.5.3 is in extras, evidently
<jrb>		matthias: oh, interesting
<kris>		kris@nasibal:~$ apt-cache search gazpacho
<kris>		gazpacho - GTK+ User Interface Designer
<kris>		it's also in debian sarge
<jrb>		aha.  I removed extras from my yum repos
 *	matthias hates logos that make him suddenly feel hungry
<zuh>		 The (currently) released version doesn't really reflect the quality of what's in svn, so you'll probably be no wiser from packaged versions...
<matthias>	thanks for the hint
<jrb>	(there are 1000 lines of changelog after the last release)
<jdahlin>	yes, you should really try svn
<matthias>	kris: that patch looks ok to me
<jdahlin>	we're getting fairly close to a new stable release
<kris>		matthias: will commit then
<kris>		matthias: thanks!
<matthias>	lets declare the meeting closed, see you next week
<tml>		bye everyone


Meeting ended July 19, 16:56 EDT