Time |
Nick |
Message |
15:14 |
meeting |
Meeting started Tue Jul 3 15:14:06 2012 UTC. The chair is gonzalo_. Information about MeetBot at http://wiki.debian.org/MeetBot. |
15:14 |
|
Useful Commands: #action #agreed #help #info #idea #link #topic #endmeeting |
15:14 |
gonzalo_ |
thanks |
15:14 |
|
#topic new members |
15:14 |
|
if anybody is new, this is the moment to say "hello!" |
15:15 |
|
ok.... |
15:16 |
|
#topic shell port |
15:16 |
|
manuq, want tell us how is doing? |
15:17 |
manuq |
gonzalo_: soo |
15:18 |
|
we are delaying the patch that removes hippo |
15:18 |
|
because I found a few easy to fix visual things while using collaboration, in the neighbourhood view |
15:19 |
gonzalo_ |
manuq, fixes after removing hippo? |
15:19 |
manuq |
gonzalo_: while removing hippo |
15:20 |
|
...and erikos will be out of action for a few weeks, so the review process would be less interesting without him |
15:20 |
|
as he knows better the work being done |
15:21 |
gonzalo_ |
manuq, are youworking in a branch? |
15:21 |
|
dogi has quit IRC |
15:21 |
manuq |
so, I'll keep working on a branch, doing the port to gtk3 |
15:21 |
|
gonzalo_: yes |
15:21 |
gonzalo_ |
manuq, but remove hippo is previous to port to gtk3, right? |
15:22 |
silbe |
manuq: it would be great if we could have a clean hippo-canvas base for the gtk3 port, not mixing hippo-removal and gtk3 port patches in a single branch |
15:22 |
manuq |
gonzalo_: yes |
15:22 |
|
silbe: of course |
15:22 |
gonzalo_ |
silbe, +1 |
15:23 |
manuq |
silbe: so yes, I'll do it in a separate branch |
15:24 |
gonzalo_ |
manuq, i would like to test and collaborate when I finish with Paint work (I hope this week) |
15:24 |
manuq |
gonzalo_: ok! |
15:24 |
silbe |
manuq: BTW, what would you currently expect a) when you're going to start posting hippo-removal patches, b) when you're going to start posting GTK3 port patches, c) when you might be done with most of the GTK3 port? I don't need fixed dates, just a rough estimation so I can base my own schedules on it. |
15:24 |
|
manuq: great, thanks! |
15:26 |
gonzalo_ |
silbe, the propoosal to replace the CanvasIcon is already reviewed? |
15:26 |
manuq |
silbe: so, I expect 3 weeks from now to post the hippo-removal patches (when eriko's back) |
15:27 |
|
I also expect most of the work for the GTK3 port is done at that time |
15:27 |
silbe |
gonzalo_: no, that's still on my list. |
15:27 |
manuq |
and if there are changes after the review of the hippo-removal, I can rebase my work |
15:28 |
gonzalo_ |
silbe, this change is important to have it decided |
15:28 |
silbe |
ok, so we'd review it all in one large batch it seems |
15:28 |
humitos |
silbe: I'm working a lot on porting activities to gtk3 |
15:29 |
gonzalo_ |
silbe, are you in touch with the people in ac porting the activities? |
15:29 |
silbe |
gonzalo_: who has carefully evaluated that change inside OLPC? |
15:29 |
gonzalo_ |
silbe, erikos was doing that |
15:29 |
silbe |
gonzalo_: right now not more than you are |
15:30 |
|
gonzalo_: anybody else? dsd? Has benzea looked at it? |
15:31 |
|
Cerlyn <Cerlyn!~ALEIN99-94-166-254.lightspeed.miamfl.sbcglobal.net> has joined #sugar-meeting |
15:33 |
gonzalo_ |
silbe, i don't know, manuq do you know? |
15:33 |
manuq |
we could ask benzea, he has been looking at sugar patches lately |
15:34 |
gonzalo_ |
ther are info here http://wiki.sugarlabs.org/go/F[…]ell/Hippo_Removal |
15:35 |
silbe |
gonzalo_: I don't have any objections on the approach myself, but I'm not a GTK guru. It would be nice to get some feedback from someone who knows the GTK pitfalls well to be sure we're not following an approach where we'll need to expect further trouble down the road. |
15:35 |
dnarvaez |
manuq: on which branch/repo are you doing the gtk3 port? |
15:36 |
gonzalo_ |
silbe, ok, we can try ping benzea to have more feedback |
15:36 |
silbe |
i.e. if someone experienced with designing custom GTK widgets (which this is) can't think of any problem we might run into, it's good enough for me |
15:36 |
|
we don't need that person to be sure that we will _not_ run into trouble later |
15:36 |
dnarvaez |
I had a look and the general approach seems fine to me |
15:36 |
|
(didn't review in detail) |
15:37 |
manuq |
dnarvaez: it wasn't started |
15:37 |
silbe |
dnarvaez: how much experience designing custom GTK widgets do you have? |
15:38 |
dnarvaez |
silbe maybe not as much as benzea but I have some :) |
15:39 |
silbe |
dnarvaez: ok, that's a good start. Let's try to have benzea take a look, but if he's not going to have time to do that anytime soon, we'll merge it anyway. |
15:39 |
benzea |
huh? |
15:39 |
silbe |
I'm only trying to reduce the amount of extra work we have to do if/when we realise we need to take a different approach |
15:40 |
|
benzea: ah, there you are. It would be great if you could take a look at https://patchwork.sugarlabs.org/patch/1547/ and tell us what you think about the approach it takes. |
15:40 |
benzea |
looks |
15:40 |
dnarvaez |
manuq: oh ok, maybe post it when you start :) |
15:40 |
benzea |
juts opened his irc window |
15:41 |
manuq |
silbe: the port of the CanvasIcon was done in Brno hackfest, by Walter Bender, Daniel Drake and Raul Gutierrez Segales, and many gnome devs were around |
15:41 |
benzea |
ah, yes, I wanted to look at it on sunday, and then evolution was missing the mail in the cache (and I was offline) :-/ |
15:41 |
silbe |
benzea: feel free to take your time with it, you can just reply to the email. |
15:42 |
|
manuq: ah, that's good to know indeed. |
15:42 |
manuq |
silbe: yeah, is in the commit message |
15:42 |
silbe |
manuq: it mentions previous authors, but not the hackfest and the presence of gnome devs. |
15:43 |
manuq |
silbe: right |
15:44 |
silbe |
ok, I'd suggest to get on with the meeting. We can handle the patch on sugar-devel, no need to do it right now. |
15:44 |
gonzalo_ |
silbe, yeah |
15:44 |
benzea |
yeah |
15:45 |
gonzalo_ |
benzea, thanks, it's important have it defined to continue with other tasks |
15:45 |
|
ok, anything else about sugar-port? |
15:45 |
silbe |
+1, thanks to benzea for reviewing this patch |
15:45 |
gonzalo_ |
volunteers? :) |
15:47 |
|
ok, can we continue? |
15:47 |
|
3 |
15:47 |
|
2 |
15:47 |
|
1 |
15:47 |
manuq |
1 |
15:47 |
|
:) |
15:47 |
gonzalo_ |
#topic port of activities |
15:48 |
manuq |
has seen a lot of progress! |
15:48 |
gonzalo_ |
yeah, a lot of progress, updated state is here http://wiki.laptop.org/go/Acti[…]ies/PortingToGtk3 |
15:49 |
silbe |
0 |
15:49 |
gonzalo_ |
sadly dirakx is not here |
15:49 |
manuq |
aguz send a patch for review, ImageViewer port |
15:50 |
|
who is the maintainer? |
15:50 |
benzea |
outch, just thinking about measure |
15:51 |
gonzalo_ |
manuq, i think aleksey maintain ImageViewer |
15:51 |
benzea |
there will be a huge performance hit there, I expect; unless maybe a custom C module is written currently |
15:51 |
manuq |
gonzalo_: ok, I think I can review it |
15:51 |
|
aguz said was a present for my birthday :) |
15:51 |
gonzalo_ |
benzea, due to signal display or gstreamer? |
15:52 |
dnarvaez |
benzea: where would the hit be? |
15:52 |
gonzalo_ |
manuq, :) |
15:52 |
manuq |
benzea: can you provide details? :) |
15:52 |
benzea |
gonzalo_, dnarvaez: display, I did quite some stuff there to get it fast |
15:52 |
|
it uses gdk_draw_lines so that it does not need a loop in python for drawing |
15:52 |
dnarvaez |
oh I thought you was talking about the shell port :) |
15:53 |
benzea |
nah, just the measure activity |
15:53 |
manuq |
dnarvaez: Measure Activity :) |
15:53 |
gonzalo_ |
benzea, and lines in cairo should be slower? |
15:54 |
benzea |
gonzalo_: well, for one drawing the lines with cairo is likely slower, the second thing is that we need a python loop |
15:54 |
|
ie. call cr.line_to() |
15:54 |
gonzalo_ |
yeah |
15:54 |
benzea |
well, I wanted to play with that anyways for some time now ;-) |
15:54 |
gonzalo_ |
benzea, great :) |
15:55 |
|
i am porting Paint to cairo, and wondering how will resolve the bucket functionality.... |
15:55 |
benzea |
what is it using currnetly? |
15:55 |
gonzalo_ |
c code |
15:56 |
benzea |
hm, keep it? |
15:56 |
manuq |
gonzalo_: with a python implementation fallback, right? |
15:56 |
gonzalo_ |
manuq, yes |
15:56 |
manuq |
gonzalo_: yes, why can't be kept? |
15:56 |
gonzalo_ |
benzea, but now is using a pixmap to work |
15:56 |
benzea |
gonzalo_: cairo does not provide such image manipulations routines, you need to use an ImageSurface and access its data (don't forget to call flush/mark_dirty) |
15:57 |
|
gonzalo_: uh, pixbuf you mean? |
15:57 |
gonzalo_ |
benzea, no, pixmap |
15:57 |
benzea |
a server side pixmap? how do you do a bucket fill with that? |
15:58 |
gonzalo_ |
benzea, Paint use pixmaps as canvas to all the operations |
15:58 |
benzea |
for me a pixmap is a GdkPixmap, and the data is located on the X server, no direct access for the application; don't a fill there sounds rather inefficient |
15:59 |
gonzalo_ |
benzea, really use GdxPixmap.get_image and process that |
16:00 |
benzea |
oh |
16:01 |
gonzalo_ |
benzea, but I think is because all the operations are done in this canvas pixmap (actually two pixmaps) |
16:01 |
benzea |
I would probably have done a gtk.gdk.Pixbuf.get_from_drawable to not require roundtrips to the X server |
16:01 |
|
and then upload it after the opation |
16:01 |
gonzalo_ |
benzea, ok, i will try this |
16:01 |
manuq |
benzea: interesting |
16:02 |
benzea |
gonzalo_: with cairo, create an image surface of the same size, and render the remote surface into it (using OPERATOR_SOURCE) |
16:02 |
gonzalo_ |
benzea, what is the remote surface in this case? |
16:03 |
benzea |
gonzalo_: the drawable |
16:03 |
gonzalo_ |
ahh, ok |
16:04 |
benzea |
if you want to keep using a Pixmap that is |
16:05 |
|
ok, enough discussion about Paint? |
16:06 |
|
thinks that any more discussions should probably be done outside of the meeting |
16:06 |
manuq |
ok, anything else? |
16:06 |
gonzalo_ |
benzea, Pixmap are deprecated in gtk3 http://developer.gnome.org/gtk[…]02.html#id1401440 |
16:06 |
manuq |
anyone want to raise issues on the gtk3 port of activities? |
16:07 |
|
igod has quit IRC |
16:07 |
benzea |
gonzalo_: yeah, but doing eg. a cairo.create_similar_surface() on the window is the same as a pixmap |
16:07 |
gonzalo_ |
benzea, ok, but is a surface |
16:07 |
benzea |
gonzalo_: just a different wrapper for the same thing :-) |
16:08 |
gonzalo_ |
i am removing the use of GdkPixmaps to port to gtk3 later |
16:08 |
|
and yes, i am using cairo surfaces to replace the pixmaps :) |
16:09 |
|
ok, dnarvaez are you here? |
16:09 |
dnarvaez |
yes |
16:09 |
gonzalo_ |
i have a question but is not about porting activities |
16:09 |
|
you are trying to start activities in your bot to test they are working, right? |
16:10 |
|
dnarvaez, ^ |
16:10 |
dnarvaez |
gonzalo_: right, I'm using dogtail to start them and then close, one by one |
16:10 |
gonzalo_ |
dnarvaez, starting from listview? |
16:10 |
dnarvaez |
yes |
16:10 |
gonzalo_ |
wow :) |
16:10 |
|
danielf_ <danielf_!~danielf@r186-49-32-236.dialup.adsl.anteldata.net.uy> has joined #sugar-meeting |
16:11 |
dnarvaez |
:) |
16:11 |
manuq |
dnarvaez: they can be also started using a command |
16:11 |
gonzalo_ |
i read your mail about problems and I was wondering if you can use sugar-activity to start them |
16:12 |
|
dnarvaez, but i don't know id is useful for your tests |
16:12 |
dnarvaez |
I could use the script but then I would still need a way to make sure it started properly and close it. More in general I;m trying to get dogtail to work because I think we can use it to do more advanced ui testing |
16:13 |
manuq |
dnarvaez: that would be very nice |
16:13 |
dnarvaez |
it would be nice if the activities itself would include ui testing scripts |
16:13 |
gonzalo_ |
dnarvaez, yeah, or have a repo with the tests |
16:13 |
|
danielf_ has quit IRC |
16:14 |
Cerlyn |
dnarvaez: Is there any particular reason you chose dogtail? |
16:14 |
gonzalo_ |
dnarvaez, when you have this working send a mail to sugar-devel |
16:14 |
dnarvaez |
gonzalo_: will do :) |
16:15 |
gonzalo_ |
dnarvaez, thanks by working on that |
16:16 |
dnarvaez |
Cerlyn: not quite, it was easy to get to work on fedora. I don't care much about dogtail, more about the accessibility approach. dogtail is really just very simple convenience on the top of pyatspi |
16:17 |
gonzalo_ |
dnarvaez, can you access the toolbar controls with dogtail and identify the widgets? |
16:17 |
dnarvaez |
gonzalo_: yes, my script uses the stop button on the toolbar |
16:18 |
|
http://buildbot.sugarlabs.org/[…]hell_3/logs/stdio |
16:18 |
gonzalo_ |
dnarvaez, great, open many posibilities |
16:18 |
dnarvaez |
gonzalo_: you can see what the tree looks like here ^ |
16:19 |
|
that's list view |
16:20 |
gonzalo_ |
dnarvaez, crazy stuff :) |
16:20 |
|
but interesting |
16:20 |
Cerlyn |
dogtail also can be hooked into the i18n files |
16:21 |
gonzalo_ |
Cerlyn, ? |
16:21 |
Cerlyn |
So it can know "Browse" and "Navagar" are the same if you set LANG |
16:22 |
|
the restriction on what it sees being it has to get back to AT-SPI somehow, so hippocanvas items are not as easily automatable |
16:22 |
gonzalo_ |
Cerlyn, but hippo is going away |
16:23 |
|
ok, we promissed keep this meeting short :) |
16:23 |
|
anybody have anything more to share/ask? |
16:27 |
benzea |
doesn't look like it |
16:27 |
gonzalo_ |
ok :) |
16:27 |
|
5 |
16:27 |
|
..... |
16:27 |
|
4 |
16:28 |
|
3 |
16:28 |
benzea |
2 |
16:28 |
gonzalo_ |
1 |
16:28 |
manuq |
1 |
16:28 |
gonzalo_ |
#endmeeting |
16:28 |
meeting |
Meeting ended Tue Jul 3 16:28:45 2012 UTC. Information about MeetBot at http://wiki.debian.org/MeetBot. (v 0.1.4) |
16:28 |
|
Minutes: http://meeting.sugarlabs.org/s[…]-03T15:14:06.html |
16:28 |
|
Log: http://meeting.sugarlabs.org/s[…]12-07-03T15:14:06 |