14:34:44 #startmeeting 14:34:44 Meeting started Thu Apr 28 14:34:44 2011 UTC. The chair is API. Information about MeetBot at http://wiki.debian.org/MeetBot. 14:34:44 Useful Commands: #action #agreed #help #info #idea #link #topic. 14:35:08 #topic Aline 14:35:23 well, this is just 14:35:27 share that Aline 14:35:38 was accepted on the Women 14:35:41 outreach program 14:35:47 she will work on 14:35:49 documentation 14:36:02 as some already know, she started with accerciser 14:36:22 joanie, will be her co-mentor 14:36:25 additionally 14:36:33 Meg Ford was also accepted 14:36:42 and will work also on a11y stuff 14:36:45 specifically 14:36:49 art a11y 14:36:52 on the themes 14:36:56 joanie, something else? 14:36:58 David Bolter of Mozilla being the other co-mentor. 14:37:08 great 14:37:25 But everyone should feel free to pitch in 14:37:35 Aline also has some coding ideas 14:37:44 yes this is good 14:37:46 as we said 14:37:46 ATK implementation/custom widget, etc 14:37:56 if she is trying to do something 14:38:01 using the atk doc 14:38:02 which is more the area of expertise of those who are not me ;-) 14:38:06 or the a11y doc in general 14:38:16 if she is not able to do that, means that the doc is incomplete 14:38:25 is a good test-case for the documentation 14:38:25 (or wrong) 14:38:49 ok, so, using meetboot 14:38:58 btw: this is the first meeting trying to use it 14:39:08 #action Joanie will follow up with Aline to see what the next steps are 14:39:16 #info Aline was accepted on the Women Outreach Program 14:39:34 #info Meg Ford was also accepted 14:39:45 Hi everybody 14:39:48 Sorry I'm late 14:39:50 #info both will work on a11y stuff (documentation, art) 14:39:58 alibezz, we are finishing your point ;) 14:40:14 but basically is reporting that you and Meg Ford were accepten on the WOP 14:40:21 and will work on a11y related tasks 14:40:23 hi alibezz 14:40:30 alibezz, do you want to add something? 14:41:08 For the record.... 14:41:25 #agreed The team is thrilled to have Aline 14:41:47 h 14:42:00 ups sorry 14:42:45 alibezz, ? 14:43:10 I'm glad to work with you all 14:43:22 :-) 14:43:31 ;), ok thanks 14:43:35 is Meg Ford also working with arts, isn't it? 14:43:43 alibezz, yes 14:43:47 she will work to 14:43:51 improve the a11y themes 14:44:13 is she here? 14:44:29 * API looking people connected 14:44:30 hmm 14:44:31 no 14:44:47 And she's not been in touch with us directly (to my knowledge) 14:45:11 well, I guess that this would happen eventually 14:45:17 oh 14:45:21 anyway, she already have an mentor 14:45:27 is David here? 14:45:34 alibezz, no 14:45:35 My other mentor:) 14:45:38 oh ok 14:45:51 nope. I could ask him to join (he's in #accessibility on mozilla's irc serve). 14:45:56 sorry I'm late... Guess I lost my own point 14:46:14 alibezz, no problem 14:46:16 soo 14:46:24 anyother question or comment? 14:46:31 can we move to next point? 14:46:34 * joanie nods 14:46:40 +1 14:46:47 +1 14:47:01 #item ATK Hackfest 14:47:08 ups 14:47:09 #topic ATK Hackfest 14:47:13 thanks 14:47:21 well, it is really near! 14:47:30 I will write a post about that 14:47:32 and again 14:47:40 please people, update the wiki 14:47:43 and do your homework 14:47:54 mgorse was doing a good work today 14:47:59 reviewing some of the bugs 14:48:04 +1 14:48:14 #action update wiki 14:48:26 API need to add a name 14:48:27 I think 14:48:32 #action do the homework: review atk bugzilla 14:49:04 hmm 14:49:07 somethink like 14:49:11 #action Joanie will coordinate with API on drafting a schedule 14:49:35 and in the case that it is a general thing? 14:49:41 dunno 14:49:45 but we shall see shortly 14:49:45 add tota11y 14:49:46 All? 14:49:56 +1 aleiva 14:50:07 joanie, btw, nicknames are case sensitive 14:50:35 anyway, we can hack the minutes later 14:50:35 not sure if they are for the bot but yeah 14:50:40 so about this topic 14:50:48 any question, doubt, etc? 14:50:54 ah, btw 14:50:55 panic? 14:50:58 about theopen day 14:51:07 fer aleiva msanchez and me 14:51:23 remember that you are volunteers for that 14:51:32 #action aleiva will prepare his presentation on Orca soon. 14:51:33 yup, I need to ping msanchez 14:51:37 ;-) 14:51:43 #action API will sent the schedule for the open day 14:51:50 and last information 14:52:07 remember that this open day 14:52:17 will have the benefict that a local LUG 14:52:28 GPUL would donate some money to the hackfest 14:52:45 so most of the lunches will be covered by this 14:52:51 so again 14:52:55 anything else? 14:52:58 doubts, questions? 14:53:25 only 14:53:38 I leave that afternoon as wiki said 14:53:53 add my talk first or second 14:53:58 yes, np 14:53:59 thanks ;) 14:54:03 it would be better if the second 14:54:12 in order to make the "introduction to gnome accessibility" 14:54:17 before "orca is cool" 14:54:21 +1 14:54:22 perfect 14:54:24 so if it is second 14:54:30 that introduction will be short 14:54:45 so lets move 14:54:48 #topic Testing 14:55:04 joanie, I guess that you added this topic 14:55:13 Yeah, for a few reasons 14:55:34 One is just that I really want us to have an accessible 3.0.2 release 14:55:40 as accessible as it can be anyway 14:55:56 so even though we always say it, and all know it needs doing 14:56:09 we really, really need to be actively testing this stuff 14:56:31 filing bugs, harassing people outside of the team whose fixes (or approvals of patches) we require 14:56:34 etc. 14:56:37 jhernandez|eee, you have just arrived and the best moment 14:56:46 it is planned a new spin for the a11y distro? 14:57:11 of course, this week we're going to have the fouth one 14:57:38 so 14:58:06 I'll need all updated tarballs 14:58:38 that's all 14:58:54 jhernandez|eee, well a question 14:58:57 for example 14:58:58 in my case 14:59:04 all the patches included on the 14:59:10 custom gnome shell 14:59:14 were committed on master 14:59:20 nice! 14:59:22 woo hoo! 14:59:23 but was after the 3.0.1 release 14:59:27 congrats! 14:59:29 so it wouldn't be there 14:59:35 until 3.0.2 release 14:59:36 sooo 14:59:47 so tarball is still needed 14:59:47 as the plan is testing all the stuff for 3.0.2 14:59:55 I guess that I should keep that tarball 15:00:01 right? 15:00:07 API could you roll a new one 15:00:08 * API joanie thinks that yes 15:00:14 joanie, a new one? 15:00:16 ah 15:00:24 a 3.0.1 but with the patches? 15:00:25 because I think other keyboard related stuff went in 15:00:32 really? 15:00:35 well, not sure 15:00:39 but just in case 15:00:41 some of my bugs were closed as dupes 15:00:48 ah ok 15:00:50 with commits being present 15:00:54 prior to your work 15:01:04 well, so trying to use this meetboot thing 15:01:09 of course, if the other commmits don't have a11y-related patches, o 15:01:10 ok 15:01:24 #info we really need to test stuff towards 3.0.2 release 15:01:41 #action jhernandez|eee will create a new spin soon 15:01:52 #action people should update their tarballs 15:02:03 #action joanie will create a new Orca tarball 15:02:14 #action API will create a new gnome-shell tarball 15:02:18 joanie, I think that you are still "people" 15:02:19 API: but, I'll considering if some module has a strong depend on gnome-shell 3.0.2 (i think is not the case) it could be a problem 15:02:22 ;) 15:02:29 I'm not convinced 'people' is gonna work 15:02:31 jhernandez|eee, hmmm 15:02:36 well, but that spin 15:02:45 is a spin to test things before the testing distro 15:02:56 yes 15:02:57 so that wouldn't include any dependency to any 3.0.2 release 15:02:58 right? 15:03:06 right! 15:03:10 ok 15:03:16 s/3.0.2/3.0.1/ 15:03:22 sorry for confusing 15:03:24 well, so anohter question, comment, information, suggestion? 15:03:24 :s 15:03:26 jhernandez|eee, np 15:03:30 ok 15:03:57 well, so anohter question, comment, information, suggestion? 15:04:04 * joanie notes the next topic might be redundant 15:04:39 * API checking 15:04:55 #topic GNOME 3.0.2 15:05:02 #info already discussed on previous topic 15:05:05 ;) 15:05:08 :-) 15:05:11 joanie, do you want to add something? 15:05:20 nope, I'm set 15:05:22 thanks 15:05:39 well, we are doing it in a really good time 15:05:47 so lets go to the next 15:05:58 #topic CSPI vs libatspi 15:06:03 I added this topic today 15:06:17 mostly because the chatting we have yesterday mgorse and tbsaunde 15:06:19 on #a11y 15:06:33 I thought 15:06:34 wrongly 15:06:42 that the core C library used by pyatspi2 15:06:47 using gobject introspection 15:07:04 has the purpose to be used just internally 15:07:16 but mgorse and tbsaunde told me yesterday that it would be possible to use it 15:07:19 on a C program 15:07:35 although mgorse said that the C library was not exactly CSPI 15:07:36 so 15:07:39 mgorse, one question 15:07:56 if you want to create a AT using C 15:08:05 and you use libatspi 15:08:17 there is something that you can't do that you can do with pytaspi2? 15:08:35 does libatspi lacks any functionality 15:08:47 Hmm findDescendant is the one thing I can think of off-hand 15:08:48 * API in related news, not sure if this is the proper name for that library 15:09:00 which was the case with cspi as well 15:09:13 well, I'm asking that 15:09:17 because if this is the case 15:09:26 that has to be implemented in pyatspi I think, and libatspi doesn't have an equivalent 15:09:39 API: well, that's what the .so is named so I'd say its the right name ;) 15:09:42 there is any reason to keep talking about CSPI? 15:10:00 why can say that CSPI is obsolete and the new C API is libatspi? 15:10:09 tbsaunde, thanks ;) 15:10:58 API: I think that makes sense. Someone could write a compatibility layer, but I don't have plans to do that myself, and the only reason would be for compatibility with apps using cspi, but those apps could be ported to libatspi if someone wanted to 15:11:15 mgorse, well, not sure if makes sense a compatibility layer 15:11:17 I mean 15:11:24 if someone is interested to 15:11:34 create a AT program use C 15:11:40 if it was already using CSPI 15:11:48 it would be just move to a new API 15:11:49 after all 15:11:57 Orca in the past also needed to do that 15:12:05 yeah 15:12:08 due the at-spi1 python bindings 15:12:24 on that "old times" with different python bindings 15:12:30 soo 15:12:49 for example, we could suggest that in order to integrate dasher on gnome 3 15:12:59 he can move from CSPI to libatspi 15:13:04 mgorse, what do you think? 15:13:20 yeah, that should work 15:13:42 ok, and in the same way 15:13:43 this 15:13:45 https://live.gnome.org/Accessibility/ThreePointTwo/RelatedProjects#CSPI 15:13:48 can be removed 15:13:55 and the final conclusion is: 15:13:56 * joanie nods 15:14:08 "CSPI is dead, migrate to libatspi" 15:14:53 question: Is Patrick still working on Dasher? 15:15:32 joanie: Good question. He had plans to when I saw him at AEGIS 15:15:49 he was also on FOXDEM 15:15:51 ups 15:15:52 FOSDEM 15:15:56 but I forgot to ask 15:16:00 I should send him an email--he might not know about libatspi for instance 15:16:17 mgorse, I had the plan to send that to the list 15:16:18 I haven't seen him on IRC in a long time 15:16:28 so I can add him on CC 15:16:30 mgorse, btw 15:16:34 ok 15:16:42 this meas that it would be good to add libatspi documentation 15:16:45 Someone has an ai, but who? 15:16:50 on the gnome developers page 15:16:54 me 15:16:57 joanie, hold on 15:17:06 * joanie apologizes 15:17:06 we will write all that meetboot stuff at the end 15:17:17 API: I thought it was already there, but I haven't looked in depth at it 15:17:25 mgorse, ok 15:17:31 so lets now conclude this 15:17:49 #info libatspi provides access to any AT program based on C 15:18:00 #info you can do most of the things that you can already do with pyatspi2 15:18:41 #agreed instead of provide a CSPI compatible with at-spi2, people should use/migrate to libatspi 15:18:55 #action API will update the wiki removing CSPI as a issue 15:19:12 #action API will send a mail about this agreement on the list, including patrick on CC 15:19:30 #action API documentation page should be reviewed 15:19:32 anything else? 15:19:44 somewhat tangential: 15:19:54 anyone know what to use if writing an AT in javascript? 15:20:27 c -> libatspi; python -> pyatspi2; javascript -> ? 15:20:51 and, it's possible there is no answer... 15:20:59 hmm 15:21:02 js using gtk or js for the web? 15:21:12 mgorse, as pyatspi2 is based on gobject introspection 15:21:18 js using gtk (or embedded inside gnome-shell) 15:21:24 I guess that there are a gir file and so on 15:21:25 right? 15:21:34 API, there should be. good point. 15:21:48 Hmm. Does Javascript support gobject-introspection? If so, then in theory you can use the libatspi typelib somehow, but I don't think anyone has tried so far 15:22:00 mgorse, it does. 15:22:06 gjs does, that is. 15:22:06 I mean that I could call atk methods using javascript on gnomeshell 15:22:10 mgorse, yes 15:22:15 mgorse in theory you can 15:22:18 this is how gnome shell does 15:22:27 I mean that gnome-shell ui is written in javascript 15:22:34 and they call gtk, clutter, st calls 15:22:38 you can use looking glass in gs 15:22:41 I was also being able to call atk 15:22:50 with some issues, but able to 15:22:51 to test it fastly 15:23:01 API, yep. gs-mag does very simlilar things, but I've yet to actually try to hit the atspi layer. 15:23:05 mgorse, so if there is a gir file for libatspi 15:23:16 similar things to what gnome-shell does. 15:23:21 I guess that you can do that 15:23:38 clown, so you would like to call libatspi methods on your gs-mag code? 15:24:02 API, I'm thinking about the bug that joanie filed 15:24:08 well, as mgorse said that it is possible 15:24:12 for doing focus tracking in gs-mag 15:24:30 clown, would you like to test if calling at-spi2 via libatspi 15:24:33 and which Dan suggested doing focus traking in gnome-shell itself. 15:24:34 with javascript is feasible? 15:24:49 clown, Im thinking on a new action item 15:24:54 and Im searching a volunteer ;) 15:24:56 yes essentially. 15:25:12 but my thinking is based on an assumption 15:25:24 #action clown will investigate if it is feasible to call libatspi using javascript (ie: on gnome-shell) 15:25:26 that focus tracking needs to use atspi to track, well, focus events. 15:25:38 will the action is just "investigate", so no pressure ;) 15:25:53 can we conclude this topic then? 15:25:53 * clown new mantra: no pressure, no pressure... 15:26:06 #topic miscellaneous time 15:26:11 I see some new faces on the meeting 15:26:13 in summary 15:26:20 we usually use the last minutes of the meeting 15:26:30 just in case someone want to ask, report 15:26:34 something *short* 15:26:37 not included on the agenda 15:26:49 so, anyone wants to add something on this miscellaneous time? 15:27:19 API, I looked at the agenda earlier in the week, and saw something about you reporting from the release team? Just curious... 15:27:22 will it be needed/worth to make other at-spi bindings for other languages? 15:27:56 * clown suggests dealing with webczat's question first. 15:28:25 actually, gnome things are just python/c/js so... 15:28:28 webczat: at face value: "yes" :-) 15:28:55 presumably there's something for Java (who uses Java anymore?) 15:28:58 I made libatspi's method names like what seems to be standard for C libraries, but the original pyatspi's names were more Pythonic, so pyatspi has a lot of renaming for instance 15:29:53 clown: Many people don't like java, but others still use it. 15:30:05 Someone could, ie, use libatspi directly in Python using pygi (ie, "from gi.repositories import Atspi", ...) 15:30:09 webczat: I like java... 15:30:31 clown: i also do. 15:30:39 s/repositories/repository/ 15:30:45 clown: there are worse languages 15:30:48 clown, hmm, that was last week agenda 15:31:04 oh, right, API. Never mind. 15:31:07 but I didn't see any libatspi bindings to java. there is an atk wrapper for swing and java-access-bridge that is probably deprecated but that isn't it 15:31:27 webczat, summary 15:31:35 we are not too many people working on a11y 15:31:38 tbsaunde: "there are worse languages" yep. C++... 15:31:43 the good thing of the gobject introspection 15:31:46 based binding 15:31:55 is that we just need to create that on C 15:31:57 and "magically" 15:32:08 we will have python and javascript (we hope) bindings 15:32:14 java bindings would be awesome 15:32:16 but AFAIK 15:32:21 there is not one with time to do that 15:32:23 and AFAIK 15:32:30 right now 15:32:41 there aren't any JAVA bindings based on gobject introspection 15:32:51 * clown looks to see if Peter Korn is around. Nope. I volunteer Peter for the Java bindings. 15:32:52 Fwiw, I found a few bugs and memory leaks in pygobject while working on pyatspi, so there's no guarantee that there won't be similar for Javascript. 15:33:08 API: you could use jni, its an ffi for java 15:33:26 tbsaunde, well, I could 15:33:27 but sorry 15:33:30 yeah, java classes can load native libs 15:33:33 i don't have time to check that 15:33:42 and check if it is working 15:33:43 ie 15:33:52 we don't know anything 15:33:55 about the current state of JAW 15:34:04 so 15:34:05 API: yeah, not suggesting *you* do anything 15:34:10 tbsaunde, yes I know 15:34:20 but you I used *me* as a example 15:34:32 I mean that if someone wants to use Java 15:34:36 API: I mean if someone wanted to use libatspi from java they don't need gobject-introspection to do it, they just use jni 15:34:45 he should find the way to do that 15:35:11 tbsaunde, ok, good to know 15:35:12 anyway 15:35:16 webczat, about your question 15:35:19 "other bindings" 15:35:41 are you thinking just in Java, or do you have other languages in mind? 15:36:20 Actually I was just curious. 15:36:48 webczat, well again 15:36:50 in summary 15:36:53