mozilla :: #accessibility

13 Feb 2017
14:02davidbyo
14:05tbsaundeheya
14:13davidbhi hi
14:14davidbtbsaunde: how goes the world of debugging?
14:14davidbI'm wondering if we have nuclear options.
14:16tbsaundedavidb: was meaning to ask if we see any change from fridays patch
14:16tbsaundeif not yes I'm planning on doing some rather nuclear things
14:19davidbtbsaunde: at first look I think the crashes are reduced, except maybe for ClearChildDoc
14:20tbsaundeeh heck I'll just fire the missels and if we figure out what's going on someday we can undo the mess
14:25davidbtbsaunde: what is the collateral damage assessment?
14:34tbsaundedavidb: some impact on perf, but its uncelar to me how big a deal it is, esp compared to IPC and stuff
14:45bhearsumhi folks, kindof random question about html inputs. i've got some fields that are integers, but i'm thinking of changing them to type="text" because they sometimes steal mouse scroll events (and the values change). are there any a11y downsides to doing that?
14:45davidbtbsaunde: yzen's tbpl push uncovered some crashes... might be something useful in there. He's around at lunch and pm.
14:46yzentbsaunde hi, this kinda looks related to the changes in that bug: https://treeherder.mozilla.org/logviewer.html#?job_id=76519310&repo=try&lineNumber=2763
14:46davidbbhearsum: I don't think so.
14:47bhearsumexcellent
14:48davidbtbsaunde: MarcoZ is a good one to ping to try it out
14:48davidbs/tbsaunde/bhearsum
14:49bhearsumah, i'm not sure it's worthwhile to go to that trouble...this is actually for balrog's admin interface - he'd have to spin up an entire development environment to give it a go
14:51tbsaundeyzen: that's curious, and maybe reasonable given all the weird ways shutdown stuff can go, but I don't know that it helps with understanding my problems
14:53yzentbsaunde also i guess not that unexpected but as part of the test (early in the beginning) i get an assert when handling a11y event, when we don't have the ipc doc yet..
14:53bhearsumdavidb: thank you :)
14:53davidbbhearsum: yw and ty for checking!
14:54tbsaundeyzen: well, could be the ipcDoc has been shutdown
14:55yzennot that it's not constructed yet ?
14:55bhearsumdavidb: np!
14:56tbsaundeyzen: well, could be either
14:57tbsaundewould maybe help to know what sort of event it is, but all I think we can really say is ipcDoc is null and it would be nice if it wasn't
15:05yzentbsaunde it is a doc laoded event
15:05yzens/laoded/loaded
15:06tbsaundethat would make the not yet created theory somewhat more likely I think, but could well be either
15:20b4nhey
15:20b4nanything I shuold do to push https://bugzilla.mozilla.org/show_bug.cgi?id=1308908 forward?
15:20firebotBug 1308908 NEW, nobody@mozilla.org Failure to update accessible name when last unread mail is marked as read
15:23b4ntbsaunde maybe? :)
15:24tbsaunde302 davidb
15:26alexarnaudMarcoZ: Hey, how are you? I'm working with texou and b4n at Hypra. We work hardly to make free software and particularly Debian GNU/Linux, Compiz, Mate, LibreOffice and Firefox.
15:29alexarnaudMarcoZ: At this time, b4n works to increase accessibility of the Thunderbird calendar.
15:29alexarnaudtbsaunde: Who decide of the priority about a11y stuff?
15:31MarcoZalexarnaud: The TB Calendar is no longer part of the Mozilla corporation's objectives AFAIK. So you'll have to talk to the Calendar team community members about this.
15:31alexarnaudThere is a huge bug reporting on the Firefox BTS for a long time but I don't understand why no one fix it at this time.
15:31alexarnaudhttps://bugzilla.mozilla.org/show_bug.cgi?id=1170242
15:31firebotBug 1170242 NEW, nobody@mozilla.org Setting the caret via AtkText should scroll content into view
15:31alexarnaudMarcoZ: I just tell you our involvement for the first contact :).
15:32davidbb4n, alexarnaud: our top priority right now is to not break accessibility catastrophically for all users of multiprocess Firefox.
15:32MarcoZalexarnaud: For the bug you cite above, tbsaunde or surkov are actually the engineers to talk to here.
15:32alexarnaudMarcoZ: I want to know what is the process for the bug fixing? Who decide the priority and why huge bugs (in my opinion) couldn't be skipped after years.
15:33davidbalexarnaud: is bug 1170242 a regression? It would be great to know when it regressed.
15:33b4ndavidb: good thing I guess :)
15:33alexarnaudIt's not a reproach, just I want to understand for a better collaboration with you.
15:34alexarnaudI ofter use Firefox nightly on GNU/Linux :).
15:34alexarnaud*often
15:34davidbalexarnaud: I'm probably a good person to ask about priorities.
15:35alexarnauddavidb: I don't know if the bug I mentioned is a regression but it's a bug that affect both blinds and visual-impaired so the people that use a screen reader and screen magnifier.
15:36davidbIt sounds like a regression, because I thought we used to scroll to the caret.
15:37davidbI wonder if APZ introduced a bug
15:37davidbSee also: https://mzl.la/2kKZZBR
15:37alexarnauddavidb: We would like to have it fixed or to have a help for fixing it.
15:39alexarnaudDo you plan to fixing it? As I can see it also affect sighted people, isn't it?
15:39davidbIf we don't close the bug, we probably want to fix it.
15:39davidb(we have 1578 a11y bugs open)
15:40alexarnaudI mean is it schedule or in a TODO list of someone ^^.
15:40* davidb tries to remember what the layout.accessiblecaret.always_show_when_scrolling config does
15:40alexarnauddavidb: Oh, it's huge. How many people works in this area?
15:40davidbalexarnaud: not enough ;)
15:40alexarnauddavidb: I know, unfortunatly..
15:41alexarnaudAt this time I only help you on QA testing on GNU/Linux :).
15:41davidbappreciated
15:42alexarnauddavidb: If you or someone could add technical information about the bug we could imagine to work on it because it's a high priority for us.
15:43davidbalexarnaud: I would always ask if there is a regression window. That way if we know what broke it, we may be able to assign the bug outside our team.
15:43alexarnaudBut its depends of the time b4n spends on the calenda accessibility in TB.
15:44alexarnaudIs there any bisect process in Firefox?
15:44davidbhttp://mozilla.github.io/mozregression/quickstart.html
15:44alexarnaudThanks.
15:44davidbyw and ty
15:45alexarnaudso if I understand well for you the bug is not related to a11y?
15:45alexarnaudThanks a lot for your help :) !
15:46davidba11y relies on non-a11y code that can break
15:46alexarnaudI understand that.
15:46alexarnauddavidb: could you help b4n and show him the right way?
15:46* davidb nods
15:47davidbI'm not the best person...
15:47alexarnauddavidb: Who is the best for his question ?
15:47davidbfor mozregression, maybe wlach
15:48davidbor parkouss
15:51alexarnaudtbsaunde: It looks like b4n fix a code you have worked on : https://bugzilla.mozilla.org/show_bug.cgi?id=1308908
15:51firebotBug 1308908 NEW, nobody@mozilla.org Failure to update accessible name when last unread mail is marked as read
15:52alexarnaudWe hope it could be integrated in the next TB release because I assume it's a minor fix.
15:52alexarnaudThanks again for your help david :) !
16:00tbsaundealexarnaud: flag me for review and I'll try and take a look
16:36b4ntbsaunde: hum, how should I do that?
16:37tbsaundeb4n: open up details for the attachment and set r? flag to tbsaunde
16:40b4ntbsaunde: okay thanks, should be done
17:50firebotNew Core - Disability Access APIs bug 1339128 filed by tbsaunde+mozbugs@tbsaunde.org.
17:50firebothttps://bugzil.la/1339128 NEW, nobody@mozilla.org reference parent proxy by id
19:03firebotNew Core - Disability Access APIs bug 1339156 filed by tbsaunde+mozbugs@tbsaunde.org.
19:03firebothttps://bugzil.la/1339156 NEW, nobody@mozilla.org add an assert to RemoveChildDoc()
19:57firebotNew Core - Disability Access APIs bug 1339168 filed by jmathies@mozilla.com.
19:57firebothttps://bugzil.la/1339168 NEW, nobody@mozilla.org Crash in mozilla::a11y::DocAccessibleParent::RemoveChildDoc
21:53davidbtbsaunde: I guess your hash table stuff can go on bug 1336770 which is our highest priority now
21:53firebothttps://bugzil.la/1336770 NEW, tbsaunde+mozbugs@tbsaunde.org Crash in nsTArray_Impl<T>::IndexOf<T> | mozilla::a11y::ProxyAccessibleBase<T>::ClearChildDoc
21:56tbsaundedavidb: well, part of it is already landed in bug 1339128
21:56firebothttps://bugzil.la/1339128 NEW, nobody@mozilla.org reference parent proxy by id
21:56tbsaundeI guess I can use that one for the rest or just file another
21:59davidbyeah
14 Feb 2017
No messages
   
Last message: 163 days and 17 hours ago