[15:54:40]
***
kaffee has joined chat #gtk-devel.
[15:55:57]
***
jrb has joined chat #gtk-devel.
[15:56:09]
***
owen has joined chat #gtk-devel.
[15:57:17]
<matthias>
is it time already ?
[15:57:23]
<jrb>
almost
[15:57:53]
***
tml has joined chat #gtk-devel.
[15:58:08]
***
owen has left chat #gtk-devel (Leaving).
[15:58:25]
***
owen has joined chat #gtk-devel.
[16:00:28]
***
charlesr has joined chat #gtk-devel.
[16:04:23]
<matthias>
ok, should we start ?
[16:04:55]
***
ebassi|dinner is now known as ebassi.
[16:05:27]
<owen>
Sure
[16:05:53]
<matthias>
to continue with last weeks action items, I actually looked at the different proposals for a druid widgets
[16:06:41]
<matthias>
but I did not get around to writing a review mail to gtk-devel-list yet. The feedback is in the bug though
[16:06:58]
<matthias>
and Carlos said that he want to work on incorporating it
[16:08:03]
<owen>
It woudl be good ot write it up for the list, certainly
[16:08:13]
<matthias^
I still want to do that
[16:08:34]
<jrb>
hrm. no alex
[16:09:14]
<matthias^
do you know how far along his canvas plans are ?
[16:10:00]
<jrb^
no. He is only in the exploratory stage.
[16:12:35]
<matthias>
owen: Hans pointed out that i probably misnamed the two new c files for the tray icon.
[16:13:04]
<matthias>
gtktrayicon.c should probably be gtktrayicon-x11.c, and gtkstatusicon-x11.c should be gtkstatusicon.c
[16:13:42]
<matthias>
can you move them on the server ?
[16:15:04]
<owen^
Do you want them copied or just moved
[16:15:44]
***
jean has joined chat #gtk-devel.
[16:15:54]
<matthias>
moving should be fine
[16:16:32]
<owen^
Done
[16:17:41]
<ebassi>
hate to reply one week late, but I know there were questions on the recent files storage, esp. regarding the libxml2 dependency
[16:18:50]
<matthias^
I want to do a thorough review of the recent-files proposal on the gtk-devel-list later this week
[16:19:10]
<matthias>
but the general feeling last week was that a libxml dependency is a nono for gtk
[16:19:48]
<ebassi+
okay. as for the dependency, I understand the issue.
[16:20:37]
<owen^
Other than the weird stuff like xincludes of http files, is there any reason we couldn't parse xbel with gmarkup?
[16:20:45]
<ebassi>
matthias: since gmarkup is too simple for the xbel format, i was wondering if expat (upon which we depend indirectely) would be okay
[16:20:47]
<owen>
(The subsset that people actually include)
[16:20:55]
<owen>
the subset that people actually use
[16:21:14]
<owen>
ebassi: We dont' depend on expat cross-platform, just on Unix
[16:21:41]
<ebassi^
I know, but gmarkup for user data is a bit too simple
[16:21:50]
<owen^
What in particular?
[16:21:57]
<matthias+
any concrete things that are missing from GMarkup ?
[16:22:38]
<ebassi>
owen: namespaces, DTDs in the passthrough section
[16:22:54]
<ebassi>
and that it seems a bit too fragile for user data
[16:23:10]
<owen^
I'm not sure what you mean by user data
[16:23:17]
<owen>
It's a pretty robust parser
[16:23:19]
<matthias>
its not really user data, its only written by tools...
[16:23:48]
<owen>
SOmeone could obviously go out of their way to break it, but it does parse a well-deifned xml subset
[16:24:08]
<owen>
ebassi: namespaces are a little bit dificult.y DTDs could be skipped pretty easily
[16:24:21]
<ebassi^
I thought the same
[16:24:50]
<owen>
does expat handle namespaces?
[16:24:58]
<ebassi^
yes
[16:25:48]
<owen>
It might be worth investigating what it would take to add that to gmarkup ... xml namespaces aren't all that complex
[16:26:15]
<owen>
I'm really don't want to start depending on expat in more of the stack, and libxml2 is huge
[16:27:11]
<jdahlin>
Xinclude?
[16:27:46]
<owen>
You could keep API compat with something like g_markup_parse_context_parse_resolve (char *tag, char **namespace, char **element);
[16:27:53]
<matthias>
jdahlin: GMarkup doesn't handle files, so that would probably a layer up...
[16:28:06]
<owen+
We are not including a http library in GLib...
[16:28:30]
<matthias>
a callback-mechanism might be possible
[16:28:43]
<owen>
jdahlin: for local files, xinclude doesn't seem that intimidating, once you have namespaces
[16:29:05]
<owen>
Of course, there is a slippery slope here...
[16:29:33]
<matthias>
dragged into the interweb of xml standards...
[16:29:36]
<ebassi>
owen: also, namespaces would mean that we could add our own meta-data to XBEL, while keeping a shared layer between us, KDE, XFCE and others
[16:29:46]
<jdahlin>
I haven't thought about it too much, I just know it would be nice to have an include feature for glade files
[16:30:06]
<jdahlin>
could be done in the parser on top of GMarkup of course.
[16:30:26]
<owen>
I don't think expat supports xinclude
[16:30:55]
<owen>
If xinclude requires strict nesting it isnt' really required either ... you can just do it at the app layer
[16:31:00]
<jrb>
jdahlin: what would we include in glade files?
[16:31:04]
***
jean has left chat #gtk-devel (Leaving).
[16:31:09]
<jdahlin>
jrb: other glade files
[16:31:24]
<jrb^
hrm. sounds mildly dubious.
[16:31:31]
* jrb
wonders how gazpacho would show that
[16:32:04]
<owen>
OK, I think we're drifting here
[16:32:08]
<matthias>
jrb: would templates be done in that way ?
[16:32:18]
<jdahlin+
well it's quite similar to overlays in feature, and menu merging in uimanager could also use it
[16:32:27]
<jdahlin>
err, the overlay feature in xul
[16:32:51]
<jrb>
matthias: that's not how I was planning on doing templates. But I'm not actually writing templates, so it's kinda up to the gazpacho guys
[16:33:54]
<jdahlin>
can the druid widget be called assistant instead of druid or wizard?
[16:34:10]
<matthias>
ok, do we have a volunteer to a) investigate what XML features are essential for reasonable xbel support, but not supported by GMarkup ?
[16:34:30]
<matthias>
jdahlin: thats what the existing patches already use
[16:34:34]
<jrb+
that's the plan, afaik
[16:34:43]
<ebassi>
matthias: okay, I'll do it
[16:34:57]
<jdahlin>
matthias, jrb: good, druid/wizard are pretty bad names
[16:35:44]
<ebassi>
matthias: if it works, we could also switch the FIleChooser bookmarks to the same format, and share them across applications and environments
[16:35:47]
<matthias>
DV will rotate if he hears we consider doing a partial implementation of xinclude over a partial implementation of namespaces over an xml subset parser...
[16:36:05]
<ebassi^
I know :-)
[16:39:31]
<matthias>
yosh: are you around ? I wondered if you had any thought on the tons of warnings we get from the atomic ops
[16:39:33]
***
pbor_ has joined chat #gtk-devel.
[16:39:49]
<matthias>
yosh: is the right fix to make the arguments volatile ?
[16:48:31]
<matthias>
owen: did you put any thought into timeframes for the the printing work ? Will it be best to wait for cairo 1.2 with non-experimental pdf and ps backends ?
[16:50:36]
<owen^
Haven't really put thought into that yet, no. I think the cairo part is the easy part though... the harder part is UI and spooling, and that could start with what is already there in cairo now
[16:51:51]
<jrb^
I randomly put in printing dialog into the hackfest at Boston
[16:52:22]
<jrb>
s/randomly/somewhat arbitrarily/ -- it might be a bad idea
[16:52:57]
<owen^
Well, sounds somewhat parallelizable. Though we'd need to get some pre-work done
[16:54:18]
<jrb>
yeah.
[16:54:39]
<jrb>
owen: seemed a better choice than, say, gnome-app
[16:55:12]
<matthias+
can you outline that pre-work ?
[16:55:28]
<owen^
Well:
[16:55:55]
<owen>
- A first stab at the API
[16:56:17]
<owen>
- Maybe an initial stupid implementation so that the framework is there so people can take and extend
[16:56:29]
<owen>
pieces of it
[16:57:49]
* jrb
assumes that requirements are first
[16:57:52]
<jrb>
oh wait, I'm wrong
[16:57:54]
<jrb>
we need a webpage first (=
[16:58:06]
<owen>
To do the first stab at the API, you need to do some thinking about use cases and API review, of course
[16:58:24]
<owen>
jrb: A logo!
[16:59:22]
***
federico has joined chat #gtk-devel.
[16:59:29]
<federico>
brilliant - did I miss the meeting?
[16:59:45]
<jrb^
it's slowing down a bit
[17:00:10]
<jrb>
we were idly discussing working on printing at the boston GNOME summit
[17:04:24]
<matthias>
ok, I'm running out of steam. Need to spend some time with the kids now, anyway.
[17:06:29]
***
jpe has left chat #gtk-devel (Ping timeout: 600 seconds).
[17:09:22]
***
kaffee has left chat #gtk-devel.
[17:14:29]
***
ijuma has joined chat #gtk-devel.
[17:16:27]
***
ijuma is now known as ijuma_home.