Attachment '20110804_log.txt'
Download 1 14:36:07 <API> #startmeeting
2 14:36:07 <tota11y> Meeting started Thu Aug 4 14:36:07 2011 UTC. The chair is API. Information about MeetBot at http://wiki.debian.org/MeetBot.
3 14:36:07 <tota11y> Useful Commands: #action #agreed #help #info #idea #link #topic.
4 14:36:30 <API> #topic possibly add a children-invalidated even
5 14:36:39 <API> well, this is somewhat a technical topic
6 14:36:54 <API> mgorse, ?
7 14:36:58 <API> you added that point
8 14:37:11 <mgorse> yeah, I added it in response to an email that fregl sent out
9 14:37:23 <mgorse> since, for kde, when removing a child, it no longer has access to the child's index
10 14:37:46 <mgorse> so he suggested having an event that would just say that children are invalidated
11 14:37:53 <fregl> hi, great
12 14:37:57 <API> insted of children-remove?
13 14:38:05 <fregl> I was just wondering how to tackle this
14 14:38:12 <mgorse> but this would mean that ATs might not see children-changed::remove events, for instance, so need to discuss it
15 14:38:22 <fregl> my problem is mostly that passing the child index along makes things hard
16 14:38:25 <API> well, fwiw some weeks ago we talked with Company, and he is not sure about the index stuff
17 14:38:34 <API> as some widgets don't have inherently a order
18 14:38:38 <mgorse> another option might be for the app-side AT-SPI implementor to send -1 for the detail, and then libatspi would fill in the index if it has the accessible cached
19 14:38:51 <mgorse> but, at least for MANAGES_DESCENDANTS objects, it won't have the child cached
20 14:38:59 <fregl> mgorse: that would work for me
21 14:39:20 <joanie> so would these be MANAGES_DESCENDANTS objects fregl?
22 14:39:23 <fregl> for manages descendants I would think we shouldn't send child removed/added
23 14:39:28 <fregl> joanie: no
24 14:39:47 <fregl> it is a general thing in how object deletion works in the case of Qt in many cases
25 14:40:00 <API> fregl, so that -1 is an option for you?
26 14:40:13 <API> I mean, send -1 on the index if it is not possible to send the proper one
27 14:40:17 <fregl> the parent get's it's pointer to the child removed and the child still knows the parent when I get the notification
28 14:40:26 <fregl> API: yes, the -1 as detail would be perfect
29 14:40:43 <fregl> and when it's not cached, all should be fine anyway as far as I understand
30 14:40:44 <API> ok, so I guess that we can assume that in general:
31 14:40:46 <joanie> and having at-spi deal with it would be perfect
32 14:40:48 <API> and update the documentation:
33 14:40:50 <API> http://developer.gnome.org/atk/unstable/AtkObject.html#AtkObject-children-changed
34 14:41:06 <API> adding that -1 will be sent if it is not possible to know the index
35 14:41:19 <API> mgorse, do you agree?
36 14:41:22 <fregl> sweet
37 14:41:22 <API> hmm, stupid question
38 14:41:27 <API> you proposed that ;)
39 14:41:27 <mgorse> Having AT-SPI fill in the index might not always work or be easy, though
40 14:41:40 <fregl> well, what is the index used for?
41 14:41:50 <mgorse> ie, if a children-changed:add event was send for the object, would it include the object path of the added child?
42 14:41:53 <joanie> mgorse: we can punt in the worst case
43 14:42:19 <joanie> fregl: depends on the app
44 14:42:24 <mgorse> if not, then the child won't be cached, unless AT-SPI were to ask for the child, which I'm not sure is what we want, either
45 14:42:28 <joanie> often we don't care
46 14:43:14 <joanie> in the case of impress I think the index is going to prove handy in terms of being able to communicate slides and placeholders that were just removed
47 14:43:22 <fregl> I thought the whole purpose of this function was to just let the app know that the thing is no longer available
48 14:43:33 <mgorse> I guess we could always do that for now, have KDE send -1 for the detail, and AT-SPI would fill in the index if it has it cached, and deal with it later if it winds up causing problems for someone
49 14:43:38 <joanie> but what thing fregl?
50 14:43:51 <joanie> without the index we know that one of the children of the object went away
51 14:44:08 <joanie> if we care which one (i.e. to say 'slide 5 deleted')
52 14:44:09 <API> mgorse, as I said Company is not also sure about that index
53 14:44:10 <fregl> joanie: no, you also get the path of the removed child
54 14:44:11 <joanie> we don't know
55 14:44:24 <joanie> the path?
56 14:44:29 <fregl> joanie: yes, the path
57 14:44:30 <joanie> but not the index?
58 14:44:31 <API> although I guess that in the same way that we can document that -1 should be used if not index is available
59 14:44:34 <fregl> the dbus path
60 14:44:44 <joanie> who is we?
61 14:44:45 <API> we should also said that this would be the lesser evil
62 14:44:49 <joanie> I'm an AT
63 14:44:59 <fregl> joanie: at-spi dbus
64 14:45:03 <fregl> so the other side
65 14:45:14 <mgorse> fregl: When a child is added, do you send the path? I remember you talking about not wanting to instantiate objects unless they're needed, and getting a path requires instantiating an object
66 14:45:43 <fregl> mgorse: actually we worked around that by patching the qt dbus implementation for this
67 14:46:16 <fregl> I think it is very silly anyway - right now I send: parentpath children-changed:add index 0 childpath
68 14:46:24 <mgorse> oh okay. If you send the object path and a children-changed event when an object is added, then it should be cached, unless it has MANAGES_DESCENDANTS
69 14:46:24 <fregl> so you get redundant information
70 14:46:52 <fregl> ok
71 14:47:03 <fregl> I really need to read the atspi code more
72 14:47:10 <fregl> but all seems fine now
73 14:47:41 <fregl> right now I send -1 as index for object removed already and that probably doesn't work
74 14:47:54 <mgorse> Anyway, I guess we can conclude this / add a #info
75 14:48:00 <fregl> great
76 14:48:46 <mgorse> #info If an app sends -1 for the detail for a children-changed, we will assume that the index is unknown
77 14:48:49 <API> yes please
78 14:49:01 <fregl> my current state is that I want to release a beta and get people to test for real now. I got startup time and memory usage to almost nothing.
79 14:49:16 <mgorse> #info mgorse will modify libatspi to fill in the index when possible (ie, when it has the child cached, which it will not for trees with MANAGES_DESCENDANTS for instance)
80 14:49:29 <fregl> mgorse: you rock!
81 14:49:31 <joanie> (woo hoo fregl! Congrats!!!)
82 14:49:44 <API> #info AtkObject:children-changed documentation will be updated, as this could be also used on ATK implementors
83 14:49:54 <API> well, anything else?
84 14:49:55 <joanie> by whom API?
85 14:49:59 <API> I can do that
86 14:50:02 <fregl> tables/trees with millions of lines cause no delay any more, that was what worried me :D
87 14:50:14 <API> well in fact it is more an action item
88 14:50:19 <API> but that just a nitpick
89 14:50:21 <joanie> #action Piñeiro will update the documentation
90 14:50:21 <mgorse> Oh; that's an improvement over gtk :)
91 14:50:36 <joanie> API it's a nitpick to you who doesn't deal with minutes :-P
92 14:50:42 <API> anything else in this point?
93 14:52:05 <API> ... silence ....
94 14:52:08 <API> ok, moving
95 14:52:16 <API> #topic CSUN Speaker Proposals
96 14:52:24 <joanie> API no
97 14:52:26 <API> joanie, you added this poing, AFAIK
98 14:52:27 <API> no?
99 14:52:35 <joanie> clown: added stuff
100 14:52:38 <API> hmm, last minute changes ;)
101 14:52:41 <joanie> I mean we could do this out of order
102 14:52:45 <joanie> but...
103 14:52:49 <API> #topic quick update on gs-mag
104 14:52:52 <clown> sure. I think I put my addition in the wrong place on the agenda, as it relates to 3.2 progress. If you want to postpone until we deal with 3.2 progess..
105 14:53:05 <joanie> clown: please go on
106 14:53:08 <API> no problem, I just have that page open
107 14:53:15 <API> we can go in order
108 14:53:17 <API> clown please
109 14:53:23 <clown> okay. it's fairly quick anyhow. I'll just add some #infos.
110 14:53:33 <clown> #info Joseph submitted a new patch for the zoom options dialog in universal access.
111 14:53:40 <clown> #info https://bugzilla.gnome.org/show_bug.cgi?id=643086#c27
112 14:53:40 <tota11y> 04Bug 643086: normal, Normal, ---, control-center-maint, NEW, Universal access: add dialog for zoom options
113 14:53:49 <clown> #info awaiting review.
114 14:53:53 <clown> and...
115 14:53:58 <clown> #info Re: Joseph's invert video patch for gs-mag
116 14:54:04 <clown> #info https://bugzilla.gnome.org/show_bug.cgi?id=639851#c8
117 14:54:04 <tota11y> 04Bug 639851: normal, Normal, ---, gnome-shell-maint, UNCONFIRMED, Magnifier: Add brightness and contrast functionality.
118 14:54:09 <clown> #info suggestion is to move it to Clutter.
119 14:54:40 <clown> that last one is a bit of a surprise -- it would be cool if anyone who used Clutter had access to it.
120 14:54:47 <API> brightness and contrast funcionality on clutter?
121 14:55:02 <clown> yes, and lightness inversion.
122 14:55:05 <API> something like a property on the actor or the stage?
123 14:55:14 <clown> yes, exactly.
124 14:55:25 <clown> there already is, in clutter, various tint and other effects.
125 14:55:38 <clown> and an abstract class for any effect.
126 14:55:49 <API> well, the good thing is that would allow to access to those effects to more apps
127 14:55:55 <API> more easily I mean
128 14:55:58 <clown> I just implemented a specific object for the effects needed for the magnifier.
129 14:56:05 <clown> yes, exactly API.
130 14:56:38 <clown> #action Joseph to reply on bug to say, "yes, let's move these effects to clutter".
131 14:56:42 <clown> done.
132 14:57:08 <API> ok, more questions, comments, doubts in this point?
133 14:58:30 <API> #topic CSUN Speaker Proposals
134 14:58:31 <API> joanie, ?
135 14:58:37 <joanie> Thanks API
136 14:58:44 <joanie> I'll #info to be efficient.
137 14:59:14 <joanie> #info While we are still waiting for the budget to be approved, the treasurer and Board have not yet told us we are insane.
138 14:59:34 <heidi> Just to clarify, are we talking about the conference end of Feb to March in San Diego
139 14:59:35 <heidi> ?
140 14:59:39 <joanie> #info We anticipate being able to send two speakers to CSUN, the key being "speakers"
141 14:59:41 <joanie> heidi yes
142 14:59:49 <heidi> THanks
143 15:00:06 <API> heidi, he have been talking about that conference last meeting, we decided to not use the budget for a booth
144 15:00:07 <joanie> #info CSUN call for papers typically opens in September. It's now August.
145 15:00:17 <API> but it would be still interesting to send peoople to talk about GNOME
146 15:00:30 <joanie> #info Now might be the time for anyone who would make a good speaker to consider speaking and thinking about topics.
147 15:00:33 <joanie> (done)
148 15:01:37 <API> I will use the "lets use native english speakers" poker ;)
149 15:01:46 <joanie> hahaha
150 15:02:01 <joanie> I will suggest that is nonsense.
151 15:02:16 * clown wonders how many native english speakers are present.
152 15:02:24 <joanie> exactly
153 15:02:30 <jjmarin> lalala
154 15:02:40 <joanie> jjmarin: ;-)
155 15:02:42 * clown there's a native english singer present.
156 15:02:54 <prlw1> I'm a tenor :-)
157 15:03:02 <API> well, it seems that there aren't too many people there
158 15:03:10 <API> anyway, we don't have all the people here
159 15:03:13 * heidi looked at this conference last year and it is in direct conflict with the main CS educators conference and can't go
160 15:03:17 <API> for example, eitan usually goes there
161 15:03:24 <clown> well, it's something for me to contemplate while on vacation next week.
162 15:03:33 <joanie> clown: +1
163 15:03:36 <API> so how about send a mail to gnome-list abou this?
164 15:04:02 <API> and that clown comment is also a +1 to my proposal of use the ml
165 15:04:11 <joanie> #action Joanie will send a mail to gnome-accessibility-list suggesting people be thinking about presenting at CSUN.
166 15:05:03 <joanie> that ai is brought to you by the suffix -ing
167 15:05:47 <prlw1> I could redo my fosdem talk on dasher/opengazer/nomon/ticker ?
168 15:06:10 <joanie> prlw1: I've not seen the talk, but potentially, yes.
169 15:06:25 <joanie> like I said, call for papers is not for a month
170 15:06:31 <joanie> but time travels quickly
171 15:06:43 <joanie> so I'm asking people to start pondering the possibilities now
172 15:07:58 <clown> joanie, so the steps are (1) propose a paper, (2) get confirmed as a speaker, (3) apply to gnome-foundation?
173 15:08:10 <joanie> yup
174 15:08:26 <clown> any chance gf will say "no" at step 3?
175 15:08:41 <clown> or, you could be a confirmed speaker, but not receive funding.
176 15:08:43 <joanie> assuming our budget gets approved, which I think it will
177 15:08:52 <clown> ok
178 15:08:53 <joanie> we have budgeted two speakers
179 15:08:56 <joanie> from our team
180 15:09:07 <joanie> including the possibility they might need to travel from Europe
181 15:09:42 <mgorse> Are we budgeting for people to attend the whole conference? Or just come one night in order to give a presentation?
182 15:09:46 <joanie> and I cannot imagine our team going crazy and submitting tons of papers that get accepted.
183 15:10:10 <clown> i can imagine it. but I doubt it will happen that way :-)
184 15:10:10 <joanie> mgorse: more than a night, less than the full conference
185 15:10:20 <joanie> hah
186 15:10:59 <joanie> for fy2013 we budgeted much more for CSUN
187 15:11:25 <joanie> in the hopes that we'll have a real "story to tell" then
188 15:11:58 * clown once upon a time ...
189 15:12:07 <joanie> in a galaxy far, far away
190 15:12:17 <joanie> API the natives are getting restless
191 15:12:20 <joanie> I think we can move on
192 15:12:28 <API> yes, we don't have much time
193 15:12:31 <heidi> On a related note, does anyone know the dates for the GNOME Summit in Boston this fall?
194 15:12:38 <API> do you want some #infos for the last chitchatting?
195 15:12:42 <joanie> heidi: almost certainly columbus day weekend
196 15:12:48 <joanie> API not really
197 15:12:49 <joanie> lol
198 15:12:56 <heidi> joanie: Thanks, last year it wasn't for some reason...
199 15:12:58 <API> ok
200 15:13:06 <joanie> piggybacking on another event
201 15:13:15 <API> #topic Testing Distro
202 15:13:21 <mgorse> heidi: Linux Plombers was in Boston last year, and it looks like it moves around, so it won't be there this year
203 15:13:26 <API> is there something to talk about it?
204 15:13:28 <joanie> Yeah, I added that one
205 15:13:35 <heidi> joanie: Thanks
206 15:13:41 <joanie> I'm going to be my "direct" self I'm afraid :-/
207 15:13:48 <joanie> so.... testing distro was super useful
208 15:13:52 <joanie> and all kinds of awesome
209 15:14:06 <joanie> but I think most of us (all of us but clown???) are no longer using it
210 15:14:22 <joanie> and Javi is having difficulties spinning the "last" one
211 15:14:38 <joanie> so I'm wondering if it is time to move on w.r.t. the testing distro
212 15:14:48 <clown> yes, I am waiting for the latest spin -- "we" promised it to AEGIS europe for their user testing.
213 15:14:48 <joanie> clown? others?
214 15:15:10 <clown> alas, we need it.
215 15:15:14 <joanie> aleiva: jjmarin: is jhernandez in shouting distance?
216 15:15:37 <clown> is korn here?
217 15:15:50 <API> what kind of problems with the testing distro is jhfernandez having?
218 15:16:08 <clown> he was also very interested in having it as a basis for some other a11y stuff.
219 15:16:23 <joanie> clown: Is he willing to pay Emergya to do this work?
220 15:16:24 <jjmarin> joanie: nop. he's a at home. He's a little bit sick I'm afraid
221 15:16:54 <joanie> clown: because it seems to me that if the "team" who needs it is really AEGIS because Peter wants it
222 15:17:06 <joanie> and Javi is crazy busy (and sick! :-( )
223 15:17:11 <joanie> and having build problems
224 15:17:22 <joanie> clown: aleiva you see the above?
225 15:18:20 <aleiva> �sorry, rebooting
226 15:18:35 <joanie> API getting back to your question, it's come up in the minutes of previous meetings. Some of it is, I believe, packaging issues. Others is that the build service goes down a lot.
227 15:18:46 <joanie> aleiva: meaning you are back now?
228 15:18:53 <joanie> or going away?
229 15:18:56 <aleiva> just back now
230 15:20:03 <joanie> #info Given the amount of trouble Javi has been having with the testing distro "final spin," she questioned who all is using it.
231 15:20:29 <joanie> #info Joseph indicated that it is required, "promised to AEGIS europe for their user testing."
232 15:20:55 <aleiva> well, we're volunteers
233 15:21:05 <aleiva> so, when we "promise" something exactly?
234 15:21:15 <joanie> #info Joanie thinks that AEGIS user testing, while valuable, goes well beyond the scope of GNOME Accessibility Team needs and what the volunteers (i.e. Javi) can be asked to do.
235 15:21:19 <joanie> aleiva: patience
236 15:21:42 <aleiva> the mother of science
237 15:21:43 <aleiva> :-)
238 15:22:11 <jjmarin> :)
239 15:22:20 <API> well, anything else in this point?
240 15:22:29 <joanie> ugh
241 15:22:29 <aleiva> anywy, we're open to any talk about this
242 15:22:31 <aleiva> API: nop
243 15:22:35 <joanie> everyone just wait ffs
244 15:22:36 <joanie> ;-)
245 15:22:44 * API stops to poke people
246 15:23:06 <joanie> #info Joanie would encourage the AEGIS testing folks to find some other means for achieving their needs/goals in this area.
247 15:23:16 <joanie> done.
248 15:23:18 <joanie> sorry
249 15:23:27 <joanie> language is such a delicate thing
250 15:23:29 <joanie> ;-)
251 15:25:57 * API API start to poke people again
252 15:26:01 <API> ok, anything else?
253 15:26:25 <aleiva> nope
254 15:26:28 <jjmarin> nop
255 15:26:54 <API> ok,
256 15:27:00 <API> #topic miscellaneous time
257 15:27:14 <API> so, someone has something to say
258 15:27:17 <API> short
259 15:27:20 <API> not scheduled?
260 15:27:35 <aleiva> see you guys at berlin :-)
261 15:28:31 <fer> see you!
262 15:28:54 <jjmarin> \o/
263 15:30:05 * heidi waves
264 15:30:23 <API> well as is seems that nobody needs to add anything
265 15:30:31 <API> I will close the meeting
266 15:30:35 <API> #endmeeting
Attached Files
To refer to attachments on a page, use attachment:filename, as shown below in the list of files. Do NOT use the URL of the [get] link, since this is subject to change and can break easily.You are not allowed to attach a file to this page.