Attachment '20130516_log.txt'
Download 1 16:05:26 <clown> #startmeeting
2 16:05:26 <tota11y> Meeting started Thu May 16 16:05:26 2013 CET. The chair is clown. Information about MeetBot at http://wiki.debian.org/MeetBot.
3 16:05:26 <tota11y> Useful Commands: #action #agreed #help #info #idea #link #topic.
4 16:05:31 * clown yay!
5 16:05:51 <clown> #info agenda: https://live.gnome.org/Accessibility/Meetings#Agenda_for_the_Next_Meeting_.289_May.29
6 16:06:15 <clown> #topic Our GSoC/OPW candidate
7 16:06:19 <jjmarin> ok, you want to be API :-)
8 16:06:39 <clown> sure. oh, the power!
9 16:06:53 <jjmarin> I was helping magpie with jhbuild
10 16:07:13 <jjmarin> #info Juanjo has been helping magpie with jhbuild
11 16:07:33 <clown> and how did that go?
12 16:07:55 <jjmarin> #info I recommend to use a recent release and she is going to switch to Fedora 19 from Ubuntu 12.10
13 16:08:16 <clown> good idea about using Fedora. How stable is 19?
14 16:08:31 <jjmarin> quite stable more than less :-)
15 16:08:42 <jjmarin> it is in beta phase
16 16:08:47 <clown> not bad.
17 16:09:29 <clown> is magpie following the directions for building gnome-shell using jhbuild?
18 16:09:35 * clown looks those up.
19 16:09:54 <jjmarin> #info Juanjo encouraged magpie to magFocusTracking try the example from pyatspi2
20 16:10:05 <clown> https://live.gnome.org/GnomeShell#Building
21 16:10:32 <jjmarin> no yet, my idea was starting with sth simple, like jhbuilding pyatspi2
22 16:10:43 <clown> okay.
23 16:10:54 <jjmarin> but, yes, that the next step
24 16:11:36 <clown> when I built gnome-shell, it sometimes took days. because back then it build practically all of gnome.
25 16:12:21 <jjmarin> clown, yes, it can be tricky
26 16:12:23 <clown> is magpie using a real machine, or running F19 in VirtualBox
27 16:12:27 <clown> ?
28 16:12:38 <clown> I would recommend a real machine.
29 16:12:45 <jjmarin> My understanfing is a real machine
30 16:13:14 <clown> okay. regarding magFocusTracker.py...
31 16:13:44 <clown> #info Joseph met with magpie in a11y where she had a number of questions about magFocusTracker.py
32 16:14:05 <clown> #info Joseph explained that it only works with python 2.x
33 16:14:29 <clown> #info magpie's big question was why Registry.start() was called at the end of the main() function
34 16:14:54 <jjmarin> #info magpie was worried about documentation and she has recollected a collection of links http://thismagpie.blogspot.com.es/2013/05/useful-gnome-platform-documentation.html
35 16:14:57 <clown> #info Joseph explained that that is used to start the main loop, but wasn't clear on exactly what that meant.
36 16:15:45 <clown> #info magpie was going to look at magFocusTracker.py and compare it with the JavaScript version that Joseph uploaded as an attachment to bugzilla 647074.
37 16:15:47 <jjmarin> that the reason she has "Python Main Loop" in the collection of links :-)
38 16:16:07 <clown> #info magpie is going to try to figure out the logic of both those pieces of code.
39 16:16:19 <clown> jjmarin, catching up to what you just typed...
40 16:16:52 * clown loading...
41 16:17:24 <clown> okay, it's not the "Python Main Loop". it's the normal main loop that every gnome application has.
42 16:17:35 * clown still loading...
43 16:18:39 <clown> oh! there is a document name "main-loop-and-events". cool.
44 16:19:26 <clown> i wonder if this would be useful for those links: https://live.gnome.org/GnomeShell/Magnification
45 16:19:40 <jjmarin> yes
46 16:20:11 <jjmarin> I think I show her this url when she talk to me about Roihandler
47 16:20:39 <clown> Yes, the RoiHandler made extensive use of the D-Bus link to the magnifier.
48 16:21:09 <jjmarin> clown, where is the documentation of pyatspi2 ?
49 16:21:13 <clown> But, there is a danger in concentrating on that D-Bus connection. It's irrelevant when running inside gnome-shell
50 16:21:34 <clown> jjmarin: I don't know. mgorse might know.
51 16:21:57 <clown> it would have been handy when I was writing magFocusTracker.py.
52 16:22:23 <jjmarin> I've only found some old doc http://people.gnome.org/~parente/pyatspi/doc/
53 16:22:53 <clown> BTW, a few weeks back when we started this GSoC with magpie, I suggested one of her first tasks was to make magFocusTracker.py work with python3. joanie was against that.
54 16:24:17 <clown> Here is what I said back then: "Here's another thought: Fix the python based magnifier focus tracker in the examples folder."
55 16:24:43 <clown> joanie replied: "You mean the one you and I couldn't get working because of the dbus issues? That to me seems like the problematic task. This is why I deliberately failed to suggest she try that."
56 16:25:17 <jjmarin> makes sense if it is a hard task
57 16:25:56 <jjmarin> is pyatspi2 are the bindings for atpi2, where are the bindinds for js ?
58 16:26:03 <clown> well, it's distracting because of the d-bus communication between python and gs-mag. That won't be used at all in the end.
59 16:26:48 <clown> you just asked a question that I was confronted with about a year ago...
60 16:27:10 <clown> okay, the answer is not simple, since there is no equivalent to a library like pyatspi2 for js.
61 16:27:27 <jjmarin> OMG !
62 16:27:36 <clown> All there is are the introspection bindings. Let me find a link.
63 16:28:11 <clown> first, go here: https://bugzilla.gnome.org/review?bug=647074&attachment=219583
64 16:28:29 <jjmarin> I have the impression that the documentation for pyatspi2 and javascript is the plain C documentation https://developer.gnome.org/libatspi/unstable/
65 16:29:10 <clown> then click the "focusTracker.js" link near the top of the page.
66 16:30:19 <clown> Yeah, I used the code in pyatspi2 to get an idea of how to call things, and the "gi" files to determine how to write the calls in js.
67 16:30:38 <jjmarin> imports.gi.Atspi;
68 16:31:03 <clown> right. that imports objects and methods that allow one to call into atspi from js.
69 16:31:33 <jjmarin> So you can to translate C documentation to JS way
70 16:31:42 <jjmarin> s/can/has
71 16:31:48 <clown> the "gi" file lists the objects, calls, and their parameters. It does some of that translation for you.
72 16:32:33 <jjmarin> clown: how it is that ?
73 16:33:15 <jjmarin> and how is done the comunication with the magnifier in gnome shell ?
74 16:33:20 <clown> yeah, I'm trying to remember where those files are. I'm not at work, and don't have access to my Fedora box.
75 16:33:27 <jjmarin> np
76 16:33:42 <jjmarin> you're in hollidays :)
77 16:33:55 <clown> the js code in that attachment is part of the gnome-shell project — it's running inside gnome shell.
78 16:35:00 <clown> my approach was to have the "focusTracker" object emit JS signals — you'll see an 'const Signals = imports.lang.signals;' near the top.
79 16:35:46 <jjmarin> yes, I realised of that, though I didn't know what signals it is about
80 16:35:50 <clown> then any other object inside gnome shell could connect a signal handler to the focusTracker to be notified of that signal, and respond as they saw fit.
81 16:36:46 <clown> at a very high level, the JS focus tracker listens for atpsi events and handle them by emitting JS events.
82 16:37:14 <clown> that is, no d-bus involved at all...
83 16:37:30 <jjmarin> so they are glib signals or something different from js?
84 16:39:05 <jjmarin> I mean sth like https://developer.gnome.org/gobject/unstable/gobject-Signals.html
85 16:39:15 <clown> the events that the focus tracker emits are not glib events. The Signals js object allows one to create any old signal you want.
86 16:40:26 <jjmarin> ok
87 16:41:10 <clown> actually, if you look at line 184, you can see that it emits a 'caret-changed' signal (that's a js signal), but also passes along the atspi event — it's called 'event' in the emti() call.
88 16:42:33 <jjmarin> ok, thank !
89 16:42:52 <clown> somehow we have to tell magpie all this...
90 16:42:58 <jjmarin> yes
91 16:43:08 <jjmarin> you or me ?
92 16:43:17 <clown> i'm on vacation...
93 16:43:24 <jjmarin> then me :-)
94 16:43:25 <clown> :-)
95 16:43:37 <clown> I could write a summary and email her.
96 16:44:02 <clown> based on on the above — this is being recorded by the meetbot.
97 16:44:59 <jjmarin> #action Juanjo will write to magpie to inform about gobject introspection arspi2 bindings for javascript and the way the focus tracker have to communicate with the zoom in gnome-shell
98 16:45:26 <clown> actually, I think it might help here if she and I went over this while talking to each other (Skype?), but not sure when...
99 16:45:46 <clown> and completely time zones.
100 16:45:54 <clown> "completely different" that is.
101 16:46:15 <jjmarin> she wasn't able to attend today to the meeting
102 16:46:26 <jjmarin> maybe in the next one she can attend
103 16:47:04 <clown> okay, I can write up a summary, but no immediately.
104 16:47:10 <clown> "not"
105 16:47:37 <clown> what it the timeline?
106 16:47:44 <clown> what "is" the timeline?
107 16:48:11 <jjmarin> http://www.google-melange.com/gsoc/proposal/review/google/gsoc2013/magpie/1
108 16:48:37 <jjmarin> 13 May - 24 May: Pre-Program Preparation
109 16:48:42 <clown> sheesh, I have to log in.
110 16:48:46 <jjmarin> 1) Set up a development environment using jhbuild
111 16:48:48 <jjmarin> and
112 16:49:02 <jjmarin> 2. Familiarise myself with AT-SPI2 by creating a pyatspi2 "example". This example will show how AT-SPI2 can be used to accomplish drag and drop and was recommended to me by my mentors. It will also serve as my second preliminary contribution.
113 16:49:51 <clown> well, all the above has nothing to do with drag and drop in python, so she doesn't need to know it immediately.
114 16:50:00 <jjmarin> exactly
115 16:50:29 <clown> so, if I had the summary by next meeting, that would be okay?
116 16:50:40 <clown> although I don't know that i will be here next meeting.
117 16:51:33 <clown> But, the summary could be written up by then.
118 16:51:39 <jjmarin> I can start writing an informal message CC API, joanie and you
119 16:51:49 <clown> Sure, that will help.
120 16:51:58 <jjmarin> and then, you can get into action when you finished your holidays
121 16:52:10 <clown> well..
122 16:52:26 <clown> I will be back at work, and goodness knows what they will have for me.
123 16:52:46 <clown> oddly, I have more time for this now, then when at work...
124 16:52:55 <jjmarin> :-)
125 16:53:02 <clown> shall I info some of this?
126 16:53:26 <jjmarin> it's up to you :-)
127 16:53:43 <clown> #info Juanjo and Joseph discussed pyatspi2, events, JavaScript api to atspi, and how the latter is poorly documented.
128 16:54:24 <clown> #info the discussion led to looking at the focus tracker javascript code (https://bugzilla.gnome.org/review?bug=647074&attachment=219583), and Joseph summarizing how it works.
129 16:55:19 <clown> #info Briefly, it listens for "focus" related atspi events and emits them as JS signals that any other gnome-shell object can connect to and handle.
130 16:55:52 <clown> #action Juanjo will summarize the discussion and email it to me, api, joanie, and magpie.
131 16:56:12 <clown> #action Joseph will look over the summary and add any needed modifications by the next meeting.
132 16:56:22 <jjmarin> :-)
133 16:56:23 <clown> how's that? anything I missed?
134 16:56:33 <jjmarin> nop
135 16:56:45 <clown> wow, I actually wrote "needed". Sheesh.
136 16:56:53 <clown> how presumptuous.
137 16:57:08 <clown> so, moving?
138 16:57:09 <jjmarin> sure there will be :-)
139 16:57:13 <jjmarin> yes
140 16:57:24 <clown> #topic W3C updates
141 16:57:31 <clown> #info nothing new to report.
142 16:57:40 <clown> questions?
143 16:57:57 <jjmarin> :-)
144 16:58:00 <jjmarin> nop
145 16:58:13 <clown> #topic Marketing
146 16:58:18 <clown> jjmarin?
147 16:58:18 <jjmarin> #info nothing new to report.
148 16:58:26 <jjmarin> questions?
149 16:58:31 * clown this is going fast now…
150 16:58:33 <clown> nope.
151 16:58:39 <jjmarin> hehe
152 16:58:46 <clown> #topic Miscellaneous Time
153 16:58:56 <clown> I don't really have anything...
154 16:59:11 <jjmarin> Have you traveled for holidays ?
155 16:59:30 <clown> oh, why are joanie and API in Toronto? Especially when I am not at work (in Toronto), and am at home.
156 16:59:54 <clown> jjmarin: I may travel next week and take a few days to vista my father.
157 16:59:58 <clown> "visit"
158 17:00:19 <jjmarin> they live in Canada as well ?
159 17:00:34 <jjmarin> s/they/he
160 17:00:44 <clown> who are they?
161 17:01:18 <clown> my father? yes, immigrated to Canada after WW2
162 17:01:43 <jjmarin> In Spanish father and parents are very similar words and I misread :-)
163 17:02:00 <clown> oh, understandable.
164 17:02:19 <clown> do you know why joanie and API are here?
165 17:02:31 <clown> where here = very close to me.
166 17:02:39 <jjmarin> nop, work, but I don't know the client
167 17:03:04 <jjmarin> maybe a mobile company
168 17:03:11 <jjmarin> but no idea
169 17:04:00 <jjmarin> Bloomberg has offices there ?
170 17:04:24 <clown> I don't know, offhand.
171 17:04:56 <clown> but, if I'd known, I have tried to meet up.
172 17:05:14 <jjmarin> I know Bloomberg is one of Igalia clients
173 17:05:17 <clown> I would have tried to meet up.
174 17:05:39 <jjmarin> It seems they are very busy there
175 17:06:04 <clown> this company: http://www.bloomberg.com/company/
176 17:06:08 <clown> ?
177 17:06:22 <jjmarin> yes
178 17:06:54 <jjmarin> Igalia people mentions lately a lot http://wingolog.org/archives/2013/05/08/generators-in-v8
179 17:06:56 <clown> it looks like a news company.
180 17:06:59 <jjmarin> yes
181 17:07:30 <jjmarin> http://wingolog.org/archives/2013/05/08/generators-in-v8
182 17:07:35 <jjmarin> http://people.gnome.org/~csaavedra/news.html
183 17:07:40 <clown> i just got a call from my dentist — I have an appointment next Thu morning.
184 17:07:49 <clown> definitely won't be at this meeting….
185 17:07:53 <jjmarin> np
186 17:08:31 <jjmarin> http://blogs.igalia.com/svillar/2013/03/22/improving-the-editing-code-in-webkit/
187 17:08:47 <jjmarin> so yes, they are working with Bloomberg a lot lately
188 17:09:12 <clown> that first one is about using JS. Interesting.
189 17:10:05 <jjmarin> yes, Andy Wingo like programming languages, he also maintains Guile (a sort of Lisp from GNU)
190 17:10:46 <jjmarin> For all these posts, I have the impression is an important client for Igalia at this moment
191 17:11:09 <clown> and that client is right in my back yard, so to speak.
192 17:11:20 <jjmarin> OMG !
193 17:12:31 <jjmarin> any, they have other clients, but they don't comment about other clients
194 17:12:45 <jjmarin> s/any/anyway
195 17:13:16 <clown> okay. well, I think this is slipping away from somewhat miscellaneous into very miscellaneous.
196 17:13:29 <jjmarin> and maybe they are under sort of professional silence
197 17:13:41 <clown> yes, could be.
198 17:14:02 <jjmarin> yes :)
199 17:15:17 <jjmarin> anythin more miscellaneous ?
200 17:15:26 <clown> not from me.
201 17:15:37 <jjmarin> me neither :-)
202 17:15:50 <jjmarin> you can end the meeting then...
203 17:15:53 <jjmarin> me hides
204 17:15:56 * clown trying to remember the magic meetbot command to do that.
205 17:16:16 <clown> #help end meeting
206 17:16:24 <clown> #help
207 17:16:35 <clown> #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.