/irc-logs / mozilla / #accessibility / 2012-04-05 / end
Options:
- # Session Start: Thu Apr 05 00:00:00 2012
- # Session Ident: #accessibility
- # [00:04] * Quits: margle (margle@moz-9542A19B.dsl.mweb.co.za) (Quit: Textual IRC Client: http://www.textualapp.com/)
- # [00:09] * Joins: davidb (davidb@471D72E.2257F909.F30C9E9E.IP)
- # [00:09] * ChanServ sets mode: +qo davidb davidb
- # [00:12] * Quits: @davidb (davidb@471D72E.2257F909.F30C9E9E.IP) (Quit: davidb)
- # [00:21] <@firebot> maxli@maxli.ca requested feedback from eitan@monotonous. org for attachment 612359 on bug 670853.
- # [00:21] <@firebot> Bug https://bugzilla.mozilla.org/show_bug.cgi?id=670853 nor, --, ---, maxli, NEW, undetermined progressmeters should expose mixed state
- # [01:19] * reuben is now known as Guest
- # [01:24] * Guest is now known as reuben
- # [01:25] <@firebot> akeybl@mozilla.com set status-firefox14 to fixed on bug 740958.
- # [01:25] <@firebot> Bug https://bugzilla.mozilla.org/show_bug.cgi?id=740958 cri, --, mozilla14, surkov.alexander, RESO FIXED, crash in nsAccessible::GetBounds
- # [01:36] * Quits: drexler (chatzilla@moz-6F22CAC7.e7jvsoe-gw.cust.sover.net) (Quit: ChatZilla 0.9.88-rdmsoft [XULRunner 1.9.0.17/2009122204])
- # [01:46] * Quits: @jprmc (jprmc@F2D29657.F60B0462.67AC9B1.IP) (Ping timeout)
- # [02:13] * Quits: @hub (hub@21B7B9F2.B87E9213.6E712CE2.IP) (Ping timeout)
- # [02:19] <@firebot> maxli@maxli.ca requested review from surkov.alexander@gm ail.com for attachment 612409 on bug 735645.
- # [02:19] <@firebot> Bug https://bugzilla.mozilla.org/show_bug.cgi?id=735645 nor, --, ---, maxli, NEW, expose sub and sup elements in text attributes
- # [02:20] <Mark_Capella> tbsaunde: thanks for the tip .... I had seen an earlier comment so i moved the method from protect to public and got a clean build (not tested yet)
- # [02:20] <Mark_Capella> is this a way to achieve the same results without changing the member declaration?
- # [02:22] <@tbsaunde> Mark_Capella: no
- # [02:22] <@tbsaunde> Mark_Capella: no
- # [02:24] <@tbsaunde> instead of if (foo->IsHTML()) static_cast<nsGenericHTMLElement*>(foo) you can do nsGenericHTMLElement* bar = nsGenericHTMLElement::FromContent(foo)
- # [02:24] <@tbsaunde> if (bar) return bar
- # [02:27] <Mark_Capella> so i can leave the member function GetURIAttr() protected / asis, and use this cast (which is public) instead yes?
- # [02:29] * Joins: jprmc (jprmc@78CFC4BC.5BCEC6DB.DA78B690.IP)
- # [02:29] * ChanServ sets mode: +o jprmc
- # [02:30] <@tbsaunde> Mark_Capella: no, casting to the right type has nothing to do with access to members
- # [02:30] <Mark_Capella> tbsaunde: assuming the answer to --^ is yes, ill try that after the clobber build im running for a different bug finishes and left you know
- # [02:30] <Mark_Capella> oh wait .... the cast is ok though in the code i thought
- # [02:32] <@tbsaunde> the cast is fine, nsGeneriHTMLElement::FromContent() > inline static cast
- # [02:37] * Quits: ehsan (ehsan@F2D29657.F60B0462.67AC9B1.IP) (Input/output error)
- # [02:39] <Mark_Capella> ok ... heres what i understand ... I need to move the GetURIAttr declaration from the protected area to the public area to make my build complete
- # [02:39] <Mark_Capella> (ive tried this and it works)
- # [02:39] <Mark_Capella> what I dont understand is what would I need to use the FromContent() member for?
- # [02:42] * Joins: davidb (davidb@471D72E.2257F909.F30C9E9E.IP)
- # [02:42] * ChanServ sets mode: +qo davidb davidb
- # [02:43] <Mark_Capella> tbsaunde: (and we may have been talking about two different casting situations just then)
- # [02:43] * Quits: jimm (jmathies@moz-7F164CA1.pn.at.cox.net) (Quit: )
- # [02:44] <@tbsaunde> Mark_Capella: in nsHTMLImageAccessible::LongDesc() you have nsIContent* foo; then bar = static_cast<nsGenericHTMLElement*>(foo); right?
- # [02:45] <Mark_Capella> nope - several iterations of code back
- # [02:45] <Mark_Capella> nsGenericHTMLElement* element =
- # [02:45] <Mark_Capella> static_cast<nsGenericHTMLElement*>(mContent.get());
- # [02:45] <Mark_Capella> then element->GetURIAttr(nsGkAtoms::longdesc, nsnull, getter_AddRefs(uri));
- # [02:46] <@tbsaunde> davidb: be happy you aren't responsible for the android tests, it seems about every last one of them failed on my try push =p
- # [02:46] * Joins: drexler (chatzilla@moz-BD26ECCD.hsd1.vt.comcast.net)
- # [02:47] <@tbsaunde> Mark_Capella: fine, the details aren't particularly interesting or important
- # [02:48] <Mark_Capella> ah ok - i thought you found a better way :)
- # [02:48] <@tbsaunde> somewhere in the patch you downcast from nsIContent to nsGenericHTMLEelement right?
- # [02:48] <@davidb> tbsaunde: heh
- # [02:48] * khuey is now known as khuey|away
- # [02:49] <@tbsaunde> Mark_Capella: and I've been saying instead of using static cast use nsGenericHTMLElement::FromContent()
- # [02:49] * Quits: @jprmc (jprmc@78CFC4BC.5BCEC6DB.DA78B690.IP) (Ping timeout)
- # [02:51] <Mark_Capella> and then to clarify, id still have to make GetURIAttr public yes?
- # [02:52] <@tbsaunde> yes
- # [02:52] <Mark_Capella> >whew< :P thanks tbsaunde
- # [02:52] <@tbsaunde> yw
- # [02:55] <@firebot> trev.saunders@gmail.com requested review from nfroyd@mozilla.com for attachment 612413 on bug 733510.
- # [02:55] <@firebot> Bug https://bugzilla.mozilla.org/show_bug.cgi?id=733510 nor, --, ---, trev.saunders, NEW, Transition boolean a11y telemetry to 'flag' telemetry as appropriate
- # [03:00] <@davidb> tbsaunde: I think the r+ carries forward for that one
- # [03:01] <@davidb> but leave it if you wish
- # [03:05] * Quits: drexler (chatzilla@moz-BD26ECCD.hsd1.vt.comcast.net) (Quit: ChatZilla 0.9.88-rdmsoft [XULRunner 1.9.0.17/2009122204])
- # [03:05] <@tbsaunde> davidb: hm, thought he didn't r+
- # [03:05] * @tbsaunde looks
- # [03:07] <@tbsaunde> davidb: yeah, guess so I guess I'll cancel and land
- # [03:07] <@davidb> yeah it looks right to me
- # [03:07] <@davidb> then i'll update my unknown consumer patch
- # [03:13] * Joins: ehsan (ehsan@F0B20A8D.8458880F.57F33CED.IP)
- # [03:14] <@firebot> trev.saunders@gmail.com cancelled review?(nfroyd@mozilla.com) for attachment 612413 on bug 733510.
- # [03:14] <@firebot> Bug https://bugzilla.mozilla.org/show_bug.cgi?id=733510 nor, --, ---, trev.saunders, NEW, Transition boolean a11y telemetry to 'flag' telemetry as appropriate
- # [03:26] <@davidb> tbsaunde: what is the weak reference bug aiming to fix exactly?
- # [03:27] <@tbsaunde> davidb: don't keep accessibles alive when they want to die
- # [03:27] <@davidb> but the weak accessible lives
- # [03:27] <@tbsaunde> and don't let ats keep gecko stuff alive
- # [03:27] <@davidb> admittedly it is smaller
- # [03:27] <@davidb> i see
- # [03:28] <@davidb> we should say this in the bug somewhere
- # [03:28] <@tbsaunde> what do you mean eak accessible?
- # [03:28] <@tbsaunde> I guess it doesn't hurt but who does it help?
- # [03:28] <@davidb> nsWeakAccessible is part of this picture right?
- # [03:29] <@davidb> it helps readers of the bug :)
- # [03:30] <@tbsaunde> davidb: no, never heard of nsWeakAccessible before
- # [03:30] <@davidb> it doesn't exist
- # [03:30] * khuey|away is now known as khuey
- # [03:30] <@davidb> I'm probably misunderstanding the bug
- # [03:30] <@tbsaunde> I guess, though I don't expect that many unfamiliar with a11y impl are going to go reading bugs for fun
- # [03:30] <@davidb> I was thinking this would be a setup like layout has for nsWeakFrame
- # [03:31] <@tbsaunde> probably something a like, but I'm not really sure how weak frames work
- # [03:31] <@davidb> well
- # [03:32] <@davidb> something like this
- # [03:32] <@tbsaunde> the purpose is similar, but I'm not sure we want to make the same trade offs they did
- # [03:32] <@davidb> here's a related change se http://hg.mozilla.org/mozilla-central/rev/f360de4b8b08
- # [03:32] <@davidb> set
- # [03:33] <@davidb> sample code
- # [03:33] <@davidb> nsWeakAccessible weakAccessible;
- # [03:33] <@davidb> weakAccessible = GetAccessible(foo);
- # [03:33] <@davidb> do stuff
- # [03:33] <@davidb> later on...
- # [03:34] <@davidb> if (weakAccessible.IsAlive())
- # [03:34] <@davidb> weakAccessible.getAccessible()->GetNameFoo()
- # [03:34] <@davidb> but I'm not clear if this is what surkov has in mind
- # [03:34] <@davidb> and IsAlive() returns false if the weakAccessible object has been cleared out
- # [03:35] <@davidb> presumable by the document or xpcom dying or something
- # [03:37] <@tbsaunde> davidb: yeah, I sort of know the api just didn't remember how they implemented seems they just keep a list of the eak frames on the preshell
- # [03:37] <@davidb> yes
- # [03:38] <@tbsaunde> I think surkov's idea is keep a hash table of accessible -> weak refs
- # [03:38] <@tbsaunde> then when the accessible wants to die it goes and wakes all the weak refs
- # [03:38] * Joins: jprmc (jprmc@moz-7F2FF3EB.cpe.net.cable.rogers.com)
- # [03:38] * ChanServ sets mode: +o jprmc
- # [03:38] <@davidb> it sounds very similar
- # [03:39] <@davidb> wakes?
- # [03:39] <@tbsaunde> davidb: in purpose yes
- # [03:39] <@tbsaunde> *wacks
- # [03:39] <@davidb> hmm
- # [03:39] <@davidb> well it should clear out the weak references raw ptr to the accessible
- # [03:39] <@tbsaunde> I'm not completely sure its needed, but we'll see
- # [03:39] <@tbsaunde> davidb: yeah, that's what I meant by wack
- # [03:39] <@davidb> yeah I'll need to see more code
- # [03:40] <@davidb> ok
- # [03:40] <@davidb> you and your fancy science terms :)
- # [03:41] <@davidb> it is advised to use weak frames sparingly, since there is overhead
- # [03:41] <@davidb> i need to think about our case but i need to understand what surkov is wanting here first
- # [03:41] <@tbsaunde> I don't think I need to see more code, I just want to see if we can't deal with all the cases we have in less complicated ways first
- # [03:41] <@davidb> this doesn't have to be complicated
- # [03:41] <@davidb> but it might be less performant
- # [03:41] <@davidb> that's my main concern
- # [03:42] <@davidb> it really depends what the idea is
- # [03:42] <@tbsaunde> using a weak ref will be slightly worse than a raw pointer, but shouldn't be bad
- # [03:42] <@davidb> usage, right.
- # [03:42] <@davidb> but there is more overhead
- # [03:42] <@tbsaunde> I'm not very concerned about perf
- # [03:42] * Joins: muralisr92 (chatzilla@moz-3BBB9F1C.dynip.nus.edu.sg)
- # [03:43] <@davidb> interesting
- # [03:43] <@tbsaunde> just a look up in a hash table so not much at all
- # [03:43] <@davidb> for each accessible
- # [03:44] <@tbsaunde> but I'm not completely sure how surkov plans to handle the ABA problem
- # [03:44] <@davidb> imagine a page mostly mutating
- # [03:44] <@davidb> ABA?
- # [03:45] <@davidb> i'm more worried about lifecycle
- # [03:45] <@tbsaunde> its a concurrency thing where a list item gets removed freed then reinserted
- # [03:45] <@tbsaunde> and someone else thinks its the original thing at that address not the new one
- # [03:45] <@davidb> OK so not the Alberta Bicycle Association
- # [03:45] <@davidb> ah ok
- # [03:46] <@davidb> that can get complicated
- # [03:46] <@tbsaunde> weak refs shouldn't really matter for mutations since we don't use them internally other than shutting an accessible down will be slightly slower
- # [03:46] <@davidb> good
- # [03:47] <@tbsaunde> but if we can just delete the accessible it might end up being faster if we don't need to nul out members
- # [03:47] <@davidb> i don't see how we can do that
- # [03:47] <@tbsaunde> why?
- # [03:48] <@davidb> we can't give out raw ptrs to garbage
- # [03:48] <@davidb> keep in mind I'm thinking nsWeakAccessible
- # [03:48] <Mark_Capella> or the American Bankers Association
- # [03:48] <@tbsaunde> why would we be doing that
- # [03:48] <@davidb> Mark_Capella: or the American Basketball Assoc
- # [03:48] <Mark_Capella> well we all have our reference points :)
- # [03:48] <@davidb> tbsaunde: i think i need to stop thinking about it until i understand the plan
- # [03:49] <@davidb> Mark_Capella: weak
- # [03:49] <@davidb> :)
- # [03:49] <Mark_Capella> holy crap you type fast dude
- # [03:49] <Mark_Capella> and tell bad puns
- # [03:49] <@davidb> ha! i'm not fast.
- # [03:49] <@tbsaunde> davidb: I sort of thought comment 2 explained it well enough, but ok
- # [03:49] <@davidb> it doesn't for me
- # [03:50] <@tbsaunde> davidb: I don't think surkov has suggested just deleting accessibles when we are killing that's me suggesting maybe we can stop ref counting them
- # [03:50] <@davidb> yeah I know it is you :)
- # [03:51] <@davidb> tbsaunde: what does "Make XPCOM object inherited from weak reference class" mean?
- # [03:51] <@davidb> this is silly - I should wait for surkov :)
- # [03:51] <@tbsaunde> but if it lets us stop refcounting accessibles then I suspect it'll be a perf win oerall
- # [03:51] <@davidb> his words
- # [03:51] <@davidb> absolutely
- # [03:52] <@davidb> now i wish we were having a work week
- # [03:52] <@tbsaunde> class xpcAccessibleTable : public WeakAcc { };
- # [03:52] <@davidb> ok
- # [03:53] <Mark_Capella> work weak
- # [03:53] <@tbsaunde> or maybe public WeakRefHolder
- # [03:53] <@davidb> WeakAcc is tbd
- # [03:53] <@davidb> ok
- # [03:53] <@davidb> Mark_Capella: groan
- # [03:53] <Mark_Capella> B)
- # [03:53] <@davidb> heheh
- # [03:53] <Mark_Capella> ok ill stop - gotta bail for a bit
- # [03:53] <@davidb> hey np
- # [03:57] <@davidb> tbsaunde: I look forward to September
- # [03:57] <@tbsaunde> davidb: make sense?
- # [03:57] <@tbsaunde> heh
- # [03:57] <@davidb> tbsaunde: yes it does.
- # [03:57] <@davidb> roughyl
- # [03:57] <@davidb> roughly
- # [03:58] <@tbsaunde> I suppose on the other hand weak could make the weak ref a member which would make killing the pointer easier, but doesn't solve the need for getting xpcom wrappers
- # [03:58] <@davidb> i think a member is clearer
- # [03:59] <@davidb> the wrapper stuff i need to think on
- # [03:59] <@tbsaunde> but possibly less useful
- # [03:59] <@davidb> in this case yes
- # [03:59] * Joins: drexler (chatzilla@moz-BD26ECCD.hsd1.vt.comcast.net)
- # [04:00] <@davidb> you gonna land the telem?
- # [04:00] <@tbsaunde> yeah, fairly soon on inbound
- # [04:00] <@davidb> nice
- # [04:01] <@davidb> ok i'm gonna check crash-stats and call it a night
- # [04:03] <@tbsaunde> ok
- # [04:03] * ctalbert is now known as ctalbert|afk
- # [04:03] <@tbsaunde> pushed :)
- # [04:03] * @davidb dances a little
- # [04:04] <@tbsaunde> heh
- # [04:04] <@tbsaunde> how does crash sts look?
- # [04:07] * Quits: ehsan (ehsan@F0B20A8D.8458880F.57F33CED.IP) (Input/output error)
- # [04:07] <@davidb> pretty good
- # [04:07] <@davidb> that get bounds fix seemed to take
- # [04:09] <@tbsaunde> yeah, saw bugmail
- # [04:11] <@davidb> hmmm i have an old patch to bail on dirty frames when creating an accessible
- # [04:11] <@davidb> i seem to recall the patch breaks us which is surprising - need to check that out again
- # [04:11] * @davidb is cleaning house
- # [04:11] * @davidb qdeletes
- # [04:15] <@davidb> ok cool I'm gonna try about put this laptop down.
- # [04:15] <@davidb> ciao!
- # [04:15] * Quits: @davidb (davidb@471D72E.2257F909.F30C9E9E.IP) (Quit: davidb)
- # [04:21] <@tbsaunde> Mark_Capella: my bet would be is that kNamespaeID_XHTML is wrong aand you need to use kNamespaceID_None, but that's just a guess and a debugger is your friend
- # [04:24] * Joins: ehsan (ehsan@F0B20A8D.8458880F.57F33CED.IP)
- # [04:24] * Joins: aaronlev (aaronlev@BCFC8969.9EBA4B5B.A641CD9D.IP)
- # [04:38] * Joins: surkov (surkov@B9538DF7.B7015738.33A1AC3C.IP)
- # [04:38] * ChanServ sets mode: +o surkov
- # [05:27] <Mark_Capella> tbsaunde: yah - that fixed it
- # [05:28] <@tbsaunde> rk$/nick tbsaunde|afk
- # [05:28] * tbsaunde is now known as tbsaunde|afk
- # [05:57] <Mark_Capella> surkov: im always looking to improve my c++ ... if you think bug678429 is something I can handle I'll be glad to try
- # [05:57] <@firebot> Bug https://bugzilla.mozilla.org/show_bug.cgi?id=678429 nor, --, ---, nobody, NEW, implement weak reference for accessible objects
- # [06:04] * Joins: jhk (jigneshhk1@8E6C34C1.A3F9767A.1C37C358.IP)
- # [06:05] <@surkov> Mark_Capella: yes, it's not hard but it should be very interesting
- # [06:05] * Quits: jhk (jigneshhk1@8E6C34C1.A3F9767A.1C37C358.IP) (Connection reset by peer)
- # [06:05] * Joins: jhk (jigneshhk1@8E6C34C1.A3F9767A.1C37C358.IP)
- # [06:06] <@firebot> markcapella@twcny.rr.com changed the Status on bug 678429 from NEW to ASSIGNED.
- # [06:07] * Quits: jhk (jigneshhk1@8E6C34C1.A3F9767A.1C37C358.IP) (Ping timeout)
- # [06:07] * Joins: jhk (jigneshhk1@8E6C34C1.A3F9767A.1C37C358.IP)
- # [06:08] * Quits: jhk (jigneshhk1@8E6C34C1.A3F9767A.1C37C358.IP) (Connection reset by peer)
- # [06:08] * Quits: muralisr92 (chatzilla@moz-3BBB9F1C.dynip.nus.edu.sg) (Quit: ChatZilla 0.9.88.1 [Firefox 11.0/20120310010316])
- # [06:08] * Joins: jhk (jigneshhk1@8E6C34C1.A3F9767A.1C37C358.IP)
- # [06:09] <@firebot> surkov.alexander@gmail.com requested review from trev.saunders@gmail .com for attachment 612328 on bug 740696.
- # [06:09] <@firebot> surkov.alexander@gmail.com granted feedback for attachment 612328 on bug 740696.
- # [06:09] <@firebot> Bug https://bugzilla.mozilla.org/show_bug.cgi?id=740696 nor, P1, ---, hub, NEW, [Mac] we have children that are expired.
- # [06:12] <@firebot> surkov.alexander@gmail.com cancelled feedback?(eitan@monotonous. org) for attachment 612038 on bug 670853.
- # [06:12] <@firebot> Bug https://bugzilla.mozilla.org/show_bug.cgi?id=670853 nor, --, ---, maxli, NEW, undetermined progressmeters should expose mixed state
- # [06:20] * tbsaunde|afk is now known as tbsaunde
- # [06:23] <@tbsaunde> surkov: David wanted me to look at bug 708927 any thoughts?
- # [06:23] <@firebot> Bug https://bugzilla.mozilla.org/show_bug.cgi?id=708927 nor, --, ---, nobody, NEW, Intermittent test_focus_menu.xul | Test timed out, sometimes followed by tens of thousands of gA11yE
- # [06:26] <@surkov> tbsaunde: (David always try to feed something :) ) well, you need to look at which test fails (see log, last info record says what events were registered and iirc it should say what last event was handled). Since these are menu tests then probably events are fired in async manner and that's a problem. But first of all you need to look at log.
- # [06:26] <@surkov> if log is not clear then probably it makes sense to improve it so it provides more info
- # [06:26] <@surkov> or you can enable debugging output on this test and wait for a failure
- # [06:29] <@tbsaunde> surkov: yeah, ok
- # [06:30] <@tbsaunde> surkov: what do you mean always trying to feed something/
- # [06:30] <@surkov> tbsaunde: finding a work
- # [06:34] <@tbsaunde> ah
- # [06:36] <@firebot> surkov.alexander@gmail.com granted review for attachment 612440 on bug 652635.
- # [06:37] <@firebot> Bug https://bugzilla.mozilla.org/show_bug.cgi?id=652635 nor, --, ---, markcapella, ASSI, Fallback missing @longdesc to aria-describedby pointing to <a href>
- # [06:39] <@firebot> trev.saunders@gmail.com changed the Target Milestone on bug 733510 from --- to mozilla14.
- # [06:39] <@firebot> Bug https://bugzilla.mozilla.org/show_bug.cgi?id=733510 nor, --, mozilla14, trev.saunders, NEW, Transition boolean a11y telemetry to 'flag' telemetry as appropriate
- # [06:48] * khuey is now known as khuey|away
- # [06:58] <@firebot> surkov.alexander@gmail.com requested review from jonas@sicking.cc for attachment 612440 on bug 652635.
- # [06:59] <@firebot> Bug https://bugzilla.mozilla.org/show_bug.cgi?id=652635 nor, --, ---, markcapella, ASSI, Fallback missing @longdesc to aria-describedby pointing to <a href>
- # [07:01] * Quits: peteb-away (ptbrunet@moz-E9B02845.austin.res.rr.com) (Client exited)
- # [07:15] * Quits: drexler (chatzilla@moz-BD26ECCD.hsd1.vt.comcast.net) (Quit: ChatZilla 0.9.88-rdmsoft [XULRunner 1.9.0.17/2009122204])
- # [07:24] * tbsaunde is now known as tbsaunde|afk
- # [07:27] * Joins: hub (hub@moz-E2FCA694.figuiere.net)
- # [07:27] * ChanServ sets mode: +o hub
- # [07:28] * Joins: MarcoZ (Daily@moz-C3875970.dip.t-dialin.net)
- # [07:28] * ChanServ sets mode: +o MarcoZ
- # [07:30] <@firebot> surkov.alexander@gmail.com granted review for attachment 612409 on bug 735645.
- # [07:30] <@firebot> Bug https://bugzilla.mozilla.org/show_bug.cgi?id=735645 nor, --, ---, maxli, NEW, expose sub and sup elements in text attributes
- # [07:35] <@firebot> surkov.alexander@gmail.com changed the Target Milestone on bug 733510 from mozilla14 to ---.
- # [07:35] <@firebot> Bug https://bugzilla.mozilla.org/show_bug.cgi?id=733510 nor, --, ---, trev.saunders, NEW, Transition boolean a11y telemetry to 'flag' telemetry as appropriate
- # [07:36] * Quits: aaronlev (aaronlev@BCFC8969.9EBA4B5B.A641CD9D.IP) (Quit: aaronlev)
- # [07:45] * Quits: @firebot (firebot@moz-67BF9FE6.carolina.res.rr.com) (Ping timeout)
- # [08:00] * Joins: firebot (firebot@moz-67BF9FE6.carolina.res.rr.com)
- # [08:00] * ChanServ sets mode: +o firebot
- # [08:08] <@firebot> markcapella@twcny.rr.com cancelled review?(jonas@sicking.cc) for attachment 612440 on bug 652635.
- # [08:08] <@firebot> Bug https://bugzilla.mozilla.org/show_bug.cgi?id=652635 nor, --, ---, markcapella, ASSI, Fallback missing @longdesc to aria-describedby pointing to <a href>
- # [08:09] <Mark_Capella> ah drats
- # [08:10] <@firebot> markcapella@twcny.rr.com requested review from jonas@sicking.cc for attachment 612460 on bug 652635.
- # [08:26] <fxa90id> hi
- # [08:27] * Quits: fxa90id (fxa90id@moz-829BEFC9.neoplus.adsl.tpnet.pl) (Quit: Leaving)
- # [08:35] * Quits: @hub (hub@moz-E2FCA694.figuiere.net) (Ping timeout)
- # [08:40] * Joins: victorporof (victorporo@AAF12FE3.67CDC207.79933D60.IP)
- # [08:41] * Quits: victorporof (victorporo@AAF12FE3.67CDC207.79933D60.IP) (Ping timeout)
- # [08:57] * Joins: margle (margle@moz-9542A19B.dsl.mweb.co.za)
- # [09:07] * Joins: icaaq (Adium@moz-FCAABB4F.cust.telenor.se)
- # [09:10] * Quits: jhk (jigneshhk1@8E6C34C1.A3F9767A.1C37C358.IP) (Connection reset by peer)
- # [09:12] * Quits: margle (margle@moz-9542A19B.dsl.mweb.co.za) (Quit: Computer has gone to sleep.)
- # [09:16] * Joins: jhk (jigneshhk1@8E6C34C1.A3F9767A.1C37C358.IP)
- # [09:21] * Quits: icaaq (Adium@moz-FCAABB4F.cust.telenor.se) (Connection reset by peer)
- # [09:29] * Joins: margle (margle@moz-9542A19B.dsl.mweb.co.za)
- # [09:42] * Quits: margle (margle@moz-9542A19B.dsl.mweb.co.za) (Quit: Computer has gone to sleep.)
- # [09:52] * Joins: icaaq (Adium@moz-FCAABB4F.cust.telenor.se)
- # [10:50] * Joins: victorporof (victorporo@622F294A.8E4D8E22.4A6B528C.IP)
- # [10:52] <@firebot> surkov.alexander@gmail.com changed the Target Milestone on bug 735645 from --- to mozilla14.
- # [10:52] <@firebot> surkov.alexander@gmail.com granted in-testsuite on bug 735645.
- # [10:52] <@firebot> Bug https://bugzilla.mozilla.org/show_bug.cgi?id=735645 nor, --, mozilla14, maxli, NEW, expose sub and sup elements in text attributes
- # [10:54] <@firebot> surkov.alexander@gmail.com cancelled review?(trev.saunders@gmail .com) for attachment 612149 on bug 741398.
- # [10:54] <@firebot> surkov.alexander@gmail.com requested review from trev.saunders@gmail .com for attachment 612476 on bug 741398.
- # [10:54] <@firebot> Bug https://bugzilla.mozilla.org/show_bug.cgi?id=741398 nor, --, ---, surkov.alexander, ASSI, make ARIA state map extensible
- # [11:00] <@firebot> eitan@monotonous.org cancelled feedback?(l10n@mozilla.com) for attachment 609578 on bug 739498.
- # [11:00] <@firebot> Bug https://bugzilla.mozilla.org/show_bug.cgi?id=739498 nor, --, ---, eitan, NEW, Add Javascript layer for mobile accessibility
- # [11:07] <@firebot> eitan@monotonous.org granted feedback for attachment 612359 on bug 670853.
- # [11:07] <@firebot> Bug https://bugzilla.mozilla.org/show_bug.cgi?id=670853 nor, --, ---, maxli, NEW, undetermined progressmeters should expose mixed state
- # [11:17] * Quits: icaaq (Adium@moz-FCAABB4F.cust.telenor.se) (Ping timeout)
- # [11:19] * Mark_Capella is now known as Mark_Capella|naptime
- # [11:20] * Joins: icaaq (Adium@moz-FCAABB4F.cust.telenor.se)
- # [11:43] <@firebot> New Core - Disability Access APIs bug 742657 filed by surkov.alexander@gmail.com.
- # [11:44] <@firebot> Bug https://bugzilla.mozilla.org/show_bug.cgi?id=742657 nor, --, ---, nobody, NEW, move nsAccUtils::GetRoleMapEntry(nsINode *aNode) into nsARIAMap
- # [11:49] <@firebot> New Core - Disability Access APIs bug 742659 filed by surkov.alexander@gmail.com.
- # [11:49] <@firebot> Bug https://bugzilla.mozilla.org/show_bug.cgi?id=742659 nor, --, ---, nobody, NEW, move nsARIAMap::UniversalStatesFor to aria namespace
- # [12:11] * Joins: fxa90id (fxa90id@moz-83BBC510.neoplus.adsl.tpnet.pl)
- # [12:19] * Quits: jhk (jigneshhk1@8E6C34C1.A3F9767A.1C37C358.IP) (Connection reset by peer)
- # [12:19] * Joins: jhk (jigneshhk1@8E6C34C1.A3F9767A.1C37C358.IP)
- # [12:21] * Quits: icaaq (Adium@moz-FCAABB4F.cust.telenor.se) (Connection reset by peer)
- # [12:28] * tbsaunde|afk is now known as tbsaunde
- # [12:32] * Joins: icaaq (Adium@moz-FCAABB4F.cust.telenor.se)
- # [12:44] * Quits: icaaq (Adium@moz-FCAABB4F.cust.telenor.se) (Ping timeout)
- # [12:44] * Joins: icaaq (Adium@moz-FCAABB4F.cust.telenor.se)
- # [12:57] <@tbsaunde> surkov: so, why do we seem to think that uia COM will work on ARM but msaa ia2 won't?
- # [12:57] <@surkov> tbsaunde: via can be not COM
- # [12:57] <@surkov> not IAccessibleEx of course
- # [12:57] <@surkov> but native UIA impl
- # [12:59] * Quits: victorporof (victorporo@622F294A.8E4D8E22.4A6B528C.IP) (Connection reset by peer)
- # [12:59] <@tbsaunde> surkov: and how would we implement such a thing, and how would implementing IWhatever help that?
- # [12:59] <@surkov> tbsaunde: that IWhatever is just not com interface if I get tight
- # [13:01] * Quits: shorlander (shorlander@moz-853043D6.dhcp.insightbb.com) (Connection reset by peer)
- # [13:01] <@tbsaunde> surkov: how? isn't it defined to inherit from IUnknown?
- # [13:01] * Joins: shorlander (shorlander@moz-853043D6.dhcp.insightbb.com)
- # [13:01] <@tbsaunde> and how does getting control patterns work without QI?
- # [13:01] <@surkov> tbsaunde: well, yes, those interface are COM
- # [13:02] <@surkov> that's what I heard actually and didn't look deeply
- # [13:03] <@tbsaunde> surkov: I haven't looked deeply, but it sounds crazy at a glance
- # [13:03] <@surkov> tbsaunde: ok, let's remove that part
- # [13:03] <@tbsaunde> surkov: ok
- # [13:03] <@tbsaunde> surkov: also, am I crazy or don't we implement IServiceProvider on all accessibles already?
- # [13:04] <@surkov> we implement
- # [13:04] <@surkov> the point is to extend its implementation
- # [13:04] <@surkov> bad wording I guess
- # [13:04] <@tbsaunde> surkov: sure
- # [13:04] <@tbsaunde> jhk: ^
- # [13:05] <@tbsaunde> surkov: what is you reasoning behind using IAccessibleEx instead of answering WM_GGETOBJECT with a IRawFragmentProviderRoot or whatever?
- # [13:06] <@tbsaunde> I've looked a little, but hadn't come to a opinion
- # [13:06] <@surkov> tbsaunde: for tree crawling?
- # [13:06] <@tbsaunde> surkov: I don't understand what you mean
- # [13:07] <@surkov> well, say uia crawls the our MSAA tree and it wants to get some UIA patterns, so they query IAccessibleEx and then get provider and other stuffs
- # [13:07] <@surkov> I think it works this way
- # [13:08] <@surkov> anyway just following recommendations
- # [13:08] <@surkov> btw, UIA provides own mechanism for events
- # [13:08] <@surkov> I"m not sure if WM_GETOBJECT is used there
- # [13:08] <@surkov> we have many questions more than answers
- # [13:09] <@tbsaunde> surkov: yeah, that is the way you do IAccessiblEx
- # [13:10] <@tbsaunde> but if you look at the msdn docs on implementing uia not based on msaa they talk about answering WM_GETOBJECT so I'm pretty sure its used for getting tree roots :)
- # [13:10] <@surkov> ok
- # [13:10] <@tbsaunde> but yes more questions than answers, and we don't actually need to decide now I think
- # [13:11] <@surkov> yep
- # [13:14] <@tbsaunde> jhk: your dates seem kind of funny to me the bigest thing was that control patterns will be a lot more work than IAccessibleEx, but probably doesn't really matter since we haven't decided what we actually want to do.
- # [13:15] <jhk> tbsaunde: yes It can change as per requirements...
- # [13:15] <jhk> and surkov "let's remove " which part?
- # [13:16] * @tbsaunde find the idea of him correcting english rather funny
- # [13:17] <@surkov> jhk: about COM stuff on ARM devices
- # [13:19] <jhk> surkov: ok
- # [13:24] * Quits: icaaq (Adium@moz-FCAABB4F.cust.telenor.se) (Ping timeout)
- # [13:25] * Joins: icaaq (Adium@moz-FCAABB4F.cust.telenor.se)
- # [13:28] * Quits: icaaq (Adium@moz-FCAABB4F.cust.telenor.se) (Ping timeout)
- # [13:31] * Joins: icaaq (Adium@moz-FCAABB4F.cust.telenor.se)
- # [13:35] <@firebot> trev.saunders@gmail.com granted review for attachment 612328 on bug 740696.
- # [13:35] <@firebot> Bug https://bugzilla.mozilla.org/show_bug.cgi?id=740696 nor, P1, ---, hub, NEW, [Mac] we have children that are expired.
- # [13:36] * Quits: icaaq (Adium@moz-FCAABB4F.cust.telenor.se) (Quit: Leaving.)
- # [13:51] <@firebot> maxli@maxli.ca requested review from askalski@mozilla.co m for attachment 612503 on bug 741699.
- # [13:51] <@firebot> Bug https://bugzilla.mozilla.org/show_bug.cgi?id=741699 nor, --, ---, maxli, NEW, telemetry for uiautomation.dll
- # [14:02] <@surkov> tbsaunde: what does densify mean actually?
- # [14:04] * Joins: icaaq (Adium@moz-7D09279.cust.bredbandsbolaget.se)
- # [14:08] <@tbsaunde> surkov: de ns ify
- # [14:08] <@tbsaunde> surkov: so remove ns
- # [14:16] * Mark_Capella|naptime is now known as Mark_Capella
- # [14:20] <@tbsaunde> surkov: make sense?
- # [14:31] <@surkov> yes
- # [14:32] <@tbsaunde> surkov: think it makes sense to file more bugs to rename stuff now or do we have enough good first bugs?
- # [14:33] <@surkov> tbsaunde: I think we should have one always
- # [14:35] * Joins: peteb-away (ptbrunet@moz-E9B02845.austin.res.rr.com)
- # [14:35] <@tbsaunde> surkov: ok, what next? html or xul form controls?
- # [14:36] <@surkov> tbsaunde: base/nsARIAGridAccessible should be easy
- # [14:36] <@surkov> let's finish with base classes
- # [14:36] <@tbsaunde> surkov: yeah, you have to deal with the wrap class but sure
- # [14:36] <@surkov> right
- # [14:37] <@tbsaunde> what should the wrap class be named? ARIAGridWrap?
- # [14:38] <@surkov> ARIAGridAccessibleWrap?
- # [14:39] <@tbsaunde> and namespace it too right?
- # [14:39] <@surkov> yes
- # [14:40] <@tbsaunde> ok
- # [14:45] <@tbsaunde> surkov: don't you find it weird attrValuesArray is of pointers to pointers to atoms?
- # [14:45] <@firebot> surkov.alexander@gmail.com changed the Target Milestone on bug 670853 from --- to mozilla14.
- # [14:45] <@firebot> surkov.alexander@gmail.com granted in-testsuite on bug 670853.
- # [14:45] <@firebot> Bug https://bugzilla.mozilla.org/show_bug.cgi?id=670853 nor, --, mozilla14, maxli, NEW, undetermined progressmeters should expose mixed state
- # [14:45] <@surkov> tbsaunde: yep
- # [14:45] <@surkov> I do
- # [14:46] <@surkov> do we have other option?
- # [14:47] <@tbsaunde> surkov: not that I'm aware of
- # [14:47] <@surkov> same with me
- # [14:48] <@surkov> tbsaunde: btw, are you going to file next densify bug?
- # [14:48] <@tbsaunde> sure
- # [14:48] <@surkov> ok
- # [14:49] <@tbsaunde> just got distracted into reviewing the aria stuffs
- # [14:50] <@tbsaunde> surkov: one more question do we want to do aria grid and grid cell at same time or break them up?
- # [14:51] <@surkov> perhaps the same time
- # [14:51] <@tbsaunde> seems like it would be nice if they could have thee own files, but not sure that's really possible
- # [14:55] <@firebot> New Core - Disability Access APIs bug 742695 filed by trev.saunders@gmail.com.
- # [14:55] <@firebot> Bug https://bugzilla.mozilla.org/show_bug.cgi?id=742695 nor, --, ---, nobody, NEW, densify nsARIAGrid
- # [14:58] * Quits: fxa90id (fxa90id@moz-83BBC510.neoplus.adsl.tpnet.pl) (Quit: Leaving)
- # [15:14] * Joins: davidb (davidb@471D72E.2257F909.F30C9E9E.IP)
- # [15:14] * ChanServ sets mode: +qo davidb davidb
- # [15:19] * Quits: nhirata (nhirata.bu@moz-585D5EFC.hsd1.ca.comcast.net) (Quit: nhirata)
- # [15:22] * Quits: jhk (jigneshhk1@8E6C34C1.A3F9767A.1C37C358.IP) (Connection reset by peer)
- # [15:23] <@tbsaunde> surkov: why does nsAccessible.h need to include nsARIAMap.h? Role()?
- # [15:24] <@surkov> tbsaunde: some inlines
- # [15:25] <@firebot> trev.saunders@gmail.com granted review for attachment 612476 on bug 741398.
- # [15:25] <@firebot> Bug https://bugzilla.mozilla.org/show_bug.cgi?id=741398 nor, --, ---, surkov.alexander, ASSI, make ARIA state map extensible
- # [15:25] <@tbsaunde> surkov: ok, I wonder if its worth making a file accessible-inl.h or something for inline members
- # [15:26] <@surkov> that'd be good
- # [15:27] <@tbsaunde> on the other hand maybe we should break up nsAccessible some how since theres several class where some of the members re unused
- # [15:28] <@tbsaunde> anyway got to run
- # [15:28] * tbsaunde is now known as tbsaunde|afk
- # [15:31] * Quits: ehsan (ehsan@F0B20A8D.8458880F.57F33CED.IP) (Input/output error)
- # [15:32] * Joins: aaronlev (aaronlev@BCFC8969.9EBA4B5B.A641CD9D.IP)
- # [15:32] <@davidb> surkov: jimm and I were talking a little bit last night. So far he hadn't been able to hook up some simple UIA.
- # [15:32] <@davidb> He had questions about performance and whether we can have modes.
- # [15:32] <@davidb> (A topic we've discussed before of course)
- # [15:33] <@surkov> davidb: what was a problem?
- # [15:33] <@davidb> Anyways wanted you to know he's active on it right now - we might need to move forward faster than GSoC
- # [15:33] <@davidb> surkov: no specific problem yet, just concern over the engine instantiation for all tablets etc.
- # [15:34] <@davidb> soft keyboard usage
- # [15:59] <@firebot> markcapella@twcny.rr.com requested feedback from surkov.alexander@ gmail.com for attachment 612525 on bug 739612.
- # [15:59] <@firebot> Bug https://bugzilla.mozilla.org/show_bug.cgi?id=739612 nor, --, ---, markcapella, ASSI, Cleanup A11y tests and test-suite organization
- # [16:03] <@davidb> MarcoZ: I was looking at crash stats last night and it appears to me we are pretty stable. Have you seen anything?
- # [16:06] * Joins: clown (clown@67828CC7.C1A51174.9D42CF23.IP)
- # [16:06] * clown is now known as clown_mtg
- # [16:07] <@davidb> surkov: I'd love an detailed explanation of how the weak references will work, and I'd like to be in the review queue for that bug.
- # [16:07] <@davidb> no rush
- # [16:08] <@MarcoZ> davidb: The number of reports I got from that particular community member have constantly gone down in the past 4 weeks. He's not reported new a11y crashes since shortly after CSUN.
- # [16:08] <@surkov> davidb: ok, if you want to review then steal the one from me, once you have a patch then you see how it works
- # [16:08] <@MarcoZ> I myself haven't seen any crashes lately, either. The ones I saw were ones that had been filed before.
- # [16:08] * Joins: jhk (jigneshhk1@8E6C34C1.A3F9767A.1C37C358.IP)
- # [16:08] <@MarcoZ> (and I skipped last weekend where we had the short-lived crash regression).
- # [16:09] <@davidb> surkov: yeah, if I'm too slow to steal, please ask me for feedback in near the end (if you remember) - I don't want to miss this one.
- # [16:09] <@davidb> MarcoZ: ok cool, is the community member testing nightly?
- # [16:10] <@MarcoZ> No, he's testing Aurora.
- # [16:12] <@MarcoZ> And the patches we up-lifted were ones he was seeing or had reported initially, too.
- # [16:12] <@surkov> davidb: actually I would love if you are involved from the start
- # [16:13] * clown_mtg is now known as clown
- # [16:14] <@davidb> surkov: then I will follow the bug.
- # [16:15] <@tbsaunde|afk> davidb: I'm curious why are you so interested?
- # [16:16] <@davidb> tbsaunde|afk: It gets my neurons firing.
- # [16:16] <@davidb> for some reason.
- # [16:16] <@tbsaunde|afk> heh
- # [16:16] <@davidb> MarcoZ: ok cool.
- # [16:17] * Joins: jimm (jmathies@moz-7F164CA1.pn.at.cox.net)
- # [16:18] <@MarcoZ> Hi jimm, welcome!
- # [16:18] <jimm> Hey MarcoZ, how ya doing?
- # [16:19] <@surkov> davidb: ok, cool
- # [16:19] <@davidb> jimm: is there a FF windows 8 roadmap/plan?
- # [16:20] <jimm> davidb: most everything we have thus far is on https://wiki.mozilla.org/Firefox/Windows_8_Integration
- # [16:20] <jimm> no official schedule yet
- # [16:21] <jimm> Asa has been working on getting resources together / planning
- # [16:21] <@davidb> OK
- # [16:21] <jimm> davidb: question for you, does accessibility expect all the calls into it to be on the main thread?
- # [16:22] <@davidb> except for plugins I think so
- # [16:22] <@davidb> (windowed plugins)
- # [16:22] <@davidb> surkov, tbsaunde|afk ^
- # [16:22] <@davidb> jimm: it gets murky for me around how COM works under the hood
- # [16:23] <jimm> the reason I ask is I ran into an off main thread exception in the cycle collector collecting an accessible object
- # [16:23] <@surkov> jimm: we actively working with frame tree which requires us to be called from main thread, right?
- # [16:23] <jimm> metro uses multiple threads for different things. thankfully UI rendering/events are all on a single thread.
- # [16:23] <@davidb> jimm: do we have a wiki of all the threads FF uses?
- # [16:23] <jimm> surkov: not sure
- # [16:23] <@davidb> I recall networking using threads
- # [16:23] <jimm> just got things hooked up and am debugging.
- # [16:24] <@davidb> nice!
- # [16:24] <jimm> davidb: no not yet
- # [16:24] <@surkov> mm, ok, I don't have answer then
- # [16:24] <jimm> davidb: we don't know what those threads are yet but that's a good idea
- # [16:25] * Joins: ehsan (ehsan@F2D29657.F60B0462.67AC9B1.IP)
- # [16:25] <jimm> davidb: the real main thread is owned by metro, we never see it except on startup
- # [16:25] <@davidb> We get calls from in-process injected dlls, as well as OOP
- # [16:25] <@davidb> interesting
- # [16:25] <jimm> there is a worker thread for rendering and events
- # [16:25] <jimm> there is also a suspend/resume thread for certain key events related to app state
- # [16:25] <@davidb> This is a notion of a privileged (or synonym?) thread in metro i hear?
- # [16:26] <@surkov> a11y relies on WM_GETOBJECT event
- # [16:26] * @davidb digs for notes
- # [16:26] <jimm> and now it seems there might be an accessibility thread, although I've just started looking at this. can't confirm that yet.
- # [16:26] <@davidb> surkov, yeah, winproc is main thread right?
- # [16:26] <@surkov> jimm said they use worker thread for events
- # [16:26] <jimm> in metro there is no wm_getobject because there is no window/window procedure
- # [16:26] <jimm> no hwnds either
- # [16:26] <@surkov> oh, I see
- # [16:27] <jimm> metro owns the window
- # [16:27] <jimm> we are embedded basically
- # [16:27] <@surkov> jimm: is there COM there?
- # [16:27] <@davidb> jimm: does metro manage a system wide a11y thread?
- # [16:27] <@davidb> or process
- # [16:27] * @davidb is confused
- # [16:27] <jimm> surkov: it's 100% COM under the hood
- # [16:27] <@surkov> ok
- # [16:28] <jimm> davidb: trying to figure that out now by itemizing threads / careful placement of IsMainThread debug output
- # [16:28] <@davidb> Mick (NVDA) said that their IPC tricks are broken in metro.
- # [16:28] <@davidb> cool
- # [16:29] <@davidb> funny how reverse engineering always trumps documentation
- # [16:29] <jimm> btw, I did manage to get UiaProviderFromIAccessible working.
- # [16:29] <@davidb> nice
- # [16:29] <jimm> the problem was a lack of an HWND for accessiblity from nsWindow's GetNativeData
- # [16:29] <@MarcoZ> davidb: jimm, I once heard that UIA is purposely inheriting nothing from MSAA, not even WM_GETOBJECT. They have the capability to bridge to MSAA, yes, but UIA doesn't inherit anything from it AFAIR:
- # [16:32] <jimm> for our new UIA lib we'll need to avoid the use of native windows/handles/messaging. I'm guessing that will be pretty easy as UIA isn't like IAccessible.
- # [16:34] * @MarcoZ nods.
- # [16:36] <@firebot> surkov.alexander@gmail.com requested review from marco.zehe@googlema il.com for attachment 612525 on bug 739612.
- # [16:36] <@firebot> Bug https://bugzilla.mozilla.org/show_bug.cgi?id=739612 nor, --, ---, markcapella, ASSI, Cleanup A11y tests and test-suite organization
- # [16:37] <@firebot> surkov.alexander@gmail.com changed the Target Milestone on bug 739889 from --- to mozilla14.
- # [16:37] <@firebot> Bug https://bugzilla.mozilla.org/show_bug.cgi?id=739889 nor, --, mozilla14, markcapella, ASSI, densify nsOuterDocAccessible
- # [16:38] <@davidb> surkov: it seems we are setting the target milestone when landing in inbound now? is that a new normal? i thought normally they did it at merge
- # [16:39] <@surkov> davidb: they asked it for
- # [16:39] <@davidb> hmm. ok.
- # [16:39] <@davidb> i guess it now means: please merge this in trunk for 14
- # [16:39] <@davidb> instead of: this was merged for 14
- # [16:40] <@MarcoZ> davidb: Yes, it's a recent change in policy.
- # [16:41] <@davidb> ok
- # [16:41] <@davidb> i did notice it happening but never bothered to ask
- # [16:42] <@davidb> surkov: so you want to block the merge on bug 733510?
- # [16:42] <@firebot> Bug https://bugzilla.mozilla.org/show_bug.cgi?id=733510 nor, --, ---, trev.saunders, NEW, Transition boolean a11y telemetry to 'flag' telemetry as appropriate
- # [16:43] <@surkov> davidb: if everybody agree then we can file new patch into inbound
- # [16:44] <@davidb> surkov: I wouldn't bother, but if you feel strongly that's fine.
- # [16:44] <@davidb> Note we will continue to get data for the old names on different trains
- # [16:44] <@surkov> davidb: I think I can live with that but not sure why you decided to go this way
- # [16:44] <@davidb> so there will be a clash
- # [16:44] <@surkov> davidb: but those data is not useful?
- # [16:45] <@davidb> surkov: comment 8
- # [16:45] <@surkov> you said you can remove that data?
- # [16:45] <@davidb> yeah but i was thinking at that time it would stop coming in
- # [16:45] <@davidb> based on the old way
- # [16:45] <@surkov> I see
- # [16:45] <@davidb> but the same named data will keep coming in from different trains - which sounded like a headache for the metrics guys
- # [16:46] <@surkov> ok, so we should live with these names
- # [16:46] <@davidb> this way we keep all the data and can eol the old data later
- # [16:46] <@davidb> ok cool
- # [16:46] <@davidb> surkov: want to flip it? or me?
- # [16:47] <@surkov> flip it? what does it mean?
- # [16:47] <@davidb> set the target mileston
- # [16:47] <@davidb> e
- # [16:47] <@davidb> (you removed it)
- # [16:48] <@firebot> dbolter@mozilla.com changed the Target Milestone on bug 733510 from --- to mozilla14.
- # [16:49] <@davidb> surkov: anyways. done
- # [16:49] <@surkov> k
- # [16:55] <@davidb> ok
- # [16:56] <@davidb> surkov: I want to understand what you're asking for on bug 729154
- # [16:56] <@firebot> Bug https://bugzilla.mozilla.org/show_bug.cgi?id=729154 nor, --, ---, dbolter, NEW, Telemetry for a11y instantiation by unknown cause.
- # [16:56] <@surkov> davidb: comment #?
- # [16:56] <@davidb> surkov: 7
- # [16:57] <@surkov> davidb: merge enums
- # [16:57] <@davidb> you want a combined structure but i thought we decided that was less clear
- # [16:57] * khuey|away is now known as khuey
- # [16:58] <@surkov> I don't recall this
- # [16:59] <@davidb> surkov: it complicates code. (bug 678965)
- # [16:59] <@firebot> Bug https://bugzilla.mozilla.org/show_bug.cgi?id=678965 nor, --, mozilla12, dbolter, RESO FIXED, telemetry for injected screen reader dll's
- # [16:59] <@surkov> davidb: comment #?
- # [16:59] <@davidb> e.g. comment 27
- # [17:00] <@davidb> Nathan advised doing it this clear way.
- # [17:00] <@surkov> mm, sorry, don't see how it's related with having one enum idea
- # [17:00] <@davidb> then i must not be understanding
- # [17:01] <@davidb> ok so you want to get rid of the bitflag?
- # [17:01] <@davidb> ?
- # [17:01] <@davidb> surkov: ^
- # [17:01] * Joins: margle (margle@moz-9542A19B.dsl.mweb.co.za)
- # [17:01] <@davidb> that must be it
- # [17:01] <@surkov> davidb: yep, it seems it's enough to have one enum, iirc tbsaunde agree
- # [17:01] <@davidb> ok
- # [17:01] <@davidb> i misunderstood "get rid telemetry enum"
- # [17:02] * Joins: nhirata (nhirata.bu@moz-C03D0C61.vlan426.asr1.sfo1.gblx.net)
- # [17:02] <@davidb> i can agree with this too
- # [17:02] <@davidb> surkov: still i'm not sure what it buys us
- # [17:03] <@surkov> davidb: it means merge it to modes enum, have one enum for both cases
- # [17:03] <@davidb> it merges "modes" and "consumers" into one enum - which to me does not clarify things
- # [17:03] <@davidb> i realize there is overlap, but the non-overlap is the problem i have
- # [17:03] * Joins: fxa90id (fxa90id@moz-83BBC510.neoplus.adsl.tpnet.pl)
- # [17:04] <fxa90id> surkov, yes
- # [17:04] <fxa90id> sry
- # [17:04] <@davidb> also, our mode code usage of the bit flag is sort of nice
- # [17:04] * Quits: margle (margle@moz-9542A19B.dsl.mweb.co.za) (Quit: Computer has gone to sleep.)
- # [17:04] * Quits: aaronlev (aaronlev@BCFC8969.9EBA4B5B.A641CD9D.IP) (Quit: aaronlev)
- # [17:05] <@firebot> marco.zehe@googlemail.com cancelled review?(marco.zehe@googlema il.com) for attachment 612525 on bug 739612.
- # [17:05] <@firebot> Bug https://bugzilla.mozilla.org/show_bug.cgi?id=739612 nor, --, ---, markcapella, ASSI, Cleanup A11y tests and test-suite organization
- # [17:05] <@surkov> davidb: well, we have enum for running ATs, some of them have specific modes
- # [17:05] <@davidb> yeah
- # [17:05] <@surkov> it's clear enough, for me
- # [17:06] <@davidb> surkov: in bug 678965 you didn't want the mode data reported - is that still the case?
- # [17:06] <@davidb> because this is going to get unnecessarily complex i think
- # [17:06] <@surkov> davidb: what is mode data?
- # [17:06] <@davidb> comment 27
- # [17:06] <@surkov> IA2Off?
- # [17:06] <@davidb> yeah
- # [17:07] <@surkov> yeah, you need to have enum like JAWSNumberBlaBla and then have IA2Off mode { if JAWSNumberBlabla then true; }
- # [17:07] <@davidb> i don't understand that :)
- # [17:08] <@davidb> i'm thinking about the reporting part
- # [17:08] <@davidb> right now it is very simple.
- # [17:10] * clown is now known as clown_mtg
- # [17:13] <@davidb> surkov: do you know what I mean?
- # [17:14] <@surkov> not sure
- # [17:16] <@davidb> darn
- # [17:16] <@davidb> well i want to land this so i guess i'll try the merge
- # [17:17] * Joins: askalski (akuda@moz-6A36EC49.ip.abpl.pl)
- # [17:17] * ChanServ sets mode: +o askalski
- # [17:20] * Joins: hub (hub@moz-E2FCA694.figuiere.net)
- # [17:20] * ChanServ sets mode: +o hub
- # [17:21] <@davidb> surkov: you still don't like the idea of the modes to be reported right?
- # [17:21] <@davidb> i just want to be clear
- # [17:22] <@surkov> davidb: it doesn't make sense to report modes because modes are turned off and on deeding on running AT
- # [17:22] <@surkov> davidb: about landing, you still don't have review?
- # [17:22] <@davidb> surkov: i'm not sure i understand the concern unless you are worried about noise?
- # [17:23] <@davidb> I don't have a peer review.
- # [17:23] <@surkov> reporting modes means duping information
- # [17:23] <@surkov> you need to get one
- # [17:23] <@davidb> surkov: you cancelled review :)
- # [17:23] <@surkov> davidb: because I wan't happy with patch approach I guess
- # [17:23] <@davidb> right
- # [17:24] <@davidb> so that's why i need to try the merge right?
- # [17:24] <@surkov> I don't r- if the patch is correct but not nice
- # [17:24] <@surkov> yeah, that should be nice
- # [17:24] <@davidb> i mean at the end of the day i have huge respect for you and trev so i don't need to agree :)
- # [17:24] <@surkov> and easy :)
- # [17:25] <@surkov> well, you need to be agree if you think the approach is right
- # [17:25] <@surkov> otherwise you shouldn't :)
- # [17:25] <@davidb> surkov: i might put the unreported modes up in high numbers to keep them out of the way
- # [17:25] <@davidb> we can't always agree
- # [17:25] <@davidb> looks at {}'s ;)
- # [17:25] <@surkov> we can :)
- # [17:25] <@davidb> we might not always agree expediently :)
- # [17:25] <@surkov> ok
- # [17:28] * Joins: victorporof (victorporo@40A61E2F.70F4578F.79933D60.IP)
- # [17:29] * Quits: Mark_Capella (chatzilla@moz-DD0C7E4F.twcny.res.rr.com) (Quit: ChatZilla 0.9.88.1 [Firefox 13.0a2/20120402042010])
- # [17:34] * Joins: Mark_Capella (chatzilla@moz-DD0C7E4F.twcny.res.rr.com)
- # [17:34] <@davidb> oh actually the way telemetry works I'm not sure what big gaps will do
- # [17:35] * @davidb wanders over to metrics
- # [17:36] <@davidb> actually shouldn't matter really
- # [17:37] <jimm> davidb: this is the kind of threading issues I'm seeing:
- # [17:37] <jimm> mozilla::widget::winrt::UIAutomationWrap::get_ProviderOptions: IsMainThread:1 ThreadId:10C4
- # [17:37] <jimm> mozilla::widget::winrt::UIAutomationWrap::get_ProviderOptions: IsMainThread:0 ThreadId:FDC
- # [17:37] <jimm> mozilla::widget::winrt::UIAutomationWrap::Navigate: IsMainThread:0 ThreadId:FDC
- # [17:37] <jimm> ###!!! ASSERTION: nsAccessNode not thread-safe: '_mOwningThread.GetThread() == PR_GetCurrentThread()', file m:/Mozilla/elm/accessible/src/base/nsAccessNode.cpp, line 85
- # [17:38] <@davidb> this is all pretty new to me
- # [17:38] <jimm> different threads calling into our uia wrapper at different times
- # [17:38] <jimm> which end up making calls into our IAccessible interface
- # [17:38] <jimm> yeah me to :)
- # [17:38] <@davidb> i don't recall seeing that assertion
- # [17:39] <jimm> I think on desktop everything happens on the main thread
- # [17:39] <jimm> I can test to see
- # [17:39] <@davidb> that would be good to ensure
- # [17:41] <@MarcoZ> davidb, surkov: Isn't nsAccessNode only used for iSimpleDOM?
- # [17:42] <@surkov> MarcoZ: correct
- # [17:42] <@davidb> but doesn't nsAccessible derive from nsAccessNodeWrap that derives from nsAccessNode?
- # [17:43] <@surkov> davidb: correct
- # [17:43] <@davidb> i guess the question is what line 85 is doing
- # [17:43] <@davidb> or that part of the code
- # [17:43] <@surkov> davidb: addref/release?
- # [17:44] <@surkov> yes, cycle collecting addref/release
- # [17:44] <@davidb> ok
- # [17:44] <@surkov> cycle collection wants to be on main thread
- # [17:44] <@davidb> makes sense
- # [17:45] <@surkov> but we want to split msaa and internal (cycle collected) implementation
- # [17:45] <@surkov> it's long term though
- # [17:45] <jimm> something is working right yet since in inspect.exe I see a long list of UIA_E_ELEMENTNOTAVAILABLE errors and no sub tree.
- # [17:45] <@surkov> and big work
- # [17:45] <@davidb> surkov: should this big work be a priority?
- # [17:45] <@surkov> well, we have many priorities
- # [17:46] <@davidb> agreed
- # [17:46] <@surkov> and not much resources
- # [17:46] <@davidb> it is an optimization problem :)
- # [17:46] <@surkov> we are trying to put this work into good first bugs so we are getting permanent progress no it
- # [17:46] <@surkov> not super fast though
- # [17:46] <@davidb> (or constraint satisfaction)
- # [17:46] <jimm> those assertions are from nsISupports. someone is calling one of the base methods on a thread that's different from the creation thread.
- # [17:47] <@surkov> davidb: note optimization takes a time too, sometimes significant time, and nobody can confirm it will work better
- # [17:47] <@surkov> in the end
- # [17:47] <@davidb> surkov: I meant' optimizating our resources.. .but anyways… jimm is here! :)
- # [17:47] <@surkov> jimm: yes, COM and ISupports have shared ref counter
- # [17:47] <@surkov> davidb: that's what I meant too
- # [17:48] <@davidb> ok
- # [17:48] <@davidb> (let's talk about that later)
- # [17:48] * Joins: margle (margle@moz-9542A19B.dsl.mweb.co.za)
- # [17:48] <@surkov> davidb: you're thinking that human resources optimization is easier than code optimization? :)
- # [17:48] <@surkov> sure
- # [17:48] <@davidb> heheh
- # [17:49] <@davidb> surkov, jimm so who would be calling us?
- # [17:49] <@surkov> uiaautomation.dll I think
- # [17:49] <@surkov> or how is it callled
- # [17:49] <@davidb> close enough
- # [17:49] <@davidb> but that's in process
- # [17:50] <@surkov> because we have similar problems with some screen readers on winxp
- # [17:50] <@davidb> so what are the characteristics (thread wise) of injected dlls calling us
- # [17:50] <@surkov> that's why we started that work
- # [17:50] * @davidb nods
- # [17:50] <jimm> yes uiaautomation.dll
- # [17:51] <@davidb> ok
- # [17:51] <jimm> we setup the wrapper on the main thread, then uia at some point starts inspecting the document using a different thread
- # [17:51] <jimm> although from mxr it looks like those asserts are passive so everything should still be succeeding
- # [17:52] <jimm> I'll break point one of these..
- # [17:53] <jimm> yeah should be passive - http://mxr.mozilla.org/mozilla-central/source/xpcom/glue/nsISupportsImpl.h#337
- # [17:54] <jimm> I wonder why instect can't see the tree
- # [17:54] <jimm> How critical are HWNDs in accessible code? I'm currently creating dummy hwnds in widget to make accessible happy. The don't point to the real window.
- # [17:55] <jimm> *The -> They
- # [17:55] <@davidb> surkov knows that stuff deeply
- # [17:56] <@surkov> jimm: HWND is primary thing that AT relies on
- # [17:57] <jimm> 'AT'?
- # [17:57] <@surkov> jimm: like screen readers
- # [17:57] <jimm> davidb: note the same asserts don't happen on the desktop. it's all the same thread there.
- # [17:57] <@davidb> ok thanks
- # [17:58] <jimm> surkov: oh, not worried about that currently. in metro, metro is a uia client listener. all I'm working on currently is getting that part working.
- # [17:59] <jimm> surkov: since metro uses uia to do things like bring up the soft keyboard when a text edit is tapped.
- # [17:59] <@surkov> well, uia is critical thing for metro
- # [18:00] <jimm> surkov: initially when I tried to create a uia wrapper of our IAccessible using the following api - UiaProviderFromIAccessible
- # [18:00] <jimm> http://msdn.microsoft.com/en-us/library/windows/desktop/hh437315(v=VS.85).aspx
- # [18:01] <jimm> the call failed. I tracked this down to our accessible code failing due to the lack of a result from nsIWidget's GetNativeData(window handle)
- # [18:01] <@surkov> I wonder how do they find an accessible at all, since currently they use WM_GETOBJECT to get top accessible object
- # [18:02] <jimm> I hand it to them via UiaProviderFromIAccessible
- # [18:02] <jimm> they hand back IRawElementProviderSimple that wrpas our IAccessible
- # [18:02] <@surkov> how do they call into us?
- # [18:02] <@surkov> or what do they call
- # [18:02] <@surkov> I mean how do they start accessibility at all?
- # [18:02] <jimm> through IRawElementProviderSimple and the other uia interfaces their generated wrapper supports.
- # [18:03] <jimm> http://pastebin.mozilla.org/1554960
- # [18:03] <jimm> that might not make a lot of sense but..
- # [18:03] <jimm> basically it's our wrapper around their wrapper which metro talks to
- # [18:04] <@surkov> I see
- # [18:04] <jimm> their wrapper wraps our IAccessible. :)
- # [18:05] <@davidb> it is like christmas
- # [18:05] <@davidb> and IAccessible is the unwanted fruitcake
- # [18:06] <@surkov> :)
- # [18:06] <jimm> that's why they gave us UiaProviderFromIAccessible
- # [18:06] <jimm> they knew a lot of client apps still used IAccessible
- # [18:07] <@davidb> including office
- # [18:08] <@davidb> jimm: if we have questions i have a contact - i will pm you
- # [18:08] <jimm> cool
- # [18:08] <jimm> anyway, I'm thinking these dummy windows I'm using are still causing problems.
- # [18:09] <@davidb> perhaps the relationship between the windows
- # [18:10] * Joins: drexler (chatzilla@moz-6F22CAC7.e7jvsoe-gw.cust.sover.net)
- # [18:12] <@davidb> surkov: are you still awake? i can't recall our dummy window situation… if we needed a parent-child window relationship
- # [18:13] <@surkov> davidb: sort of awake, I'm still not sure what fails
- # [18:13] <@surkov> jimm: was you able to get our accessible object?
- # [18:13] <jimm> sure
- # [18:13] <@surkov> but when you try to wrap it then you fail?
- # [18:14] <jimm> that part is working too now that widget is handing back hwnds from GetNativeData
- # [18:14] <jimm> but those hwnds are fakes, so maybe that is still breaking something.
- # [18:14] <jimm> in our accessibility code
- # [18:14] <@davidb> do the IRawElementProviderSimple methods work?
- # [18:14] <@surkov> I see
- # [18:15] <jimm> davidb: yes, metro is quite busy calling those when a touch display is present.
- # [18:15] <@davidb> and good things are returned?
- # [18:15] <jimm> and those calls definitely call into our accessibility codee
- # [18:15] <@davidb> nice ok
- # [18:16] <@surkov> btw, we don't create accessible for window at all, wonder if that could be a problem
- # [18:16] <jimm> I need to check return resuts but it seems as though these calls are working
- # [18:16] <jimm> hmm
- # [18:16] <jimm> surkov: when I call UiaProviderFromIAccessible
- # [18:17] <@surkov> so on desktop Windows create an accessible for window, when we handle WM_GETOBJECT for client then we return document accessible
- # [18:17] <@surkov> we hack its parent to return window accessible created by Windows
- # [18:17] <jimm> surkov: what do I get from this
- # [18:17] <jimm> :
- # [18:17] <jimm> nsAccessible *rootAccessible = mWidget->GetRootAccessible();
- # [18:18] <jimm> IAccessible *msaaAccessible = NULL;
- # [18:18] <jimm> rootAccessible->GetNativeInterface((void**)&msaaAccessible);
- # [18:18] <jimm> document?
- # [18:18] <@surkov> jimm: root document accessible
- # [18:18] <jimm> ok
- # [18:18] <jimm> when I call UiaProviderFromIAccessible I pass OBJID_CLIENT as the child id
- # [18:18] <jimm> that seems correct to me.
- # [18:18] <@surkov> hierarchy should look like: Window accessible object (created by Windows for Firefox window) -> root document accessible
- # [18:19] <jimm> yeah that looks right. I thyink we're setup that way.
- # [18:19] <Mark_Capella> surkov: re:Bug 678429 - implement weak reference for accessible objects
- # [18:19] <Mark_Capella> That's one heck of a thread you've got me following :)
- # [18:19] <@firebot> Bug https://bugzilla.mozilla.org/show_bug.cgi?id=678429 nor, --, ---, markcapella, ASSI, implement weak reference for accessible objects
- # [18:20] <@surkov> jimm: do you pass OBJIID_CLIENT for any accessible object?
- # [18:20] <@surkov> for every
- # [18:20] <@surkov> Mark_Capella: sorry, tbsaunde is guilty ;)
- # [18:20] <Mark_Capella> Lets Get im!!!
- # [18:20] <@surkov> )
- # [18:21] <Mark_Capella> Well you said easy but interesting ... Very!
- # [18:21] <jimm> I only pass that once to UiaProviderFromIAccessible. UiaProviderFromIAccessible then generates a uia wrapper around our document IAccessible, which I hand back to metro when it asks for it.
- # [18:21] <@surkov> ok
- # [18:21] <@surkov> so presumingly metro is interested in root document accessible object
- # [18:22] <@surkov> than it does all magic on its side when it traverses the tree
- # [18:22] <jimm> yes I believe so
- # [18:22] <@surkov> I assume
- # [18:29] * @MarcoZ decides he is very, very scared now. :D
- # [18:31] <jimm> surkov: is nsAccessibleWrap the document object?
- # [18:32] <@surkov> jimm: no, nsDocAccessible is document object
- # [18:32] <@surkov> root document object is nsRootAccessible
- # [18:32] <@surkov> of course they are both inherited from nsAccessibleWrap
- # [18:32] <jimm> ah ok
- # [18:32] <jimm> so from GetNativeInterface() I get nsRootAccessible?
- # [18:32] <jimm> (on that first call)
- # [18:33] <@surkov> GetNativeInterface I think returns IUnknown from nsISupports object
- # [18:33] <@surkov> but mWidget->GetRootAccessible(); return nsRootAccessible
- # [18:34] <@surkov> or it should at least in non e10s builds
- # [18:34] * Quits: fxa90id (fxa90id@moz-83BBC510.neoplus.adsl.tpnet.pl) (Ping timeout)
- # [18:41] <jimm> whatever we get back from GetNativeInterface is what we hand back to windows from WM_GETOBJEVT - http://mxr.mozilla.org/mozilla-central/source/widget/windows/nsWindow.cpp#5168
- # [18:41] * Quits: jhk (jigneshhk1@8E6C34C1.A3F9767A.1C37C358.IP) (Connection reset by peer)
- # [18:41] <jimm> so I'm assuming that's the top level nsDocAccessible
- # [18:41] * Joins: jhk (jigneshhk1@8E6C34C1.A3F9767A.1C37C358.IP)
- # [18:41] <jimm> which is what I wrap with our uia wrapper
- # [18:44] <@firebot> surkov.alexander@gmail.com changed the Target Milestone on bug 741398 from --- to mozilla14.
- # [18:44] <@firebot> Bug https://bugzilla.mozilla.org/show_bug.cgi?id=741398 nor, --, mozilla14, surkov.alexander, ASSI, make ARIA state map extensible
- # [18:49] * clown_mtg is now known as clown
- # [18:50] <jimm> surkov: you raise an interesting point - do we know if this works with e10s / fennec XUL's out of process tabs?
- # [18:51] <jimm> I guess it works on desktop fennec so it should work in metro.
- # [18:52] <@surkov> the difference on e10s is mWidget->GetRootAccessible() return nsDocAccessibleWrap instead nsRootAccessibleWrap but that shouldn't be a problem
- # [18:53] <@surkov> we were somehow accessible on e10s build
- # [18:53] <@surkov> not perfectly though
- # [18:55] <jimm> inspect.exe doesn't seem to have any trouble loading the tree on the desktop build of fennec w/remote tabs enabled so that's a good sign.
- # [18:55] <@surkov> ok
- # [19:02] * Quits: @hub (hub@moz-E2FCA694.figuiere.net) (Ping timeout)
- # [19:05] * Joins: victorporo (victorporo@E25721DF.A00C134E.79933D60.IP)
- # [19:05] * @davidb steps out
- # [19:05] * Quits: victorporof (victorporo@40A61E2F.70F4578F.79933D60.IP) (Ping timeout)
- # [19:10] * khuey is now known as khuey|away
- # [19:14] * ctalbert|afk is now known as ctalbert
- # [19:21] * Quits: @surkov (surkov@B9538DF7.B7015738.33A1AC3C.IP) (Quit: surkov)
- # [19:25] * Quits: @MarcoZ (Daily@moz-C3875970.dip.t-dialin.net) (Quit: MarcoZ)
- # [19:26] * Joins: hub (hub@21B7B9F2.B87E9213.6E712CE2.IP)
- # [19:26] * ChanServ sets mode: +o hub
- # [19:32] * Quits: @askalski (akuda@moz-6A36EC49.ip.abpl.pl) (Ping timeout)
- # [19:39] <@davidb> hi hub
- # [19:42] * Joins: askalski (akuda@moz-6A36EC49.ip.abpl.pl)
- # [19:42] * ChanServ sets mode: +o askalski
- # [19:43] * jimm is now known as jimm-lunch
- # [19:46] <@hub> 3 people in the Yahoo! a11y team were laid off
- # [19:50] <@davidb> hub: is this on the twitters? i don't check during day
- # [19:50] <@davidb> hub: who?
- # [19:52] <@davidb> oh 2000 jobs
- # [19:52] * khuey|away is now known as khuey
- # [19:52] <@firebot> ehsan@mozilla.com changed the Status on bug 734566 from ASSIGNED to RESOLVED.
- # [19:52] <@firebot> ehsan@mozilla.com set the Resolution field on bug 734566 to FIXED.
- # [19:52] <@firebot> Bug https://bugzilla.mozilla.org/show_bug.cgi?id=734566 nor, --, mozilla14, michaljev, RESO FIXED, optimize layout of TextAttrsMgr
- # [19:53] <@hub> davidb: jenisson twitted about it
- # [19:53] <@firebot> ehsan@mozilla.com changed the Status on bug 559746 from NEW to RESOLVED.
- # [19:53] <@firebot> ehsan@mozilla.com set the Resolution field on bug 559746 to FIXED.
- # [19:53] <@firebot> Bug https://bugzilla.mozilla.org/show_bug.cgi?id=559746 nor, --, mozilla14, maxli, RESO FIXED, add a11y mochitests for <input type="tel">
- # [19:53] <@hub> davidb: amongst the 2000 layoffs
- # [19:54] <@hub> davidb: https://twitter.com/#!/yahooaccess/status/187857369269665792
- # [19:55] <@davidb> thanks
- # [20:01] <@davidb> tbsaunde|afk: if you can answer my latest on bug 729154 feel free
- # [20:01] <@firebot> Bug https://bugzilla.mozilla.org/show_bug.cgi?id=729154 nor, --, ---, dbolter, NEW, Telemetry for a11y instantiation by unknown cause.
- # [20:24] * jimm-lunch is now known as jimm
- # [20:25] <@askalski> davidb, while reviewing a patch, should I recompile to test it?
- # [20:26] <@davidb> askalski: for contributors it can be a good idea
- # [20:27] <@davidb> or generally if you suspect a compile issue
- # [20:27] <@askalski> davidb, sure
- # [20:27] <@davidb> it isn't required though
- # [20:27] <@davidb> and you can always mention you didn't compile it if you like
- # [20:28] <@davidb> before landing we compile locally and/or on the try server depending on perceived risk
- # [20:30] <@askalski> davidb, I definitely need to learn about try servers
- # [20:30] <@askalski> I noticed dozens of memes about it
- # [20:31] <@askalski> right now I compile locally all the time, and do mochitests locally as well
- # [20:31] <@askalski> btw, the machine is very strong
- # [20:31] <@davidb> that's what i do
- # [20:31] <@davidb> good :)
- # [20:32] <@firebot> ehsan@mozilla.com changed the Status on bug 733510 from NEW to RESOLVED.
- # [20:32] <@firebot> ehsan@mozilla.com set the Resolution field on bug 733510 to FIXED.
- # [20:32] <@askalski> and smart with power management. my previous one did not last long on a battery while using virtualbox. now I use 1-2 virtual machines (I build on windows 7 and test MATS on windows xp) and everything works fine
- # [20:32] <@firebot> Bug https://bugzilla.mozilla.org/show_bug.cgi?id=733510 nor, --, mozilla14, trev.saunders, RESO FIXED, Transition boolean a11y telemetry to 'flag' telemetry as appropriate
- # [20:34] <@askalski> davidb, how did the meeting with microsoft go? do we imlement UIA?
- # [20:34] <@hub> *sigh* some big stuff landed on inbound between my last update and the end of my build...
- # [20:35] <@davidb> telemetry flags?
- # [20:35] <@davidb> or more wide ranging i gues
- # [20:43] <jimm> davidb: question for you if you have a sec
- # [20:43] <@davidb> jimm: yes.
- # [20:43] <jimm> trying to understand the signifigance of this event sending:
- # [20:43] <jimm> http://mxr.mozilla.org/mozilla-central/source/accessible/src/msaa/nsAccessibleWrap.cpp#1558
- # [20:44] <@davidb> ok i'm weakish on our msaa layer but am reading...
- # [20:44] <jimm> I think that's for screen readers and the like
- # [20:44] <@davidb> yeah
- # [20:44] <jimm> they might hook our window to get at those events
- # [20:44] <jimm> ok
- # [20:44] <jimm> but we (mozilla) don't hook for those afaict
- # [20:45] <jimm> we just send them
- # [20:45] <@davidb> our own? no i don't think we do
- # [20:45] <jimm> ok cool
- # [20:46] <@firebot> ehsan@mozilla.com changed the Status on bug 735645 from NEW to RESOLVED.
- # [20:46] <@firebot> ehsan@mozilla.com set the Resolution field on bug 735645 to FIXED.
- # [20:46] <@davidb> our engine has no instances of calling SetWinEventHook
- # [20:46] <@firebot> Bug https://bugzilla.mozilla.org/show_bug.cgi?id=735645 nor, --, mozilla14, maxli, RESO FIXED, expose sub and sup elements in text attributes
- # [20:48] <@firebot> ehsan@mozilla.com changed the Status on bug 670853 from NEW to RESOLVED.
- # [20:48] <@firebot> ehsan@mozilla.com set the Resolution field on bug 670853 to FIXED.
- # [20:48] <@firebot> Bug https://bugzilla.mozilla.org/show_bug.cgi?id=670853 nor, --, mozilla14, maxli, RESO FIXED, undetermined progressmeters should expose mixed state
- # [21:12] <@firebot> markcapella@twcny.rr.com cancelled feedback?(surkov.alexander@ gmail.com) for attachment 612525 on bug 739612.
- # [21:12] <@firebot> Bug https://bugzilla.mozilla.org/show_bug.cgi?id=739612 nor, --, ---, markcapella, ASSI, Cleanup A11y tests and test-suite organization
- # [21:29] <@askalski> davidb, is uiautomation part of UIA?
- # [21:29] <@davidb> it is UIA
- # [21:29] <@askalski> davidb, good. just checking the names
- # [21:30] <@davidb> yep
- # [21:30] <@davidb> one api two purposes
- # [21:30] <@davidb> testing and a11y
- # [21:30] <@askalski> yep, it's a great way to promote a11y in business, really
- # [21:30] <@davidb> yeah
- # [21:30] <@davidb> the testing side of it has made it really stable
- # [21:30] <@davidb> mostly it is MS using it though
- # [21:30] <@davidb> s
- # [21:30] <@davidb> so far
- # [21:31] <@askalski> getting these two things bound on most popular platform is a huge advantage
- # [21:31] <@davidb> LDTP uses atk/atspi similarly
- # [21:31] <@askalski> LDTP?
- # [21:31] <@davidb> linux desktop testing project
- # [21:31] <@davidb> http://ldtp.freedesktop.org/wiki/
- # [21:32] <@davidb> but I don't mean to distract you
- # [21:32] <@askalski> I was reviewing a no-brainer patch, it loks very much the same as referenced one
- # [21:33] <@askalski> so I just made sure the names match, now I will check for compilation errors (very unlikely) and do r+
- # [21:33] <@davidb> the telemetry one? yeah
- # [21:33] <@davidb> looks good
- # [21:33] <@askalski> I guess there is no point in debugging to see if it triggers
- # [21:33] <@askalski> since I have no setup ready
- # [21:33] <@davidb> then no
- # [21:34] <@askalski> I mean - I have build environment running etc, but I don't know how to run debugger for ff in VS
- # [21:36] * Quits: peteb-away (ptbrunet@moz-E9B02845.austin.res.rr.com) (Ping timeout)
- # [21:36] <@davidb> for this patch eye-balling is ok
- # [21:36] <@davidb> make sure it isn't destructive
- # [21:43] * Quits: jhk (jigneshhk1@8E6C34C1.A3F9767A.1C37C358.IP) (Connection reset by peer)
- # [21:43] * Joins: jhk (jigneshhk1@8E6C34C1.A3F9767A.1C37C358.IP)
- # [21:45] * Joins: icaaq1 (Adium@moz-7D09279.cust.bredbandsbolaget.se)
- # [21:45] * Quits: icaaq (Adium@moz-7D09279.cust.bredbandsbolaget.se) (Connection reset by peer)
- # [21:49] * Quits: victorporo (victorporo@E25721DF.A00C134E.79933D60.IP) (Quit: Linkinus - http://linkinus.com)
- # [21:52] * Joins: peteb-away (ptbrunet@moz-E9B02845.austin.res.rr.com)
- # [21:59] * Quits: drexler (chatzilla@moz-6F22CAC7.e7jvsoe-gw.cust.sover.net) (Quit: ChatZilla 0.9.88-rdmsoft [XULRunner 1.9.0.17/2009122204])
- # [22:23] <@askalski> davidb, once I spoke with the guy from google on his testing experience
- # [22:23] <@askalski> he told me, that he started his adventure with getting a lot of historical data from test slaves
- # [22:23] <@davidb> makes sense
- # [22:24] <@askalski> does a11y has it's own test harness running on testbots, or am I building the first one to be deployed
- # [22:24] <@askalski> and mochitests is all we had so far
- # [22:26] <@askalski> because if we do have any historical data (especially mochitests) I wrote a script several weeks ago I could just run to find which tests did vary most over the time. that would mean that they are either very good or very bad tests :)
- # [22:27] <@davidb> yep i haven't forgotten that script
- # [22:27] <@davidb> in general we have bugs for failures that get comments with a freq relative to their happenings
- # [22:27] <@davidb> i'm on a phone meeting atm
- # [22:28] <@askalski> no problem
- # [22:29] <@askalski> just let me know if you ever hear about an option to get to historical data, I would love to dig in it for a moment, just to have in mind how they look like to have something to think about in spare time :)
- # [22:34] <@davidb> good to know :)
- # [22:34] <@davidb> askalski: are you off friday and monday?
- # [22:35] <@askalski> hmm, not sure yet. I have to write something for Uni and help family with holiday preparation
- # [22:35] <@askalski> if I find any free time I will check in :)
- # [22:36] <@askalski> davidb, and generally, if you don't mind, I prefer catching up some time during weekend - you know, gyms are closed, professors don't answer e-mails, perfect conditions to work :)
- # [22:37] <@askalski> *weekends
- # [22:37] <@askalski> as long as I do not have to synchronize of course
- # [22:37] <@davidb> sure
- # [22:37] <@askalski> and am not blocked
- # [22:40] <@askalski> you know what I love about working at Mozilla?
- # [22:40] <@askalski> the "no problem" policy. once I first applied, and told them I want to relocate a lot of times during first semester they just said "OK, we can do that", which is like a miracle to me
- # [22:42] <@askalski> this is a huuuge advantage, the company should advertise itself as "human corporation"
- # [22:42] <@askalski> or something like this
- # [22:42] <@askalski> (I hope it's not an idiom or something like "speclenium")
- # [22:44] <@davidb> :)
- # [22:44] * Quits: margle (margle@moz-9542A19B.dsl.mweb.co.za) (Quit: Textual IRC Client: http://www.textualapp.com/)
- # [22:44] <@davidb> it is all about 'getting it done'
- # [22:47] * Quits: jhk (jigneshhk1@8E6C34C1.A3F9767A.1C37C358.IP) (Ping timeout)
- # [22:49] * Joins: mwcampbell (matt@moz-94197894.ks.ok.cox.net)
- # [22:50] <mwcampbell> Is there an explanation somewhere of how a screen reader with a virtual buffer should treat the "application" ARIA role?
- # [22:50] <mwcampbell> What exactly is the role for?
- # [22:53] <mwcampbell> More specifically, if the body element of a document has the application role, should a screen reader deactivate its virtual buffer and let all keystrokes pass through to the page?
- # [22:57] <mwcampbell> This is apparently the behavior of NVDA in Firefox; I just want to make sure it's correct.
- # [23:02] * Joins: jhk (jigneshhk1@8E6C34C1.A3F9767A.1C37C358.IP)
- # [23:03] <clown> happy easter, davidb!
- # [23:03] * Joins: fxa90id (fxa90id@moz-83BBC510.neoplus.adsl.tpnet.pl)
- # [23:03] <@davidb> clown: you too!
- # [23:04] <@davidb> and to A
- # [23:04] <clown> ttfn
- # [23:04] <@davidb> ttgn
- # [23:04] <@davidb> errr g/f
- # [23:04] <clown> and to J
- # [23:04] <clown> too
- # [23:04] <@davidb> ok
- # [23:04] * clown forgets all davidb's initials.
- # [23:04] <@davidb> oh yes your J of course
- # [23:04] <clown> tah!
- # [23:04] <@davidb> N G and R :)
- # [23:05] <@firebot> markcapella@twcny.rr.com changed the Status on bug 740758 from NEW to ASSIGNED.
- # [23:05] <@firebot> Bug https://bugzilla.mozilla.org/show_bug.cgi?id=740758 nor, --, ---, markcapella, ASSI, dexpcom nsAccessible::GetValue
- # [23:05] <clown> and to your A B C D E F G H I J K L M N O P Q R S T U V W X Y Z (that should take care of it all).
- # [23:05] <@davidb> lol
- # [23:06] * Quits: clown (clown@67828CC7.C1A51174.9D42CF23.IP) (Quit: Leaving.)
- # [23:06] <@askalski> clown, "use regular expressions, Luke!"
- # [23:07] <@davidb> ok phone meeting done, time to spin in circles and turn into a short order cook!
- # [23:07] <@davidb> shazam!
- # [23:07] * Quits: @davidb (davidb@471D72E.2257F909.F30C9E9E.IP) (Quit: davidb)
- # [23:15] * Quits: mwcampbell (matt@moz-94197894.ks.ok.cox.net) (Quit: Leaving)
- # [23:16] * Quits: jhk (jigneshhk1@8E6C34C1.A3F9767A.1C37C358.IP) (Ping timeout)
- # [23:19] <@firebot> askalski@mozilla.com granted review for attachment 612503 on bug 741699.
- # [23:19] <@firebot> Bug https://bugzilla.mozilla.org/show_bug.cgi?id=741699 nor, --, ---, maxli, NEW, telemetry for uiautomation.dll
- # [23:20] * Quits: ehsan (ehsan@F2D29657.F60B0462.67AC9B1.IP) (Input/output error)
- # [23:21] * Quits: @askalski (akuda@moz-6A36EC49.ip.abpl.pl) (Quit: Wychodzi)
- # [23:23] * Joins: jhk (jigneshhk1@8E6C34C1.A3F9767A.1C37C358.IP)
- # [23:25] * khuey is now known as khuey|away
- # [23:55] <@firebot> hub@mozilla.com changed the Target Milestone on bug 740696 from --- to mozilla14.
- # [23:55] <@firebot> Bug https://bugzilla.mozilla.org/show_bug.cgi?id=740696 nor, P1, mozilla14, hub, NEW, [Mac] we have children that are expired.
- # [23:58] * Quits: peteb-away (ptbrunet@moz-E9B02845.austin.res.rr.com) (Ping timeout)
- # Session Close: Fri Apr 06 00:00:01 2012
The end :)