/irc-logs / mozilla / #accessibility / 2015-03-09 / end
Options:
Previous day, Next day
- # Session Start: Mon Mar 09 00:00:00 2015
- # Session Ident: #accessibility
- # [00:28] <@firebot> New Core - Disability Access APIs bug 1140917 filed by bugs@pettay.fi.
- # [00:28] <@firebot> https://bugzil.la/1140917 — NEW, bugs@pettay.fi — IPC Proxy for replace/insert/copy/cut/delete/paste
- # [01:17] * Quits: newtron (newtron@moz-ieu4n6.dsl.teksavvy.com) (Connection closed)
- # [02:17] * Joins: yliao (yliao@moz-aj9ejj.HINET-IP.hinet.net)
- # [02:21] * Quits: yliao (yliao@moz-aj9ejj.HINET-IP.hinet.net) (Ping timeout: 121 seconds)
- # [02:43] * Joins: yliao (yliao@moz-2hf62a.HINET-IP.hinet.net)
- # [02:43] * Quits: smaug (chatzilla@moz-7dv8tf.bb.dnainternet.fi) (Connection closed)
- # [02:44] * Joins: smaug (chatzilla@moz-7dv8tf.bb.dnainternet.fi)
- # [02:49] * Quits: smaug (chatzilla@moz-7dv8tf.bb.dnainternet.fi) (Connection closed)
- # [02:50] * Joins: smaug (chatzilla@moz-7dv8tf.bb.dnainternet.fi)
- # [03:24] * Quits: smaug (chatzilla@moz-7dv8tf.bb.dnainternet.fi) (Ping timeout: 121 seconds)
- # [06:00] * Joins: icaaq (icaaq@moz-4a4bca.cust.bredbandsbolaget.se)
- # [08:12] * Tomcat|afk is now known as Tomcat
- # [08:21] * Tomcat is now known as Tomcat|sheriffduty
- # [08:28] * Joins: ioanachiorean (ioanachiore@moz-ohb.65e.79.82.IP)
- # [09:37] * Quits: icaaq (icaaq@moz-4a4bca.cust.bredbandsbolaget.se) (Quit: 10-79 Notify coroner, over and out)
- # [09:42] * Quits: ioanachiorean (ioanachiore@moz-ohb.65e.79.82.IP) (Connection closed)
- # [09:43] * Joins: ioanachiorean (ioanachiore@moz-ohb.65e.79.82.IP)
- # [09:49] * Joins: icaaq (icaaq@moz-t231nb.cust.bredbandsbolaget.se)
- # [09:49] * icaaq is now known as icaaq|afk
- # [09:50] * icaaq|afk is now known as icaaq
- # [10:24] * Joins: agibson (agibson@moz-j04gi9.cable.virginm.net)
- # [11:36] * Quits: yliao (yliao@moz-2hf62a.HINET-IP.hinet.net) (Connection closed)
- # [12:03] * Joins: yliao (yliao@moz-aj9ejj.HINET-IP.hinet.net)
- # [12:04] * Quits: yliao (yliao@moz-aj9ejj.HINET-IP.hinet.net) (Connection closed)
- # [12:17] * Joins: Gijs (chatzilla@moz-be1.e16.80.5.IP)
- # [12:56] * agibson is now known as agibson|afk
- # [13:03] * Joins: yliao (yliao@moz-aj9ejj.HINET-IP.hinet.net)
- # [13:08] * Quits: yliao (yliao@moz-aj9ejj.HINET-IP.hinet.net) (Ping timeout: 121 seconds)
- # [13:17] * Joins: smaug (chatzilla@moz-7dv8tf.bb.dnainternet.fi)
- # [13:58] * Joins: yzen (yzen@moz-ql1r6k.cpe.pppoe.ca)
- # [13:58] * ChanServ sets mode: +o yzen
- # [14:04] * agibson|afk is now known as agibson
- # [14:07] * Quits: scott_gonzalez (scott_gonza@moz-5t7jkp.fios.verizon.net) (Client exited)
- # [14:21] * Joins: davidb (davidb@moz-i5m.05u.207.66.IP)
- # [14:21] * ChanServ sets mode: +qo davidb davidb
- # [14:22] * Joins: anvk (anovak@moz-871.mvk.11.204.IP)
- # [14:24] * Joins: scott_gonzalez (scott_gonza@moz-mhe.gsn.149.12.IP)
- # [14:33] * Joins: newtron (newtron@moz-6lh.c31.71.199.IP)
- # [14:34] * Quits: newtron (newtron@moz-6lh.c31.71.199.IP) (Quit: Leaving...)
- # [14:35] * Joins: newtron (newtron@moz-6lh.c31.71.199.IP)
- # [14:46] * Gijs is now known as Gijs_away
- # [14:51] * davidb is now known as davidb|afk
- # [14:54] * davidb|afk is now known as davidb
- # [14:55] * Joins: clown (clown@moz-v6m.c3o.211.205.IP)
- # [15:02] * icaaq is now known as icaaq|afk
- # [15:11] * davidb is now known as davidb|afk
- # [15:17] * Joins: yliao (yliao@moz-aj9ejj.HINET-IP.hinet.net)
- # [15:21] * Quits: yliao (yliao@moz-aj9ejj.HINET-IP.hinet.net) (Ping timeout: 121 seconds)
- # [15:25] * davidb|afk is now known as davidb
- # [15:25] * Quits: smaug (chatzilla@moz-7dv8tf.bb.dnainternet.fi) (Ping timeout: 121 seconds)
- # [15:41] * Joins: SteveF (chatzilla@moz-80j6qs.cable.virginm.net)
- # [16:08] * davidb is now known as davidb|afk
- # [16:11] * davidb|afk is now known as davidb
- # [16:30] * Joins: scottgonzalez (scott_gonza@moz-mhe.gsn.149.12.IP)
- # [16:33] * Joins: wlach (wlach@moz-43h9ue.members.linode.com)
- # [16:33] * Quits: scott_gonzalez (scott_gonza@moz-mhe.gsn.149.12.IP) (Ping timeout: 121 seconds)
- # [16:34] <wlach> hey, can anyone give feedback on how usable treeherder is vs. tbpl with a screen reader? https://bugzilla.mozilla.org/show_bug.cgi?id=1053279#c13
- # [16:34] <@firebot> Bug 1053279 — NEW, nobody@mozilla.org — Make the treeherder front-end more accessible
- # [16:34] <wlach> davidb: tbsaunde: ^^^
- # [16:34] <wlach> I know the current situation isn't great, but the question is whether there are serious regression vs. tbpl (if tbpl was ever usable?)
- # [16:35] <~davidb> wlach, tbsaunde is probably better able to answer that, and i recommend you needinfo :marcoz in the bug (he's PTO for a bit)
- # [16:36] <wlach> davidb: who's on pto? marcoz or tbsaunde?
- # [16:36] <~davidb> marcoz
- # [16:36] * Joins: smaug (chatzilla@moz-7dv8tf.bb.dnainternet.fi)
- # [16:38] <wlach> davidb: should I needinfo both tbsaunde and marcoz?
- # [16:39] <~davidb> wlach, marcoz is enough unless you are rushed for an answer
- # [16:41] <wlach> davidb: it's only urgent in the sense that the plan is to turn tbpl off at the end of the month (see dev.tree-management)
- # [16:41] <~davidb> ah i see
- # [16:41] <~davidb> well marco can chime in monday
- # [16:41] <wlach> ok, sounds great
- # [16:43] <@tbsaunde> wlach: I have a least one bug I need to get around to filing, I'll try to get that done
- # [16:45] <wlach> tbsaunde: cool, feel free to CC me (:wlach) on it
- # [17:01] * Tomcat|sheriffduty is now known as Tomcat|afk
- # [17:04] * Quits: scottgonzalez (scott_gonza@moz-mhe.gsn.149.12.IP) (Ping timeout: 121 seconds)
- # [17:05] * davidb is now known as davidb|afk
- # [17:06] * icaaq|afk is now known as icaaq
- # [17:13] <@tbsaunde> ia2 deprecated editable text right? did atk do the same?
- # [17:14] * Gijs_away is now known as Gijs
- # [17:14] <@tbsaunde> joanie: ?
- # [17:15] <joanie> yes?
- # [17:15] <joanie> oh, sorry, scrollback
- # [17:16] <joanie> we didn't deprecate editable text
- # [17:16] <@tbsaunde> joanie: do its operations need to be sync?
- # [17:16] <@tbsaunde> that si does the paste need to complete before the function returns?
- # [17:16] <joanie> to be honest, Orca doesn't do anything with it, so I don't know
- # [17:17] <joanie> but we didn't deprecate it
- # [17:17] <@tbsaunde> ok, lets see what api says then
- # [17:17] <joanie> shall I get him?
- # [17:17] <joanie> in fact, lemme do that
- # [17:17] <joanie> long story
- # [17:17] <@tbsaunde> joanie: please
- # [17:18] * wlach is now known as wlach|biab
- # [17:18] <joanie> but I'm trying to "hide" in #a11y
- # [17:18] <@tbsaunde> heh, sure
- # [17:19] <joanie> I pinged him, but he's not pinged back yet
- # [17:19] <joanie> and I pastebinned the scrollback
- # [17:19] <joanie> whilst we wait, I have a question for you:
- # [17:20] <@tbsaunde> I may have an answer
- # [17:20] <joanie> Is get text for sentence boundaries unimplemented, or just completely broken?
- # [17:20] <joanie> I thought you had implemented it at some point
- # [17:20] <joanie> but I'm getting start and end offsets of 0 and an empty string
- # [17:20] <joanie> for everything
- # [17:22] <@tbsaunde> joanie: pretty sure its unimplemented
- # [17:22] <joanie> if I file a bug will y'all implement it?
- # [17:22] <joanie> be honest
- # [17:23] * @tbsaunde tries to see if there's a existing sentence edge finding thing
- # [17:24] <joanie> thanks
- # [17:26] <@tbsaunde> joanie: seems like there's no existing stuff to break by sentence, so probably not on the grounds its not clear how to find boundaries
- # [17:26] <joanie> ok
- # [17:26] <joanie> I may file a bug anyway just for documentation's sake
- # [17:26] <@tbsaunde> I'm no i18n expert, but I kind of wonder how you'd implement it
- # [17:27] <@tbsaunde> joanie: I wonder if there isn't one, but documenting you want it can't hurt
- # [17:27] <joanie> dunno. But WebKit and Gtk at least have done so
- # [17:27] * Joins: infapi00 (API@moz-lnbnpf.igalia.com)
- # [17:27] <@tbsaunde> interesting maybe when surkov is board he can look at how they did it
- # [17:27] <joanie> fearless leader!
- # [17:28] <joanie> I'll shut up for now so infapi00 can get caught up and then answer your question
- # [17:28] <infapi00> tbsaunde, so short answer
- # [17:29] <infapi00> no, we didn't deprecate editable text
- # [17:29] <infapi00> what we did was deprecate some atktext methods, and add some newsssssssss
- # [17:29] <infapi00> *new methods
- # [17:29] <infapi00> in order to simplify atktext API
- # [17:29] <infapi00> so not sure if you thought that the deprecation was about editable text
- # [17:29] <@tbsaunde> infapi00: so for now I only care about the stuff that modifies text
- # [17:30] <infapi00> tbsaunde, ok
- # [17:30] <infapi00> so as I said, no deprecation on editable text
- # [17:30] <@tbsaunde> infapi00: no, ia2 did deprecate their version iirc so I thought atk might have too
- # [17:30] <@tbsaunde> infapi00: does the editing need to be sync?
- # [17:30] <joanie> i'm surprised they did
- # [17:31] <infapi00> tbsaunde, so all the editable text interface got deprecated? or just some methods?
- # [17:31] <joanie> while a screen reader doesn't need it, other tools (e.g. alternative input devices and software) would
- # [17:32] <@tbsaunde> infapi00: would need to check, but maybe just methods
- # [17:32] <infapi00> tbsaunde, so what are the replacement for those methods?
- # [17:32] <@tbsaunde> infapi00: but really the question I'm trying to answer is can I make the editing happen after the function returns and not break people
- # [17:34] <infapi00> tbsaunde, so for example
- # [17:34] <infapi00> implement atk_editable_text_insert_text
- # [17:34] * Joins: scott_gonzalez (scott_gonza@moz-6k5.198.244.168.IP)
- # [17:34] <infapi00> and the insertion happens "little after"
- # [17:34] <infapi00> the method is called?
- # [17:35] <@tbsaunde> infapi00: yes
- # [17:35] <infapi00> hmmm
- # [17:35] <infapi00> well, the interface itself doesn't specify that detail
- # [17:35] <infapi00> is a kind of corner case
- # [17:35] <infapi00> in any case
- # [17:35] <infapi00> so that means that for example
- # [17:35] <infapi00> something like
- # [17:35] <infapi00> atk_editable_text_insert_text (blah); atk_text_get_text(blah2)
- # [17:36] <infapi00> would result on get_text getting a text that doesn't include the insertion?
- # [17:37] <@tbsaunde> infapi00: yes, that would be possible
- # [17:37] <infapi00> so the only way to ensure that get_text gets the proper result, is connecting to the text changed events?
- # [17:37] <@tbsaunde> on the other hand if you do the insertion first then presumably you fire text change events from within atk_editable_text_insert
- # [17:37] <@tbsaunde> infapi00: yeah
- # [17:37] <joanie> how would an AT know that atk_editable_text_cut_text() had completed?
- # [17:38] <joanie> ah, the text-changed event again
- # [17:38] <joanie> yuck
- # [17:38] <infapi00> hmm, well
- # [17:38] <infapi00> as I said, this seems like a corner case
- # [17:38] <@tbsaunde> or you have a period of time where the text has changed, but the text change event hasn't been fired
- # [17:38] <infapi00> in the sense that it is not specified
- # [17:38] <joanie> cut text isn't a corner case infapi00
- # [17:38] <infapi00> no
- # [17:39] <infapi00> I mean
- # [17:39] <infapi00> that the corner case
- # [17:39] <joanie> fortunately Orca doesn't need it
- # [17:39] <infapi00> is about the need to have the change available as soon as the method ends
- # [17:39] <infapi00> most people would understand that
- # [17:39] <infapi00> but it is not specified on the interface
- # [17:39] <infapi00> probably corner case is not the proper words
- # [17:40] <joanie> I think that if I were implementing, say, a speech input solution
- # [17:40] <infapi00> but a "hole on the specification" or something
- # [17:40] <joanie> having to listen for the text change events
- # [17:40] <joanie> and wait
- # [17:40] <joanie> could be kinda sucky
- # [17:40] <joanie> and like you say, it's not in the documentation
- # [17:41] <infapi00> well, I guess that the final point
- # [17:41] <infapi00> is that when you call atk_text_get_text
- # [17:41] <@tbsaunde> joanie: so, the thing is the alternatives are also weird, either the event comes fter the change, or the event happens while your in cuttext
- # [17:41] <infapi00> it says that you get the current text
- # [17:42] <infapi00> so if we were literal with that definition
- # [17:42] <infapi00> then this call the method and wait is wrong
- # [17:42] <joanie> well, if you're literal with that definition, the current text lacks the inserted text ;)
- # [17:43] <infapi00> tbsaunde, so this suggested change are raised by doubts about when to emit the events?
- # [17:43] <joanie> because the editable-text docs don't specify if it's sync or not
- # [17:43] <@tbsaunde> infapi00: no, the problem is having to wait for the operation to complete before returning from the cut_text impl
- # [17:44] <@tbsaunde> I believe there's a valid question about events too, but I don't care about it right now
- # [17:45] <infapi00> well, I don't have a problem to avoid waits
- # [17:45] <infapi00> the problematic part is possible errors on following calls to get_text
- # [17:47] <@tbsaunde> if you interpret the spec to allow you to return before the operation completes then I don't think the errors are a problem, the current text is what it is, and the operation hasn't completed so it hasn't changed
- # [17:49] <joanie> by spec, are we talking about docs?
- # [17:49] <joanie> or is this part of a spec-spec?
- # [17:50] <infapi00> tbsaunde, but just a question:
- # [17:50] <infapi00> on the spec it says:
- # [17:50] <infapi00> https://developer.gnome.org/atk/unstable/AtkEditableText.html#atk-editable-text-insert-text
- # [17:50] <infapi00> position
- # [17:50] <infapi00>
- # [17:50] <infapi00> The caller initializes this to the position at which to insert the text. After the call it points at the position after the newly inserted text.
- # [17:50] <infapi00>
- # [17:50] <infapi00> so if the method returns without the text being changed
- # [17:50] <infapi00> what would be the value returned on position?
- # [17:50] <infapi00> joanie, well, yes, at least I, when I said spec, I was talking about the doc
- # [17:51] * davidb|afk is now known as davidb
- # [17:55] <joanie> sooooo.... That suggests to me it's sync -- and documented as such. Though perhaps not quite loudly enough.
- # [17:55] <joanie> Right?
- # [17:58] * Joins: Zunino (zunino@moz-7of.uh2.58.187.IP)
- # [18:01] * wlach|biab is now known as wlach
- # [18:04] * infapi00 wonders if tbsaunde is still around or not
- # [18:06] * Quits: ~davidb (davidb@moz-i5m.05u.207.66.IP) (Connection closed)
- # [18:07] <@tbsaunde> infapi00: back now
- # [18:07] <@tbsaunde> joanie: infapi00 I have to agree, it seems like that's badly doc'd as sync
- # [18:08] * Joins: davidb (davidb@moz-i5m.05u.207.66.IP)
- # [18:08] * ChanServ sets mode: +qo davidb davidb
- # [18:09] <joanie> tbsaunde: does that mean case closed?
- # [18:09] <joanie> I think infapi00 was snatched away from something
- # [18:09] <@tbsaunde> joanie: it means I'm a very sad panda
- # [18:09] <joanie> and I want to talk about dead accessibles
- # [18:10] <joanie> tbsaunde: I'm curious as to what brought up the question. Is there a bug you're fixing, code you're refactoring, something else?
- # [18:11] <~davidb> e10s
- # [18:11] <~davidb> (multi process FF)
- # [18:11] <@tbsaunde> joanie: I'm looking at making editable text work with multiple processes
- # [18:12] <@tbsaunde> and I really wanted to avoid blocking a process while teh text editing happens
- # [18:13] <joanie> well, right now I don't think there are any consumers of that API
- # [18:13] <joanie> so I'm not sure what the right thing to do is
- # [18:13] <joanie> if someone is (or plans to) rely upon that API then obviously it needs to be done correctly
- # [18:14] <joanie> by the same token, there are plenty of other accessibility-related bugs that impact Orca ;)
- # [18:14] <joanie> and I myself would prefer you work on those ;) ;)
- # [18:14] <joanie> and speaking of the multiple process stuff -- is there any chance that is why I've been seeing an increase in dead accessibles lately?
- # [18:15] <@tbsaunde> I doubt it, multiple process stuff should only effect what happens if you have multiple processes
- # [18:15] <joanie> heh
- # [18:16] * Quits: anvk (anovak@moz-871.mvk.11.204.IP) (Quit: Leaving.)
- # [18:16] <joanie> all I know is that we've got dead accessibles all over the place with the latest firefox
- # [18:16] <joanie> not sure if that's a coincidence (web authors doing something)
- # [18:16] <@tbsaunde> latest == release or nightly?
- # [18:16] <joanie> release
- # [18:17] * Joins: surkov (surkov@moz-v8g78c.cpe.teksavvy.com)
- # [18:17] * ChanServ sets mode: +o surkov
- # [18:17] <joanie> i've even seen some children-changed events where the event source is already a dead accessible by the time we process the event
- # [18:18] <joanie> children-changed *add* events
- # [18:18] <joanie> I've similarly seen children-changed:add events where the added child is a dead accessible
- # [18:19] <@tbsaunde> joanie: hm, its hard to say something without a test case
- # [18:19] <joanie> I know
- # [18:19] <joanie> that's on my to-do list for today or tomorrow
- # [18:20] <@tbsaunde> and I don't have much time for stuff that isn't multiple process stuff :(
- # [18:20] <joanie> I'll bother surkov then :)
- # [18:20] <joanie> hey, surkov is here. I didn't notice.
- # [18:20] <@tbsaunde> I really hope he helps out
- # [18:20] <@surkov> hey
- # [18:20] <@surkov> what’s up?
- # [18:20] <joanie> Can I have a sentence boundary AtkText implementation please?
- # [18:20] <joanie> :)
- # [18:21] <smaug> sounds like "regression-window-wanted"
- # [18:21] <joanie> all the cool kids do it
- # [18:21] <smaug> s/window/range/
- # [18:21] <joanie> smaug: the problem is that it could be a coincidence
- # [18:21] <joanie> web devs do things that kill accessibles -- or rather cause gecko to kill accessibles
- # [18:21] <joanie> so maybe they're just doing that thing more
- # [18:21] <@surkov> joanie, technically yes, I’m not sure how fast we could do this, that’s something that requires some investigations
- # [18:22] <joanie> surkov: understood. Filing the bug now then :)
- # [18:22] <@surkov> ok :)
- # [18:25] <@firebot> New Core - Disability Access APIs bug 1141181 filed by jdiggs@igalia.com.
- # [18:25] <@firebot> https://bugzil.la/1141181 — NEW, nobody@mozilla.org — Implement support for the sentence boundary/granularity in AtkText.
- # [18:29] * agibson is now known as agibson|afk
- # [18:30] * Joins: jongund (chatzilla@moz-emft4q.dres.uiuc.edu)
- # [18:37] * agibson|afk is now known as agibson
- # [18:41] * Quits: ~davidb (davidb@moz-i5m.05u.207.66.IP) (Connection closed)
- # [18:42] <Zunino> Hi. First time here, so I'd like to know whether it's adequate to ask questions about ARIA usage on this channel.
- # [18:43] * Quits: jongund (chatzilla@moz-emft4q.dres.uiuc.edu) (Quit: ChatZilla 0.9.91.1 [Firefox 36.0/20150222232811])
- # [18:43] * Joins: davidb (davidb@moz-i5m.05u.207.66.IP)
- # [18:43] * ChanServ sets mode: +qo davidb davidb
- # [18:44] * agibson is now known as agibson|afk
- # [18:58] * Quits: Gijs (chatzilla@moz-be1.e16.80.5.IP) (Quit: ChatZilla 0.9.91-rdmsoft [XULRunner 32.0.3/20140923175406])
- # [18:58] * Joins: Gijs (chatzilla@moz-be1.e16.80.5.IP)
- # [19:05] * Quits: SteveF (chatzilla@moz-80j6qs.cable.virginm.net) (Ping timeout: 121 seconds)
- # [19:11] * Joins: scottgonzalez (scott_gonza@moz-mhe.gsn.149.12.IP)
- # [19:14] * Joins: SteveF (chatzilla@moz-80j6qs.cable.virginm.net)
- # [19:14] * Quits: scott_gonzalez (scott_gonza@moz-6k5.198.244.168.IP) (Ping timeout: 121 seconds)
- # [19:17] * Joins: yliao (yliao@moz-aj9ejj.HINET-IP.hinet.net)
- # [19:21] * Quits: yliao (yliao@moz-aj9ejj.HINET-IP.hinet.net) (Ping timeout: 121 seconds)
- # [19:30] * Joins: anvk (anovak@moz-871.mvk.11.204.IP)
- # [19:34] <smaug> tbsaunde: if you you have any good suggestions what I should do next, just tell. Otherwise I'll add a proxy for images and links and after those perhaps start looking tables and cells
- # [19:36] * Joins: yzen_ (yzen@moz-ql1r6k.cpe.pppoe.ca)
- # [19:36] * ChanServ sets mode: +o yzen_
- # [19:37] * Quits: @yzen (yzen@moz-ql1r6k.cpe.pppoe.ca) (Connection closed)
- # [19:37] * agibson|afk is now known as agibson
- # [19:40] * icaaq is now known as icaaq|afk
- # [19:40] <@tbsaunde> smaug: sgtm
- # [19:40] <@tbsaunde> smaug: I'm working on making windows stuff
- # [19:41] <@tbsaunde> I've also been doing a bit of pto for family stuff this past week
- # [19:43] * agibson is now known as agibson|afk
- # [19:57] * Quits: SteveF (chatzilla@moz-80j6qs.cable.virginm.net) (Ping timeout: 121 seconds)
- # [20:02] * agibson|afk is now known as agibson
- # [20:03] <@firebot> New Core - Disability Access APIs bug 1141220 filed by jdiggs@igalia.com.
- # [20:03] <@firebot> https://bugzil.la/1141220 — NEW, nobody@mozilla.org — Incredibly long load time and "dead" accessible objects seen with certain sites
- # [20:03] <joanie> tbsaunde: surkov: ^^ Your test case re dead accessibles
- # [20:04] <joanie> I'll see if I can pin down when this started because it didn't used to happen
- # [20:04] <@surkov> looking
- # [20:04] <joanie> but if you could take a look in the meantime, it would be quite helpful
- # [20:06] <@surkov> joanie: can you share perf profile (https://developer.mozilla.org/en-US/docs/Tools/Performance)
- # [20:07] <joanie> surkov: lemme read and then yeah
- # [20:07] <@surkov> thank you
- # [20:07] <joanie> thank you for looking at it
- # [20:07] <joanie> it's really bad :(
- # [20:07] <joanie> and of course "orca's fault" :P
- # [20:12] <joanie> hit the attachment size limit
- # [20:12] <~davidb> joanie, you can upload profiles and just provide the link in the bug
- # [20:13] <joanie> i'm tarballing it
- # [20:13] <joanie> surkov: attached
- # [20:13] <~davidb> https://developer.mozilla.org/en-US/docs/Mozilla/Performance/Profiling_with_the_Built-in_Profiler
- # [20:13] <joanie> I used that
- # [20:13] <joanie> I saved it
- # [20:13] <~davidb> k
- # [20:13] <@surkov> thanks :)
- # [20:14] <@surkov> sorry for wrong link
- # [20:14] * agibson is now known as agibson|afk
- # [20:14] <joanie> did I do it wrong?
- # [20:14] <~davidb> (there should be an upload button in the tool, and we host the file)
- # [20:14] <joanie> oh
- # [20:14] <joanie> oh well
- # [20:14] <joanie> surkov: do you have what you need?
- # [20:15] <@surkov> checking, I thought it’s supposed to be a link
- # [20:15] <joanie> meh
- # [20:15] <joanie> ok, lemme read
- # [20:15] <~davidb> (this wiki page could be a lot clearer)
- # [20:17] * Quits: infapi00 (API@moz-lnbnpf.igalia.com) (Quit: Leaving)
- # [20:17] <joanie> and the info is different
- # [20:19] <joanie> ok, lemme do this all over and get a nightly and bla bla bla :)
- # [20:19] <@surkov> strange, it looks empty for me
- # [20:19] <joanie> i'll do it using https://developer.mozilla.org/en-US/docs/Mozilla/Performance/Reporting_a_Performance_Problem
- # [20:20] <~davidb> joanie, oh yeah that's a way better howto
- # [20:20] * ~davidb bookmarks
- # [20:20] <joanie> it's linked on the page you provided davidb :)
- # [20:20] <~davidb> oh :)
- # [20:20] <joanie> but different from the instructions surkov provided
- # [20:20] <joanie> so I'll start over
- # [20:21] <~davidb> k
- # [20:21] <joanie> anything to get this fixed
- # [20:21] <joanie> I can't hack around it
- # [20:21] <joanie> as it's not in orca
- # [20:21] <@firebot> New Core - Disability Access APIs bug 1141228 filed by mhabicher@mozilla.com.
- # [20:21] <@firebot> https://bugzil.la/1141228 — NEW, nobody@mozilla.org — Include GfxMessageUtils.h in PDocAccessible.ipdl
- # [20:28] <joanie> tool says that the profle I am trying to upload is more than the 9 MBs storage maximum
- # [20:28] <joanie> davidb: ^^
- # [20:29] <joanie> this is using the share with URL stuff
- # [20:29] <~davidb> you can select a window (start end) of the profile that looks interesting (i think)
- # [20:29] * Quits: agibson|afk (agibson@moz-j04gi9.cable.virginm.net) (Quit: )
- # [20:30] <joanie> i dunno what looks interesting ;)
- # [20:30] * joanie looks
- # [20:30] <~davidb> if you peruse.... inverting the callstack can be interesting sometimes
- # [20:31] <~davidb> look for *accessible method calls
- # [20:31] <joanie> k
- # [20:32] <joanie> the "guilty" one is GetOrCreateAccessible
- # [20:32] * Quits: athena (sid21127@moz-ge8mr4.ealing.irccloud.com) (Ping timeout: 121 seconds)
- # [20:33] * ~davidb nods
- # [20:33] <joanie> if I provide the full file maybe this time it will work
- # [20:33] * joanie tries
- # [20:33] * Joins: athena (sid21127@moz-ge8mr4.ealing.irccloud.com)
- # [20:35] <joanie> surkov: can you see if what I just attached is what you need?
- # [20:38] * Joins: SteveF (chatzilla@moz-80j6qs.cable.virginm.net)
- # [20:39] <joanie> if it is, then I'll start going back through earlier releases to see if I can figure out when this problem got introduced
- # [20:42] <@firebot> mhabicher@mozilla.com changed the Assignee on bug 1141228 from nobody@mozilla.org to mhabicher@mozilla.com.
- # [20:42] <@firebot> mhabicher@mozilla.com changed the Status on bug 1141228 from NEW to ASSIGNED.
- # [20:43] <@firebot> https://bugzil.la/1141228 — ASSIGNED, mhabicher@mozilla.com — Include GfxMessageUtils.h in PDocAccessible.ipdl
- # [20:57] * Gijs is now known as Gijs_away
- # [21:06] * davidb is now known as davidb|afk
- # [21:07] <joanie> surkov: I'm seeing this particular issue with older versions of Firefox too (e.g. 30). So I guess it's not newish after all. :-/
- # [21:07] <joanie> anyhoo, please let me know (once you have a chance to look), if the profile I uploaded has what you need.
- # [21:09] * davidb|afk is now known as davidb
- # [21:11] * Quits: SteveF (chatzilla@moz-80j6qs.cable.virginm.net) (Ping timeout: 121 seconds)
- # [21:15] <@yzen_> eeejay hi
- # [21:25] <~davidb> you will create a scope of trust... what you share with your cohort etc... you don't need to share outside.
- # [21:25] <~davidb> that said, you can ask me anything about my outcomes.
- # [21:25] <~davidb> oh wrong channel but meh
- # [21:31] * Joins: SteveF (chatzilla@moz-80j6qs.cable.virginm.net)
- # [21:32] <smaug> tbsaunde: perhaps we really need to give up with prio(high) and then use async whenever possible
- # [21:38] * Parts: clown (clown@moz-v6m.c3o.211.205.IP)
- # [21:38] <smaug> or maybe not "To avoid jank, the parent process is not allowed to send sync messages of normal priority."
- # [21:42] <@firebot> fred.wang@free.fr changed the Status on bug 1128143 from NEW to ASSIGNED.
- # [21:42] <@firebot> https://bugzil.la/1128143 — ASSIGNED, fred.wang@free.fr — Add ATK roles for MathML fractions, roots
- # [21:47] * Quits: SteveF (chatzilla@moz-80j6qs.cable.virginm.net) (Ping timeout: 121 seconds)
- # [21:48] * Quits: anvk (anovak@moz-871.mvk.11.204.IP) (Quit: Leaving.)
- # [21:52] * Joins: fredw (Icedove@moz-7gtdsj.fbx.proxad.net)
- # [21:55] <@tbsaunde> smaug: honestly I just don't know
- # [21:56] <@tbsaunde> on the other hand nobody seems to complain about events being slightly late, maybe we can just cheat and nobody wil care
- # [22:11] * Quits: ~davidb (davidb@moz-i5m.05u.207.66.IP) (Quit: Blah blah blah)
- # [22:27] <eeejay> yzen_: hey. sorry
- # [22:33] * wlach is now known as wlach|bbl
- # [22:58] <eeejay> yzen_ you around?
- # [23:02] <@yzen_> eeejay hi yeah , re my earlier ping nvm
- # [23:02] <eeejay> yzen_: question: i want the browser to announce loading/loaded
- # [23:02] <eeejay> yzen_: is there a way to do that without actually displaying text?
- # [23:02] <@yzen_> eeejay yeah
- # [23:03] <@yzen_> eeejay are you talking about gaia or accessfu?
- # [23:03] <eeejay> yzen_: gaia
- # [23:04] <@yzen_> eeejay and that corresponds to a spinner in the url bar or something like that ?
- # [23:04] <eeejay> yzen_: yeah. there is a progress bar
- # [23:06] <eeejay> yzen_: i think we might have to add something in accessfu
- # [23:07] <@yzen_> eeejay ok, afaik, there are 2 approaches (but one has an outstanding bug for accessfu): 1) have an indeterminate progress with value text being what you want above (the outstanding accessfu bug is for treating subtree progressbar similar to a live region). another one is use role status for something that is labeled with things that you want above
- # [23:08] <eeejay> yzen_: hm.. interesting
- # [23:08] <eeejay> i was using progressbar, but i guess it could be status
- # [23:08] <@yzen_> I like the accessfu solution but that wont be in 2.2 without an uplift to 37
- # [23:09] <@yzen_> status works quite nice in some place that we use it already (like updating, last updated email fetch )
- # [23:09] <@yzen_> but yeah, browser progress looks like a progress
- # [23:10] <@yzen_> eeejay what happens if you put aria live on a progress, will it announce valueText?
- # [23:10] <eeejay> yzen_: it also gets hidden immediately on load complete, so i am afraid it won't get read. i'll try.
- # [23:11] <@yzen_> eeejay i would try then changing default actions that alia live announced to include hide?
- # [23:11] <@yzen_> s/alia/aria
- # [23:13] <eeejay> hm
- # [23:17] * Joins: yliao (yliao@moz-aj9ejj.HINET-IP.hinet.net)
- # [23:17] * Quits: newtron (newtron@moz-6lh.c31.71.199.IP) (Ping timeout: 121 seconds)
- # [23:21] * Quits: yliao (yliao@moz-aj9ejj.HINET-IP.hinet.net) (Ping timeout: 121 seconds)
- # [23:25] * Quits: smaug (chatzilla@moz-7dv8tf.bb.dnainternet.fi) (Ping timeout: 121 seconds)
- # [23:37] <joanie> surkov: so... any ideas about what to do regarding the slow tree bug?
- # [23:38] <joanie> 40 seconds is a really long time
- # [23:38] <joanie> and are the dead accessibles something at-spi2 is doing, you think?
- # [23:38] <joanie> e.g. due to a timeout maybe?
- # [23:40] <joanie> and another possible clue: The bug doesn't happen if you save the content locally and load it
- # [23:44] <@surkov> joanie: I’ll give it a try on crollsplatform layer
- # [23:44] <joanie> cool
- # [23:44] <joanie> thank you
- # [23:45] <joanie> I wonder if this happens on Windows
- # [23:45] * Gijs_away is now known as Gijs
- # [23:45] <joanie> This particular test case, for me, reliably tanks
- # [23:49] <fredw> surkov: I've tried to refresh Jonathan's MathML patches using your commit for bug 1001634. However, it seems that the code has changed a lot :-(
- # [23:49] <@firebot> https://bugzil.la/1001634 — NEW, surkov.alexander@gmail.com — Add MathMLAccessible class
- # [23:49] <@surkov> fredw: I’m looking it
- # [23:59] <@surkov> fredw: it looks good, land it?
- # Session Close: Tue Mar 10 00:00:00 2015
Previous day, Next day
Think these logs are useful? Then please donate to show your gratitude (and keep them up, of course). Thanks! — Krijn