15:07:07 <API> #startmeeting
15:07:07 <Services> Meeting started Thu Mar 20 15:07:07 2014 UTC.  The chair is API. Information about MeetBot at http://wiki.debian.org/MeetBot.
15:07:07 <Services> Useful Commands: #action #agreed #help #info #idea #link #topic.
15:07:22 <API> #topic Progress towards 3.12
15:07:55 <API> #info hard code freeze is here
15:08:04 <joanie> praise zombiejesus
15:08:30 <API> #info from now, just important (criticals, crashes, etc) bugs, and every of them with RT bless are allowed
15:08:57 <API> #info in my specific case, AtkValue changes got included on 3.11.92
15:09:01 <API> hmm
15:09:15 <API> #info in API specific case, AtkValue changes got included on 3.11.92
15:09:29 <clown> congrats
15:09:29 <API> sorry, "my specific case" is not exactly #info material
15:09:53 <API> #info API was being working also on some evince bugs, but last one
15:10:09 <API> #info bug 724965
15:10:09 <Services> 04Bug http://bugzilla.gnome.org/show_bug.cgi?id=724965 normal, Normal, ---, evince-maint, UNCONFIRMED, Implement a means to get the accessible text of the next/previous page
15:10:31 <API> #info included 10 patches, so getting them into 3.11.92 needed too much review in a hurry
15:10:48 <joanie> 'twas a valiant effort though
15:10:48 <API> #info evince maintainer mentioned that he will try to review and get them into 3.12.1
15:11:07 <API> Im not sure if that is something we can mention on release notes
15:11:25 <API> anyway, joanie bugs were solved, and afaik, included on .11.92
15:11:28 <jjmarin> Well, we have a generic Evince have better accessibility
15:11:34 <API> and as Im done
15:11:45 <API> joanie, could you elaborate what I just said with pretty infos?
15:11:52 <joanie> heh
15:11:56 <joanie> surrrrreeeee
15:12:06 * joanie thinks
15:12:59 <joanie> #info There were several Evince accessibility issues (including a crash and broken AtkText stuff from the old/original accessibility implementation from years past) that are now fixed.
15:13:19 <joanie> #info Thus the generic "Evince has better accessibility" statement is indeed correct.
15:13:22 <joanie> :)
15:13:39 <joanie> #info Joanie is now working on Orca + Evo accessibility.
15:14:08 <joanie> #info Some things are working that were not, but there may still be larger issues. Thus we cannot yet say if Evo 3.12 is really accessible.
15:14:27 <joanie> Should we chat about the redhat dead accessible bug?
15:14:41 <joanie> other than that I think I'm done.
15:14:41 <API> well, in general
15:14:55 <clown> not the RDAB??
15:15:09 <joanie> yes <insert scary music here>
15:15:18 <API> #info as mentioned now is about catching critical bugs
15:15:24 * clown has no idea what the RDAB is...
15:15:49 <API> #info one one that appereared recently, although not affecting directly orca is this:
15:15:55 <API> https://bugzilla.redhat.com/show_bug.cgi?id=972257
15:15:55 <Services> 04Bug 972257: unspecified, unspecified, ---, vhumpa, ASSIGNED , GError: The name :1.0 was not provided by any .service files
15:16:12 <API> in any case, in general, it would be good to try to check last gnome
15:16:27 <API> Im pretty sure that due the hard code freeze deadline
15:16:41 <API> a lot of stuff was being pushed
15:16:44 <API> having said so
15:16:49 <API> joanie, about that specific  bug
15:16:55 <API> do we need to chat more about it?
15:17:15 <API> I thought that what we chatted some minutes ago on #a11y
15:17:15 <jjmarin> I'd like to know if API understand my bug report #726670
15:17:19 <joanie> as I mentioned in my last comment on that bug, we are discussing if something should be added to at-spi2
15:17:26 <API> setted a kin of plan
15:17:30 <API> *kind
15:17:41 <joanie> if we think something should be added to at-spi2, then we can file an upstream bug
15:17:45 <joanie> and tell downstream
15:17:52 <API> well, from #a11y, I think that what you proposed:
15:17:57 <joanie> but it wasn't clear to me if we agreed on that
15:18:20 <API> <API> mgorse, it would be possible to check if the app is valid when asked for the root?
15:18:29 <API> <mgorse> yeah, it could send a message to the app at some point and delete the app if the method returns an error saying that the app is no longer on the bus
15:18:37 <API> <joanie> maybe even check when iterating through desktop children?
15:18:37 <API> <joanie> desktop == registry desktop
15:18:37 <API> <joanie> because that is where people seem to be getting bitten by this
15:18:52 <API> there are some skips, but I think that in that way the summary is better
15:18:54 <API> so
15:18:58 <API> for me that is a good plan
15:19:02 <API> mgorse, joanie ?
15:19:08 <joanie> I think so
15:19:11 <joanie> it's up to mgorse though
15:19:46 <mgorse> It needs to be fixed, and that should at least make it stop happening, and it shouldn't take long to add, and there's no time, so...
15:19:58 <joanie> no time?
15:20:19 <mgorse> yeah, .0 needs to be out on Monday, and need to get freeze break approval from release-team, etc
15:20:34 <joanie> I don't think this is freeze break worthy
15:20:48 <mgorse> oh
15:20:57 <joanie> my question is, do you think this is worth doing at all?
15:21:15 <mgorse> it's probably an improvement over what we have, anyway
15:21:20 <joanie> k
15:21:37 <joanie> #action Joanie will file an upstream bug and notify downstream regarding the dead app issue.
15:21:37 <API> I agree that it would be good to have in place that improvement
15:21:42 <API> about freeze worthy
15:21:46 <API> as I mentioned
15:21:50 <API> this is not breaking Orca
15:22:00 <joanie> because Orca is smart :P
15:22:02 <API> is causing problems on Accerciser and on dogtail scripts
15:22:11 <joanie> jhernandez: might want to add smarts to Accerciser
15:22:12 <API> so I think that as far as this got solved
15:22:16 <jhernandez> heh
15:22:18 <API> would make dogtail scripters happy
15:22:29 <API> no matters if gets in time for 3.12
15:22:39 <jhernandez> acerciser is affected by that bug, but it doesn't make Accerciser crash
15:22:53 <joanie> no, it just makes it unusable
15:23:03 <joanie> because the tree doesn't get built
15:23:12 <joanie> so you cannot inspect or interact with any accessible at all
15:23:19 <jhernandez> yeah
15:23:27 <joanie> but, no, it doesn't crash. :P
15:23:30 <API> jhernandez, and I don't see you really hating us for suggesting that fixing this is not a freeze-break candidate
15:23:54 <API> so for now, we will assume that: a) needs to be fixed, b) lets not ask a freeze-break for now
15:24:13 <API> as with evince previously commented bug
15:24:19 <API> probably is a good candidate for 3.12.1
15:24:30 * jhernandez agrees
15:24:38 * clown nods
15:24:39 <API> joanie, talking about making dogtail scripters happy
15:24:46 <API> could you create an upstream bug for this?
15:24:58 <joanie> did you see my action item?
15:25:02 <joanie> I'm already doing it now
15:25:55 <API> joanie, obviously I didnt
15:26:04 <API> sorry I was writing at the moment
15:26:10 <API> and thanks
15:26:14 <API> having said so
15:26:19 <joanie> https://bugzilla.gnome.org/show_bug.cgi?id=726781
15:26:19 <Services> 04Bug 726781: normal, Normal, ---, at-spi-maint, UNCONFIRMED, Check for and remove "dead" applications when iterating through children of the registry's desktop
15:26:24 <API> anything else onthis topic?
15:26:37 <joanie> jhernandez got his freeze break :)
15:26:55 * joanie wonders if it is because fredp saw her snarky comments
15:27:17 <jhernandez> xDDD
15:27:37 <jhernandez> yeah, I got it
15:27:59 <jjmarin> How do you feel about  #726670  ?
15:28:12 * jhernandez owes a beer to joanie for those snarjy comments :P :P
15:28:14 <API> jjmarin, ah yeah sorry
15:28:18 <jhernandez> *snarky
15:28:21 <API> yes now I understand what you meant
15:28:25 <API> with "windows"
15:28:28 <API> in any case
15:28:35 <API> although I will take a look to that one
15:28:46 <API> I also don't see that bug as a freeze-break candidate
15:28:47 <API> imho
15:29:13 <jjmarin> Ok
15:29:26 <joanie> btw I have a board meeting so I'm multi-attending now
15:29:50 <API> joanie, I guess that you think the same about that bug, right?
15:30:04 <API> that bug == gnome-shel windows view at the overview
15:30:06 <joanie> 1 sec
15:30:08 <joanie> concall codes
15:30:41 * joanie reads
15:30:51 <jjmarin> I think this is the kind of bug that makes people think gnome shell is not accessible, but it doesn't make people feel we are more accesible is it is fixed
15:30:52 <joanie> I agree that it's not a freeze-break candidate
15:30:56 <joanie> but something that should be fixed
15:31:06 <joanie> jjmarin: no Orca user is going to navigate that way
15:31:13 <joanie> but I agree it should be fixed
15:31:21 <jjmarin> I agree
15:31:30 <joanie> alt-tab is what is most efficient
15:31:35 <jjmarin> exactly
15:31:50 <jjmarin> ok, move on
15:31:58 <API> #topic w3c updates
15:32:27 <API> #info API didn't send yet that email about what failed on extending roles at ATK
15:32:30 <API> sorry, I was busy
15:32:43 <API> eventually, I will send it
15:32:44 * clown collects his thoughts.
15:32:53 <API> having said so, giving the floor to clow
15:32:56 <API> clown
15:33:21 <clown> #info There were two new topics under discussion at this week's ARIA telecon.
15:33:57 <clown> #info They were (1) another way of making role values extensible, and (2) adding a new role for <input type="search">
15:34:11 <clown> #info nothing is resolved, but, briefly the status of each is:
15:34:46 <clown> #info Desire to allow third parties to add their roles to the role attribute as long as they leave a default fall back role
15:35:15 <clown> #info for example  <div role="epub-chapter region" …>
15:36:04 <clown> #info Here the Epub group has marked the div as their "chapter" role.  If the AT doesn't understand what that means, then it uses the "region" fall back
15:36:31 <clown> #info  Current status of this proposal is that it will be formally written up, and then discussed further.
15:36:56 <clown> #info Regarding the new search role:  HTML 5 has that new input type.
15:37:57 <clown> #info There is a question of how it is exposed through a11y APIs.  Currently, I think, type="search" comes out as a text entry role
15:38:41 <clown> #info But, if there is a new, say ATK_ROLE_SEARCH, and one wants to use that in another language such as SVG, then there need to be an ARIA role "search".
15:38:54 <clown> #info however, the deatils are still under discussion.
15:39:01 <clown> done, questions?
15:40:02 <API> well, no questions
15:40:04 <API> thanks for the info
15:40:06 <API> just a comment
15:40:10 <clown> wlcm
15:40:19 <API> that "if the AT doesn't understand what that means"
15:40:32 <API> was also one of the problems of extending roles
15:40:35 <API> on ATK
15:40:45 <API> ATs, as implementors
15:40:52 <API> want a spec to play with
15:41:01 <API> afaik, joanie can correct me
15:41:04 <API> ;)
15:41:10 <joanie> (in board meeting)
15:41:18 <clown> yes, API, your email will be apropos to this discussion.
15:41:27 <API> that "I'm extending a role" led to the AT to guess about the roles
15:41:54 <clown> was there a fallback, API?
15:42:06 <API> for me is a kind of disconnection between the spec implementors and the AT developers
15:42:06 <jjmarin> AI for ATs :-)
15:42:25 <API> clown, no,
15:42:30 <API> but I think that this will lead
15:42:47 <API> that most ATs will only take care of the fallback
15:42:51 <API> instead of guessing
15:43:05 <API> and using as default a fallback seems an odd definition of fallback to me
15:43:32 <API> but any case, this is a ranty form of my point
15:43:41 <API> eventually I will send that email with proper wording
15:43:48 <API> so after this rant
15:43:53 <API> more comments, questions doubts?
15:44:27 <jjmarin> nope
15:44:36 <clown> me neither :-)
15:45:16 <API> so moving then
15:45:19 <API> #topic Marketing
15:45:20 <API> jjmarin, ?
15:45:21 <jjmarin> I think today will be even shorter
15:45:28 <jjmarin> #info The release notes for 3.12 features a couple of short lines related to accessibility : "Accessibility has been improved for links in labels, menus and the color chooser dialog." and "Evince features improved accessibility for reading PDFs. Further accessibility improvements are planned for next release."
15:45:37 <jjmarin> done
15:45:44 <jjmarin> questions or suggestions ?
15:47:01 <API> well, yes, I forgot that release notes needs to be brief
15:47:16 <API> I mean that I was worried
15:47:27 <API> about the details of evince improvements
15:47:35 <API> how to wrote that
15:47:51 <API> but that concern is worthless, as there is no room for details
15:48:02 <API> so for me looks good
15:48:03 <API> thanks
15:48:07 <jjmarin> np
15:48:32 <jjmarin> Details are usually for major enhacements
15:48:56 <jjmarin> For example, Tagged PDF support
15:49:17 <jjmarin> and explain what Tagged PDF support is
15:49:45 <API> yes, true, but it is also true that a lot of the previous evince support needed to be rewriten
15:50:01 <joanie> (more than expected)
15:50:22 <API> in the sense that for 3.12, evince a11y support code is mostly new compared with 3.10
15:50:44 <API> not sure how much of the 3.8 code is still there
15:51:07 <API> in any case
15:51:07 <jjmarin> yes, but usually all this hard work is not very glamouruse for release notes
15:51:17 <joanie> exactly
15:51:32 <joanie> BUT we are still expected to have glamor from our team each release
15:51:36 <API> yes, users want to ride a speeder car, doesn't worry about a engine change
15:51:50 <joanie> the fact that we're working hard to make things better (and making great progress) apparently isn't enough
15:51:57 <API> so after this car simile
15:52:05 <API> more comments, questions suggestion on this topic?
15:52:43 <jjmarin> I think next release will be more verbose in the release note rlt a11y :-)
15:53:02 <API> I hope so ;)
15:53:15 <API> and I will take at the last comment
15:53:26 <API> #topic miscellaneous time
15:53:27 <jjmarin> all of you are doing the work to make it happen :-)
15:53:38 <API> something (short) not scheduled to mention?
15:54:55 <jjmarin> Switch to berserk mode and keep hacking. Go for 3.12 ! :-)
15:57:34 <clown> #info There is some suggestions regarding new UI for focus/caret tracking in the magnifier.
15:57:44 <clown> #info https://mail.gnome.org/archives/gnome-accessibility-list/2014-March/msg00003.html
15:58:27 <clown> #info Bas (a user) is in favour of keeping the mode separate for focus vs caret tracking.
15:58:43 <clown> a sample size of 1!  whee!
15:59:27 <API> well, plus your own opinoin
15:59:30 <API> and joanie's one
15:59:55 <API> I assume that that would be enough for allan day that collapsing options was not a good idea
16:00:05 <API> at least I hope for that
16:00:06 <jjmarin> I don't understand the advantage of joining them
16:00:44 <jjmarin> apart from a more compact UI
16:00:57 <clown> jjmarin, just to play  the devil's advocate, stepping back, your average user doesn't (?) distinguish between focus and caret.
16:01:27 <clown> they think in terms of "point of regard" or some such.
16:01:36 <clown> what am I looking at?
16:01:41 <clown> so to speak.
16:02:01 <jjmarin> ok, but people who need the feature know all that stuff
16:02:05 <clown> I *think* that is Allan's perspective.
16:02:10 <jjmarin> or they should :-)
16:02:12 <clown> right, jjmarin.
16:02:31 <clown> jjmarin, feel like posting to the thread with your thoughts?
16:02:46 <jjmarin> ok
16:02:52 <clown> thanks!
16:02:56 <jjmarin> np
16:03:03 <API> and over meeting time
16:03:12 <API> so unless someone jumps to add something else
16:03:17 <API> I'm closing the meeting
16:03:25 <API> #endmeeting