14:07:34 <API> #startmeeting
14:07:34 <Services> Meeting started Thu Jul  3 14:07:34 2014 UTC.  The chair is API. Information about MeetBot at http://wiki.debian.org/MeetBot.
14:07:34 <Services> Useful Commands: #action #agreed #help #info #idea #link #topic.
14:08:11 <API> #topic Progress towards 3.14
14:08:17 <API> just a small comment
14:09:14 <API> #info recently on the gnome-shell list, someone complained because it was slow
14:09:15 <API> https://mail.gnome.org/archives/gnome-shell-list/2014-June/msg00022.html
14:09:31 <API> #info Florian suspected of the (already fixed) children-changed bug
14:09:56 <API> #info he mentioned that would try to get the workaround into the COPR
14:10:02 <API> so, again
14:10:14 <API> although this was a workaround, I still think that was good to include
14:10:23 <API> as people would start to complain on the lists
14:10:39 * clown nods
14:10:44 <API> having said so, this week was busy to me so I don't have anything else to report
14:10:46 <API> next?
14:12:13 <mgorse> I think I found part of the reason that evolution is still inaccessible. Ie, the atk implementation for a text cell was dropped because it was constantly crashing and no one was maintaining it
14:12:22 <mgorse> not sure if that's worth #infoing
14:12:36 <joanie> probably worth infoing
14:13:34 <mgorse> #info mgorse was looking at evolution accessibility and realized that the dropped text cell accessible support has a major impact. Plans to file a bug to try to get it reinstated. It was dropped because it was crashing and no one maintained it
14:14:40 <joanie> And no one noticed because the rest of the a11y support was broken
14:15:10 <joanie> mgorse: did you have time yet to look into the Wayland a11y event plugin stuff?
14:17:42 <mgorse> I've tried to a little. I'm wondering if maliit is abandoned, which doesn't help. Been somewhat busy lately. I'm wondering if I'll end up proposing trying to add a click function to atk for the time being and proposing implementing it in gtk. Seems easier to do than updating weston, mutter, adding the appropriate security protocol and debating the specification for that (since afaik that isn't done), etc, although doing that might be best in the long
14:18:52 <mgorse> ie, it appears to me that the text protocol was added into weston with the intention of adding it into the wayland protocol but the latter wasn't done
14:19:06 <joanie> but if we add it to gtk, won't we also need to add it to webkitgtk and gecko and evince?
14:19:31 <API> well, about maliit: yes probably it is abandonware
14:19:35 <API> but in any case
14:19:48 <API> wayland is being also agressively pushed for mobiles
14:19:52 <mgorse> that's a good question; depends if something could, say, be added to the top-level window
14:19:54 <API> they would need a osk eventually
14:20:22 <API> and as we said
14:20:27 <API> that security protocol
14:20:35 <API> is something that they would need to add in any case
14:20:48 <API> our proposal is not writing by ourselves a security protocol
14:20:50 <API> due a11y needs
14:21:04 <API> but reuse the security protocol that would need to be added in any case
14:21:17 <mgorse> yeah, that's true; our protocol would be written first in any case
14:21:26 <API> anyway, it is true that I have been somewhat disconnected from the mailing list
14:21:27 <API> so
14:22:04 <API> #action API will do a quick research of wayland mailing lists, wikis, etc, in order to know more about any security protocol, and about all those services (like osk) that right now are not well supported
14:22:31 <API> having said so
14:22:36 <API> ah
14:22:41 <API> well, before asking for moving
14:22:48 <API> about the implementation of that click on atk
14:23:01 <API> mgorse, I guess that you are proposing a similar workaround
14:23:10 <API> to the current on-key
14:23:32 <API> so, a equivalent to
14:23:34 <API> https://developer.gnome.org/atk/stable/AtkUtil.html#atk-add-key-event-listener
14:23:41 <API> but for click mouse
14:23:43 <API> right?
14:23:55 <joanie> ?
14:24:34 <API> joanie, right now, the key events come from atk, through that method
14:24:50 <API> that is still one of the methods that unfortunately needs to be implemented by the toolkit
14:25:04 <joanie> independent of whether or not the key event is real or synthesized?
14:25:09 <mgorse> I was thinking of an atk_object_click() that would have a default implementation that would try the parent window if an implementation doesn't exist for the current object, so that a handler for a top-level window could handle requests for its descendants, but I haven't actually tried coding that
14:25:43 <API> ah ok
14:25:47 <API> then I misunderstood
14:25:49 <API> well, fwiw
14:25:53 <API> if you try to implement it
14:26:03 <API> atkobject has no pads anymore
14:26:15 <API> so if you want to add that as a virtual method, it would not be possible
14:26:16 <API> anyway
14:26:27 <API> probably a method like that fits better on AtkComponent
14:26:47 <joanie> unless it's text
14:26:48 <mgorse> I see one pad, but agree that AtkComopnent would make sense
14:27:02 <joanie> because if it's text the caret gets set
14:27:08 <joanie> and text might become unselected
14:27:25 <joanie> and what if we move the mouse?
14:27:31 <joanie> but don't click
14:27:34 <mgorse> oh, AtkComponent has no pads
14:27:43 <API> AtkComponent is an interface
14:27:44 <joanie> i.e. this is bigger than clicking
14:27:55 <API> you can add methods to interfaces without worrying of pads
14:28:00 <joanie> AtkComponent already has grab focus
14:28:04 <API> I removed the pads some releases ago
14:28:09 <joanie> it doesn't have mouse hover
14:28:18 <joanie> it doesn't have left click versus right click
14:28:37 <API> joanie, well, but AtkObject also don't have that
14:28:37 <joanie> it doesn't have any connection with text selection or text caret
14:28:51 <API> what I mean is that in the case of adding all that stuff somewhere on atk
14:28:59 <API> probably it makes sense to add it on AtkComponent
14:29:21 <API> anyway, I have the feeling that we are going too deep on this conversation
14:29:29 * clown reviews hierarchy
14:29:31 <API> taking into account that we are on the weekly meeting
14:29:43 <joanie> my bad
14:29:52 <API> no problem
14:29:56 <API> the conversation was interesting
14:30:03 <API> we can resume that at any other moment
14:31:24 <API> so if you don't mind
14:31:26 <API> I will move
14:31:28 <API> is that ok?
14:31:35 <mgorse> ok with me
14:31:36 <joanie> sure
14:31:38 <clown> fine
14:32:38 <API> #topic W3C updates
14:32:40 <API> clown, ?
14:33:02 <clown> Well, I was away on holidays.  I missed all the meetings.  I have no news.
14:33:37 <joanie> The Monday meeting was largely a deep dive into confusing spec text ;)
14:33:38 <clown> Did you go to the Mon meeting, joanie?
14:33:46 <API> no news means good news?
14:34:02 <clown> no news means I'm (blissfully) ignorant.
14:34:04 <joanie> At one point someone said "If we don't understand what this means, how can we expect anyone else to?"
14:34:28 <clown> what was it regarding?
14:34:34 <joanie> So I think you got some action items :)
14:34:46 <joanie> presentation/none role
14:34:50 <clown> oh, joy
14:35:03 <joanie> there was also some discussion about name calculation
14:35:12 <joanie> e.g. for rows, etc.
14:35:16 <clown> is it caam relatted (implemenation of none role in …. oh a11y name...
14:35:32 <joanie> it wasn't all name related
14:35:43 <joanie> point is, I don't recall anything GNOME-relevant
14:35:54 <joanie> it was more about getting our (W3C) docs right
14:36:05 <clown> okay, back to what I started to ask:  is the presentation/none issue about how to map those roles (caam related)?
14:36:53 * joanie goes back to the minutes
14:37:05 <joanie> (it was a deep dive)
14:37:09 * clown looks up minutes
14:38:49 * clown wonders why there are no topics.
14:39:04 * clown answer:  joanie wasn't the scribe.
14:39:05 <joanie> New scribe
14:39:25 <joanie> but if solving wayland is too deep
14:39:34 <joanie> I think this line of discussion is as well
14:39:35 <joanie> ;)
14:39:54 <clown> okay, I will look over those minutes later, and look for my actions.
14:41:11 <clown> thanks joanie
14:41:17 <API> moving then?
14:41:23 <joanie> sorry for not having a better memory
14:41:31 <joanie> I just remember JamesN's quote
14:41:57 <API> well, jjmarin is not here, so I will skip marketing
14:41:58 <API> so
14:42:06 <API> #topic Miscellaneous time
14:42:15 <API> something not scheduled to mention?
14:44:34 * clown *crickets*
14:44:57 <API> taking into account that the crickets don't have anything to say
14:45:00 <API> I will close the meeting
14:45:07 <joanie> there's a reason for that
14:45:09 * joanie googles a study
14:45:14 <API> thanks everybody for coming to this light meeting
14:45:21 <API> #endmeeting