/irc-logs / mozilla / #accessibility / 2012-04-05 / end

Options:

  1. # Session Start: Thu Apr 05 00:00:00 2012
  2. # Session Ident: #accessibility
  3. # [00:04] * Quits: margle (margle@moz-9542A19B.dsl.mweb.co.za) (Quit: Textual IRC Client: http://www.textualapp.com/)
  4. # [00:09] * Joins: davidb (davidb@471D72E.2257F909.F30C9E9E.IP)
  5. # [00:09] * ChanServ sets mode: +qo davidb davidb
  6. # [00:12] * Quits: @davidb (davidb@471D72E.2257F909.F30C9E9E.IP) (Quit: davidb)
  7. # [00:21] <@firebot> maxli@maxli.ca requested feedback from eitan@monotonous. org for attachment 612359 on bug 670853.
  8. # [00:21] <@firebot> Bug https://bugzilla.mozilla.org/show_bug.cgi?id=670853 nor, --, ---, maxli, NEW, undetermined progressmeters should expose mixed state
  9. # [01:19] * reuben is now known as Guest
  10. # [01:24] * Guest is now known as reuben
  11. # [01:25] <@firebot> akeybl@mozilla.com set status-firefox14 to fixed on bug 740958.
  12. # [01:25] <@firebot> Bug https://bugzilla.mozilla.org/show_bug.cgi?id=740958 cri, --, mozilla14, surkov.alexander, RESO FIXED, crash in nsAccessible::GetBounds
  13. # [01:36] * Quits: drexler (chatzilla@moz-6F22CAC7.e7jvsoe-gw.cust.sover.net) (Quit: ChatZilla 0.9.88-rdmsoft [XULRunner 1.9.0.17/2009122204])
  14. # [01:46] * Quits: @jprmc (jprmc@F2D29657.F60B0462.67AC9B1.IP) (Ping timeout)
  15. # [02:13] * Quits: @hub (hub@21B7B9F2.B87E9213.6E712CE2.IP) (Ping timeout)
  16. # [02:19] <@firebot> maxli@maxli.ca requested review from surkov.alexander@gm ail.com for attachment 612409 on bug 735645.
  17. # [02:19] <@firebot> Bug https://bugzilla.mozilla.org/show_bug.cgi?id=735645 nor, --, ---, maxli, NEW, expose sub and sup elements in text attributes
  18. # [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)
  19. # [02:20] <Mark_Capella> is this a way to achieve the same results without changing the member declaration?
  20. # [02:22] <@tbsaunde> Mark_Capella: no
  21. # [02:22] <@tbsaunde> Mark_Capella: no
  22. # [02:24] <@tbsaunde> instead of if (foo->IsHTML()) static_cast<nsGenericHTMLElement*>(foo) you can do nsGenericHTMLElement* bar = nsGenericHTMLElement::FromContent(foo)
  23. # [02:24] <@tbsaunde> if (bar) return bar
  24. # [02:27] <Mark_Capella> so i can leave the member function GetURIAttr() protected / asis, and use this cast (which is public) instead yes?
  25. # [02:29] * Joins: jprmc (jprmc@78CFC4BC.5BCEC6DB.DA78B690.IP)
  26. # [02:29] * ChanServ sets mode: +o jprmc
  27. # [02:30] <@tbsaunde> Mark_Capella: no, casting to the right type has nothing to do with access to members
  28. # [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
  29. # [02:30] <Mark_Capella> oh wait .... the cast is ok though in the code i thought
  30. # [02:32] <@tbsaunde> the cast is fine, nsGeneriHTMLElement::FromContent() > inline static cast
  31. # [02:37] * Quits: ehsan (ehsan@F2D29657.F60B0462.67AC9B1.IP) (Input/output error)
  32. # [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
  33. # [02:39] <Mark_Capella> (ive tried this and it works)
  34. # [02:39] <Mark_Capella> what I dont understand is what would I need to use the FromContent() member for?
  35. # [02:42] * Joins: davidb (davidb@471D72E.2257F909.F30C9E9E.IP)
  36. # [02:42] * ChanServ sets mode: +qo davidb davidb
  37. # [02:43] <Mark_Capella> tbsaunde: (and we may have been talking about two different casting situations just then)
  38. # [02:43] * Quits: jimm (jmathies@moz-7F164CA1.pn.at.cox.net) (Quit: )
  39. # [02:44] <@tbsaunde> Mark_Capella: in nsHTMLImageAccessible::LongDesc() you have nsIContent* foo; then bar = static_cast<nsGenericHTMLElement*>(foo); right?
  40. # [02:45] <Mark_Capella> nope - several iterations of code back
  41. # [02:45] <Mark_Capella> nsGenericHTMLElement* element =
  42. # [02:45] <Mark_Capella> static_cast<nsGenericHTMLElement*>(mContent.get());
  43. # [02:45] <Mark_Capella> then element->GetURIAttr(nsGkAtoms::longdesc, nsnull, getter_AddRefs(uri));
  44. # [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
  45. # [02:46] * Joins: drexler (chatzilla@moz-BD26ECCD.hsd1.vt.comcast.net)
  46. # [02:47] <@tbsaunde> Mark_Capella: fine, the details aren't particularly interesting or important
  47. # [02:48] <Mark_Capella> ah ok - i thought you found a better way :)
  48. # [02:48] <@tbsaunde> somewhere in the patch you downcast from nsIContent to nsGenericHTMLEelement right?
  49. # [02:48] <@davidb> tbsaunde: heh
  50. # [02:48] * khuey is now known as khuey|away
  51. # [02:49] <@tbsaunde> Mark_Capella: and I've been saying instead of using static cast use nsGenericHTMLElement::FromContent()
  52. # [02:49] * Quits: @jprmc (jprmc@78CFC4BC.5BCEC6DB.DA78B690.IP) (Ping timeout)
  53. # [02:51] <Mark_Capella> and then to clarify, id still have to make GetURIAttr public yes?
  54. # [02:52] <@tbsaunde> yes
  55. # [02:52] <Mark_Capella> >whew< :P thanks tbsaunde
  56. # [02:52] <@tbsaunde> yw
  57. # [02:55] <@firebot> trev.saunders@gmail.com requested review from nfroyd@mozilla.com for attachment 612413 on bug 733510.
  58. # [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
  59. # [03:00] <@davidb> tbsaunde: I think the r+ carries forward for that one
  60. # [03:01] <@davidb> but leave it if you wish
  61. # [03:05] * Quits: drexler (chatzilla@moz-BD26ECCD.hsd1.vt.comcast.net) (Quit: ChatZilla 0.9.88-rdmsoft [XULRunner 1.9.0.17/2009122204])
  62. # [03:05] <@tbsaunde> davidb: hm, thought he didn't r+
  63. # [03:05] * @tbsaunde looks
  64. # [03:07] <@tbsaunde> davidb: yeah, guess so I guess I'll cancel and land
  65. # [03:07] <@davidb> yeah it looks right to me
  66. # [03:07] <@davidb> then i'll update my unknown consumer patch
  67. # [03:13] * Joins: ehsan (ehsan@F0B20A8D.8458880F.57F33CED.IP)
  68. # [03:14] <@firebot> trev.saunders@gmail.com cancelled review?(nfroyd@mozilla.com) for attachment 612413 on bug 733510.
  69. # [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
  70. # [03:26] <@davidb> tbsaunde: what is the weak reference bug aiming to fix exactly?
  71. # [03:27] <@tbsaunde> davidb: don't keep accessibles alive when they want to die
  72. # [03:27] <@davidb> but the weak accessible lives
  73. # [03:27] <@tbsaunde> and don't let ats keep gecko stuff alive
  74. # [03:27] <@davidb> admittedly it is smaller
  75. # [03:27] <@davidb> i see
  76. # [03:28] <@davidb> we should say this in the bug somewhere
  77. # [03:28] <@tbsaunde> what do you mean eak accessible?
  78. # [03:28] <@tbsaunde> I guess it doesn't hurt but who does it help?
  79. # [03:28] <@davidb> nsWeakAccessible is part of this picture right?
  80. # [03:29] <@davidb> it helps readers of the bug :)
  81. # [03:30] <@tbsaunde> davidb: no, never heard of nsWeakAccessible before
  82. # [03:30] <@davidb> it doesn't exist
  83. # [03:30] * khuey|away is now known as khuey
  84. # [03:30] <@davidb> I'm probably misunderstanding the bug
  85. # [03:30] <@tbsaunde> I guess, though I don't expect that many unfamiliar with a11y impl are going to go reading bugs for fun
  86. # [03:30] <@davidb> I was thinking this would be a setup like layout has for nsWeakFrame
  87. # [03:31] <@tbsaunde> probably something a like, but I'm not really sure how weak frames work
  88. # [03:31] <@davidb> well
  89. # [03:32] <@davidb> something like this
  90. # [03:32] <@tbsaunde> the purpose is similar, but I'm not sure we want to make the same trade offs they did
  91. # [03:32] <@davidb> here's a related change se http://hg.mozilla.org/mozilla-central/rev/f360de4b8b08
  92. # [03:32] <@davidb> set
  93. # [03:33] <@davidb> sample code
  94. # [03:33] <@davidb> nsWeakAccessible weakAccessible;
  95. # [03:33] <@davidb> weakAccessible = GetAccessible(foo);
  96. # [03:33] <@davidb> do stuff
  97. # [03:33] <@davidb> later on...
  98. # [03:34] <@davidb> if (weakAccessible.IsAlive())
  99. # [03:34] <@davidb> weakAccessible.getAccessible()->GetNameFoo()
  100. # [03:34] <@davidb> but I'm not clear if this is what surkov has in mind
  101. # [03:34] <@davidb> and IsAlive() returns false if the weakAccessible object has been cleared out
  102. # [03:35] <@davidb> presumable by the document or xpcom dying or something
  103. # [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
  104. # [03:37] <@davidb> yes
  105. # [03:38] <@tbsaunde> I think surkov's idea is keep a hash table of accessible -> weak refs
  106. # [03:38] <@tbsaunde> then when the accessible wants to die it goes and wakes all the weak refs
  107. # [03:38] * Joins: jprmc (jprmc@moz-7F2FF3EB.cpe.net.cable.rogers.com)
  108. # [03:38] * ChanServ sets mode: +o jprmc
  109. # [03:38] <@davidb> it sounds very similar
  110. # [03:39] <@davidb> wakes?
  111. # [03:39] <@tbsaunde> davidb: in purpose yes
  112. # [03:39] <@tbsaunde> *wacks
  113. # [03:39] <@davidb> hmm
  114. # [03:39] <@davidb> well it should clear out the weak references raw ptr to the accessible
  115. # [03:39] <@tbsaunde> I'm not completely sure its needed, but we'll see
  116. # [03:39] <@tbsaunde> davidb: yeah, that's what I meant by wack
  117. # [03:39] <@davidb> yeah I'll need to see more code
  118. # [03:40] <@davidb> ok
  119. # [03:40] <@davidb> you and your fancy science terms :)
  120. # [03:41] <@davidb> it is advised to use weak frames sparingly, since there is overhead
  121. # [03:41] <@davidb> i need to think about our case but i need to understand what surkov is wanting here first
  122. # [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
  123. # [03:41] <@davidb> this doesn't have to be complicated
  124. # [03:41] <@davidb> but it might be less performant
  125. # [03:41] <@davidb> that's my main concern
  126. # [03:42] <@davidb> it really depends what the idea is
  127. # [03:42] <@tbsaunde> using a weak ref will be slightly worse than a raw pointer, but shouldn't be bad
  128. # [03:42] <@davidb> usage, right.
  129. # [03:42] <@davidb> but there is more overhead
  130. # [03:42] <@tbsaunde> I'm not very concerned about perf
  131. # [03:42] * Joins: muralisr92 (chatzilla@moz-3BBB9F1C.dynip.nus.edu.sg)
  132. # [03:43] <@davidb> interesting
  133. # [03:43] <@tbsaunde> just a look up in a hash table so not much at all
  134. # [03:43] <@davidb> for each accessible
  135. # [03:44] <@tbsaunde> but I'm not completely sure how surkov plans to handle the ABA problem
  136. # [03:44] <@davidb> imagine a page mostly mutating
  137. # [03:44] <@davidb> ABA?
  138. # [03:45] <@davidb> i'm more worried about lifecycle
  139. # [03:45] <@tbsaunde> its a concurrency thing where a list item gets removed freed then reinserted
  140. # [03:45] <@tbsaunde> and someone else thinks its the original thing at that address not the new one
  141. # [03:45] <@davidb> OK so not the Alberta Bicycle Association
  142. # [03:45] <@davidb> ah ok
  143. # [03:46] <@davidb> that can get complicated
  144. # [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
  145. # [03:46] <@davidb> good
  146. # [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
  147. # [03:47] <@davidb> i don't see how we can do that
  148. # [03:47] <@tbsaunde> why?
  149. # [03:48] <@davidb> we can't give out raw ptrs to garbage
  150. # [03:48] <@davidb> keep in mind I'm thinking nsWeakAccessible
  151. # [03:48] <Mark_Capella> or the American Bankers Association
  152. # [03:48] <@tbsaunde> why would we be doing that
  153. # [03:48] <@davidb> Mark_Capella: or the American Basketball Assoc
  154. # [03:48] <Mark_Capella> well we all have our reference points :)
  155. # [03:48] <@davidb> tbsaunde: i think i need to stop thinking about it until i understand the plan
  156. # [03:49] <@davidb> Mark_Capella: weak
  157. # [03:49] <@davidb> :)
  158. # [03:49] <Mark_Capella> holy crap you type fast dude
  159. # [03:49] <Mark_Capella> and tell bad puns
  160. # [03:49] <@davidb> ha! i'm not fast.
  161. # [03:49] <@tbsaunde> davidb: I sort of thought comment 2 explained it well enough, but ok
  162. # [03:49] <@davidb> it doesn't for me
  163. # [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
  164. # [03:50] <@davidb> yeah I know it is you :)
  165. # [03:51] <@davidb> tbsaunde: what does "Make XPCOM object inherited from weak reference class" mean?
  166. # [03:51] <@davidb> this is silly - I should wait for surkov :)
  167. # [03:51] <@tbsaunde> but if it lets us stop refcounting accessibles then I suspect it'll be a perf win oerall
  168. # [03:51] <@davidb> his words
  169. # [03:51] <@davidb> absolutely
  170. # [03:52] <@davidb> now i wish we were having a work week
  171. # [03:52] <@tbsaunde> class xpcAccessibleTable : public WeakAcc { };
  172. # [03:52] <@davidb> ok
  173. # [03:53] <Mark_Capella> work weak
  174. # [03:53] <@tbsaunde> or maybe public WeakRefHolder
  175. # [03:53] <@davidb> WeakAcc is tbd
  176. # [03:53] <@davidb> ok
  177. # [03:53] <@davidb> Mark_Capella: groan
  178. # [03:53] <Mark_Capella> B)
  179. # [03:53] <@davidb> heheh
  180. # [03:53] <Mark_Capella> ok ill stop - gotta bail for a bit
  181. # [03:53] <@davidb> hey np
  182. # [03:57] <@davidb> tbsaunde: I look forward to September
  183. # [03:57] <@tbsaunde> davidb: make sense?
  184. # [03:57] <@tbsaunde> heh
  185. # [03:57] <@davidb> tbsaunde: yes it does.
  186. # [03:57] <@davidb> roughyl
  187. # [03:57] <@davidb> roughly
  188. # [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
  189. # [03:58] <@davidb> i think a member is clearer
  190. # [03:59] <@davidb> the wrapper stuff i need to think on
  191. # [03:59] <@tbsaunde> but possibly less useful
  192. # [03:59] <@davidb> in this case yes
  193. # [03:59] * Joins: drexler (chatzilla@moz-BD26ECCD.hsd1.vt.comcast.net)
  194. # [04:00] <@davidb> you gonna land the telem?
  195. # [04:00] <@tbsaunde> yeah, fairly soon on inbound
  196. # [04:00] <@davidb> nice
  197. # [04:01] <@davidb> ok i'm gonna check crash-stats and call it a night
  198. # [04:03] <@tbsaunde> ok
  199. # [04:03] * ctalbert is now known as ctalbert|afk
  200. # [04:03] <@tbsaunde> pushed :)
  201. # [04:03] * @davidb dances a little
  202. # [04:04] <@tbsaunde> heh
  203. # [04:04] <@tbsaunde> how does crash sts look?
  204. # [04:07] * Quits: ehsan (ehsan@F0B20A8D.8458880F.57F33CED.IP) (Input/output error)
  205. # [04:07] <@davidb> pretty good
  206. # [04:07] <@davidb> that get bounds fix seemed to take
  207. # [04:09] <@tbsaunde> yeah, saw bugmail
  208. # [04:11] <@davidb> hmmm i have an old patch to bail on dirty frames when creating an accessible
  209. # [04:11] <@davidb> i seem to recall the patch breaks us which is surprising - need to check that out again
  210. # [04:11] * @davidb is cleaning house
  211. # [04:11] * @davidb qdeletes
  212. # [04:15] <@davidb> ok cool I'm gonna try about put this laptop down.
  213. # [04:15] <@davidb> ciao!
  214. # [04:15] * Quits: @davidb (davidb@471D72E.2257F909.F30C9E9E.IP) (Quit: davidb)
  215. # [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
  216. # [04:24] * Joins: ehsan (ehsan@F0B20A8D.8458880F.57F33CED.IP)
  217. # [04:24] * Joins: aaronlev (aaronlev@BCFC8969.9EBA4B5B.A641CD9D.IP)
  218. # [04:38] * Joins: surkov (surkov@B9538DF7.B7015738.33A1AC3C.IP)
  219. # [04:38] * ChanServ sets mode: +o surkov
  220. # [05:27] <Mark_Capella> tbsaunde: yah - that fixed it
  221. # [05:28] <@tbsaunde> rk$/nick tbsaunde|afk
  222. # [05:28] * tbsaunde is now known as tbsaunde|afk
  223. # [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
  224. # [05:57] <@firebot> Bug https://bugzilla.mozilla.org/show_bug.cgi?id=678429 nor, --, ---, nobody, NEW, implement weak reference for accessible objects
  225. # [06:04] * Joins: jhk (jigneshhk1@8E6C34C1.A3F9767A.1C37C358.IP)
  226. # [06:05] <@surkov> Mark_Capella: yes, it's not hard but it should be very interesting
  227. # [06:05] * Quits: jhk (jigneshhk1@8E6C34C1.A3F9767A.1C37C358.IP) (Connection reset by peer)
  228. # [06:05] * Joins: jhk (jigneshhk1@8E6C34C1.A3F9767A.1C37C358.IP)
  229. # [06:06] <@firebot> markcapella@twcny.rr.com changed the Status on bug 678429 from NEW to ASSIGNED.
  230. # [06:07] * Quits: jhk (jigneshhk1@8E6C34C1.A3F9767A.1C37C358.IP) (Ping timeout)
  231. # [06:07] * Joins: jhk (jigneshhk1@8E6C34C1.A3F9767A.1C37C358.IP)
  232. # [06:08] * Quits: jhk (jigneshhk1@8E6C34C1.A3F9767A.1C37C358.IP) (Connection reset by peer)
  233. # [06:08] * Quits: muralisr92 (chatzilla@moz-3BBB9F1C.dynip.nus.edu.sg) (Quit: ChatZilla 0.9.88.1 [Firefox 11.0/20120310010316])
  234. # [06:08] * Joins: jhk (jigneshhk1@8E6C34C1.A3F9767A.1C37C358.IP)
  235. # [06:09] <@firebot> surkov.alexander@gmail.com requested review from trev.saunders@gmail .com for attachment 612328 on bug 740696.
  236. # [06:09] <@firebot> surkov.alexander@gmail.com granted feedback for attachment 612328 on bug 740696.
  237. # [06:09] <@firebot> Bug https://bugzilla.mozilla.org/show_bug.cgi?id=740696 nor, P1, ---, hub, NEW, [Mac] we have children that are expired.
  238. # [06:12] <@firebot> surkov.alexander@gmail.com cancelled feedback?(eitan@monotonous. org) for attachment 612038 on bug 670853.
  239. # [06:12] <@firebot> Bug https://bugzilla.mozilla.org/show_bug.cgi?id=670853 nor, --, ---, maxli, NEW, undetermined progressmeters should expose mixed state
  240. # [06:20] * tbsaunde|afk is now known as tbsaunde
  241. # [06:23] <@tbsaunde> surkov: David wanted me to look at bug 708927 any thoughts?
  242. # [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
  243. # [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.
  244. # [06:26] <@surkov> if log is not clear then probably it makes sense to improve it so it provides more info
  245. # [06:26] <@surkov> or you can enable debugging output on this test and wait for a failure
  246. # [06:29] <@tbsaunde> surkov: yeah, ok
  247. # [06:30] <@tbsaunde> surkov: what do you mean always trying to feed something/
  248. # [06:30] <@surkov> tbsaunde: finding a work
  249. # [06:34] <@tbsaunde> ah
  250. # [06:36] <@firebot> surkov.alexander@gmail.com granted review for attachment 612440 on bug 652635.
  251. # [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>
  252. # [06:39] <@firebot> trev.saunders@gmail.com changed the Target Milestone on bug 733510 from --- to mozilla14.
  253. # [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
  254. # [06:48] * khuey is now known as khuey|away
  255. # [06:58] <@firebot> surkov.alexander@gmail.com requested review from jonas@sicking.cc for attachment 612440 on bug 652635.
  256. # [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>
  257. # [07:01] * Quits: peteb-away (ptbrunet@moz-E9B02845.austin.res.rr.com) (Client exited)
  258. # [07:15] * Quits: drexler (chatzilla@moz-BD26ECCD.hsd1.vt.comcast.net) (Quit: ChatZilla 0.9.88-rdmsoft [XULRunner 1.9.0.17/2009122204])
  259. # [07:24] * tbsaunde is now known as tbsaunde|afk
  260. # [07:27] * Joins: hub (hub@moz-E2FCA694.figuiere.net)
  261. # [07:27] * ChanServ sets mode: +o hub
  262. # [07:28] * Joins: MarcoZ (Daily@moz-C3875970.dip.t-dialin.net)
  263. # [07:28] * ChanServ sets mode: +o MarcoZ
  264. # [07:30] <@firebot> surkov.alexander@gmail.com granted review for attachment 612409 on bug 735645.
  265. # [07:30] <@firebot> Bug https://bugzilla.mozilla.org/show_bug.cgi?id=735645 nor, --, ---, maxli, NEW, expose sub and sup elements in text attributes
  266. # [07:35] <@firebot> surkov.alexander@gmail.com changed the Target Milestone on bug 733510 from mozilla14 to ---.
  267. # [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
  268. # [07:36] * Quits: aaronlev (aaronlev@BCFC8969.9EBA4B5B.A641CD9D.IP) (Quit: aaronlev)
  269. # [07:45] * Quits: @firebot (firebot@moz-67BF9FE6.carolina.res.rr.com) (Ping timeout)
  270. # [08:00] * Joins: firebot (firebot@moz-67BF9FE6.carolina.res.rr.com)
  271. # [08:00] * ChanServ sets mode: +o firebot
  272. # [08:08] <@firebot> markcapella@twcny.rr.com cancelled review?(jonas@sicking.cc) for attachment 612440 on bug 652635.
  273. # [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>
  274. # [08:09] <Mark_Capella> ah drats
  275. # [08:10] <@firebot> markcapella@twcny.rr.com requested review from jonas@sicking.cc for attachment 612460 on bug 652635.
  276. # [08:26] <fxa90id> hi
  277. # [08:27] * Quits: fxa90id (fxa90id@moz-829BEFC9.neoplus.adsl.tpnet.pl) (Quit: Leaving)
  278. # [08:35] * Quits: @hub (hub@moz-E2FCA694.figuiere.net) (Ping timeout)
  279. # [08:40] * Joins: victorporof (victorporo@AAF12FE3.67CDC207.79933D60.IP)
  280. # [08:41] * Quits: victorporof (victorporo@AAF12FE3.67CDC207.79933D60.IP) (Ping timeout)
  281. # [08:57] * Joins: margle (margle@moz-9542A19B.dsl.mweb.co.za)
  282. # [09:07] * Joins: icaaq (Adium@moz-FCAABB4F.cust.telenor.se)
  283. # [09:10] * Quits: jhk (jigneshhk1@8E6C34C1.A3F9767A.1C37C358.IP) (Connection reset by peer)
  284. # [09:12] * Quits: margle (margle@moz-9542A19B.dsl.mweb.co.za) (Quit: Computer has gone to sleep.)
  285. # [09:16] * Joins: jhk (jigneshhk1@8E6C34C1.A3F9767A.1C37C358.IP)
  286. # [09:21] * Quits: icaaq (Adium@moz-FCAABB4F.cust.telenor.se) (Connection reset by peer)
  287. # [09:29] * Joins: margle (margle@moz-9542A19B.dsl.mweb.co.za)
  288. # [09:42] * Quits: margle (margle@moz-9542A19B.dsl.mweb.co.za) (Quit: Computer has gone to sleep.)
  289. # [09:52] * Joins: icaaq (Adium@moz-FCAABB4F.cust.telenor.se)
  290. # [10:50] * Joins: victorporof (victorporo@622F294A.8E4D8E22.4A6B528C.IP)
  291. # [10:52] <@firebot> surkov.alexander@gmail.com changed the Target Milestone on bug 735645 from --- to mozilla14.
  292. # [10:52] <@firebot> surkov.alexander@gmail.com granted in-testsuite on bug 735645.
  293. # [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
  294. # [10:54] <@firebot> surkov.alexander@gmail.com cancelled review?(trev.saunders@gmail .com) for attachment 612149 on bug 741398.
  295. # [10:54] <@firebot> surkov.alexander@gmail.com requested review from trev.saunders@gmail .com for attachment 612476 on bug 741398.
  296. # [10:54] <@firebot> Bug https://bugzilla.mozilla.org/show_bug.cgi?id=741398 nor, --, ---, surkov.alexander, ASSI, make ARIA state map extensible
  297. # [11:00] <@firebot> eitan@monotonous.org cancelled feedback?(l10n@mozilla.com) for attachment 609578 on bug 739498.
  298. # [11:00] <@firebot> Bug https://bugzilla.mozilla.org/show_bug.cgi?id=739498 nor, --, ---, eitan, NEW, Add Javascript layer for mobile accessibility
  299. # [11:07] <@firebot> eitan@monotonous.org granted feedback for attachment 612359 on bug 670853.
  300. # [11:07] <@firebot> Bug https://bugzilla.mozilla.org/show_bug.cgi?id=670853 nor, --, ---, maxli, NEW, undetermined progressmeters should expose mixed state
  301. # [11:17] * Quits: icaaq (Adium@moz-FCAABB4F.cust.telenor.se) (Ping timeout)
  302. # [11:19] * Mark_Capella is now known as Mark_Capella|naptime
  303. # [11:20] * Joins: icaaq (Adium@moz-FCAABB4F.cust.telenor.se)
  304. # [11:43] <@firebot> New Core - Disability Access APIs bug 742657 filed by surkov.alexander@gmail.com.
  305. # [11:44] <@firebot> Bug https://bugzilla.mozilla.org/show_bug.cgi?id=742657 nor, --, ---, nobody, NEW, move nsAccUtils::GetRoleMapEntry(nsINode *aNode) into nsARIAMap
  306. # [11:49] <@firebot> New Core - Disability Access APIs bug 742659 filed by surkov.alexander@gmail.com.
  307. # [11:49] <@firebot> Bug https://bugzilla.mozilla.org/show_bug.cgi?id=742659 nor, --, ---, nobody, NEW, move nsARIAMap::UniversalStatesFor to aria namespace
  308. # [12:11] * Joins: fxa90id (fxa90id@moz-83BBC510.neoplus.adsl.tpnet.pl)
  309. # [12:19] * Quits: jhk (jigneshhk1@8E6C34C1.A3F9767A.1C37C358.IP) (Connection reset by peer)
  310. # [12:19] * Joins: jhk (jigneshhk1@8E6C34C1.A3F9767A.1C37C358.IP)
  311. # [12:21] * Quits: icaaq (Adium@moz-FCAABB4F.cust.telenor.se) (Connection reset by peer)
  312. # [12:28] * tbsaunde|afk is now known as tbsaunde
  313. # [12:32] * Joins: icaaq (Adium@moz-FCAABB4F.cust.telenor.se)
  314. # [12:44] * Quits: icaaq (Adium@moz-FCAABB4F.cust.telenor.se) (Ping timeout)
  315. # [12:44] * Joins: icaaq (Adium@moz-FCAABB4F.cust.telenor.se)
  316. # [12:57] <@tbsaunde> surkov: so, why do we seem to think that uia COM will work on ARM but msaa ia2 won't?
  317. # [12:57] <@surkov> tbsaunde: via can be not COM
  318. # [12:57] <@surkov> not IAccessibleEx of course
  319. # [12:57] <@surkov> but native UIA impl
  320. # [12:59] * Quits: victorporof (victorporo@622F294A.8E4D8E22.4A6B528C.IP) (Connection reset by peer)
  321. # [12:59] <@tbsaunde> surkov: and how would we implement such a thing, and how would implementing IWhatever help that?
  322. # [12:59] <@surkov> tbsaunde: that IWhatever is just not com interface if I get tight
  323. # [13:01] * Quits: shorlander (shorlander@moz-853043D6.dhcp.insightbb.com) (Connection reset by peer)
  324. # [13:01] <@tbsaunde> surkov: how? isn't it defined to inherit from IUnknown?
  325. # [13:01] * Joins: shorlander (shorlander@moz-853043D6.dhcp.insightbb.com)
  326. # [13:01] <@tbsaunde> and how does getting control patterns work without QI?
  327. # [13:01] <@surkov> tbsaunde: well, yes, those interface are COM
  328. # [13:02] <@surkov> that's what I heard actually and didn't look deeply
  329. # [13:03] <@tbsaunde> surkov: I haven't looked deeply, but it sounds crazy at a glance
  330. # [13:03] <@surkov> tbsaunde: ok, let's remove that part
  331. # [13:03] <@tbsaunde> surkov: ok
  332. # [13:03] <@tbsaunde> surkov: also, am I crazy or don't we implement IServiceProvider on all accessibles already?
  333. # [13:04] <@surkov> we implement
  334. # [13:04] <@surkov> the point is to extend its implementation
  335. # [13:04] <@surkov> bad wording I guess
  336. # [13:04] <@tbsaunde> surkov: sure
  337. # [13:04] <@tbsaunde> jhk: ^
  338. # [13:05] <@tbsaunde> surkov: what is you reasoning behind using IAccessibleEx instead of answering WM_GGETOBJECT with a IRawFragmentProviderRoot or whatever?
  339. # [13:06] <@tbsaunde> I've looked a little, but hadn't come to a opinion
  340. # [13:06] <@surkov> tbsaunde: for tree crawling?
  341. # [13:06] <@tbsaunde> surkov: I don't understand what you mean
  342. # [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
  343. # [13:07] <@surkov> I think it works this way
  344. # [13:08] <@surkov> anyway just following recommendations
  345. # [13:08] <@surkov> btw, UIA provides own mechanism for events
  346. # [13:08] <@surkov> I"m not sure if WM_GETOBJECT is used there
  347. # [13:08] <@surkov> we have many questions more than answers
  348. # [13:09] <@tbsaunde> surkov: yeah, that is the way you do IAccessiblEx
  349. # [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 :)
  350. # [13:10] <@surkov> ok
  351. # [13:10] <@tbsaunde> but yes more questions than answers, and we don't actually need to decide now I think
  352. # [13:11] <@surkov> yep
  353. # [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.
  354. # [13:15] <jhk> tbsaunde: yes It can change as per requirements...
  355. # [13:15] <jhk> and surkov "let's remove " which part?
  356. # [13:16] * @tbsaunde find the idea of him correcting english rather funny
  357. # [13:17] <@surkov> jhk: about COM stuff on ARM devices
  358. # [13:19] <jhk> surkov: ok
  359. # [13:24] * Quits: icaaq (Adium@moz-FCAABB4F.cust.telenor.se) (Ping timeout)
  360. # [13:25] * Joins: icaaq (Adium@moz-FCAABB4F.cust.telenor.se)
  361. # [13:28] * Quits: icaaq (Adium@moz-FCAABB4F.cust.telenor.se) (Ping timeout)
  362. # [13:31] * Joins: icaaq (Adium@moz-FCAABB4F.cust.telenor.se)
  363. # [13:35] <@firebot> trev.saunders@gmail.com granted review for attachment 612328 on bug 740696.
  364. # [13:35] <@firebot> Bug https://bugzilla.mozilla.org/show_bug.cgi?id=740696 nor, P1, ---, hub, NEW, [Mac] we have children that are expired.
  365. # [13:36] * Quits: icaaq (Adium@moz-FCAABB4F.cust.telenor.se) (Quit: Leaving.)
  366. # [13:51] <@firebot> maxli@maxli.ca requested review from askalski@mozilla.co m for attachment 612503 on bug 741699.
  367. # [13:51] <@firebot> Bug https://bugzilla.mozilla.org/show_bug.cgi?id=741699 nor, --, ---, maxli, NEW, telemetry for uiautomation.dll
  368. # [14:02] <@surkov> tbsaunde: what does densify mean actually?
  369. # [14:04] * Joins: icaaq (Adium@moz-7D09279.cust.bredbandsbolaget.se)
  370. # [14:08] <@tbsaunde> surkov: de ns ify
  371. # [14:08] <@tbsaunde> surkov: so remove ns
  372. # [14:16] * Mark_Capella|naptime is now known as Mark_Capella
  373. # [14:20] <@tbsaunde> surkov: make sense?
  374. # [14:31] <@surkov> yes
  375. # [14:32] <@tbsaunde> surkov: think it makes sense to file more bugs to rename stuff now or do we have enough good first bugs?
  376. # [14:33] <@surkov> tbsaunde: I think we should have one always
  377. # [14:35] * Joins: peteb-away (ptbrunet@moz-E9B02845.austin.res.rr.com)
  378. # [14:35] <@tbsaunde> surkov: ok, what next? html or xul form controls?
  379. # [14:36] <@surkov> tbsaunde: base/nsARIAGridAccessible should be easy
  380. # [14:36] <@surkov> let's finish with base classes
  381. # [14:36] <@tbsaunde> surkov: yeah, you have to deal with the wrap class but sure
  382. # [14:36] <@surkov> right
  383. # [14:37] <@tbsaunde> what should the wrap class be named? ARIAGridWrap?
  384. # [14:38] <@surkov> ARIAGridAccessibleWrap?
  385. # [14:39] <@tbsaunde> and namespace it too right?
  386. # [14:39] <@surkov> yes
  387. # [14:40] <@tbsaunde> ok
  388. # [14:45] <@tbsaunde> surkov: don't you find it weird attrValuesArray is of pointers to pointers to atoms?
  389. # [14:45] <@firebot> surkov.alexander@gmail.com changed the Target Milestone on bug 670853 from --- to mozilla14.
  390. # [14:45] <@firebot> surkov.alexander@gmail.com granted in-testsuite on bug 670853.
  391. # [14:45] <@firebot> Bug https://bugzilla.mozilla.org/show_bug.cgi?id=670853 nor, --, mozilla14, maxli, NEW, undetermined progressmeters should expose mixed state
  392. # [14:45] <@surkov> tbsaunde: yep
  393. # [14:45] <@surkov> I do
  394. # [14:46] <@surkov> do we have other option?
  395. # [14:47] <@tbsaunde> surkov: not that I'm aware of
  396. # [14:47] <@surkov> same with me
  397. # [14:48] <@surkov> tbsaunde: btw, are you going to file next densify bug?
  398. # [14:48] <@tbsaunde> sure
  399. # [14:48] <@surkov> ok
  400. # [14:49] <@tbsaunde> just got distracted into reviewing the aria stuffs
  401. # [14:50] <@tbsaunde> surkov: one more question do we want to do aria grid and grid cell at same time or break them up?
  402. # [14:51] <@surkov> perhaps the same time
  403. # [14:51] <@tbsaunde> seems like it would be nice if they could have thee own files, but not sure that's really possible
  404. # [14:55] <@firebot> New Core - Disability Access APIs bug 742695 filed by trev.saunders@gmail.com.
  405. # [14:55] <@firebot> Bug https://bugzilla.mozilla.org/show_bug.cgi?id=742695 nor, --, ---, nobody, NEW, densify nsARIAGrid
  406. # [14:58] * Quits: fxa90id (fxa90id@moz-83BBC510.neoplus.adsl.tpnet.pl) (Quit: Leaving)
  407. # [15:14] * Joins: davidb (davidb@471D72E.2257F909.F30C9E9E.IP)
  408. # [15:14] * ChanServ sets mode: +qo davidb davidb
  409. # [15:19] * Quits: nhirata (nhirata.bu@moz-585D5EFC.hsd1.ca.comcast.net) (Quit: nhirata)
  410. # [15:22] * Quits: jhk (jigneshhk1@8E6C34C1.A3F9767A.1C37C358.IP) (Connection reset by peer)
  411. # [15:23] <@tbsaunde> surkov: why does nsAccessible.h need to include nsARIAMap.h? Role()?
  412. # [15:24] <@surkov> tbsaunde: some inlines
  413. # [15:25] <@firebot> trev.saunders@gmail.com granted review for attachment 612476 on bug 741398.
  414. # [15:25] <@firebot> Bug https://bugzilla.mozilla.org/show_bug.cgi?id=741398 nor, --, ---, surkov.alexander, ASSI, make ARIA state map extensible
  415. # [15:25] <@tbsaunde> surkov: ok, I wonder if its worth making a file accessible-inl.h or something for inline members
  416. # [15:26] <@surkov> that'd be good
  417. # [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
  418. # [15:28] <@tbsaunde> anyway got to run
  419. # [15:28] * tbsaunde is now known as tbsaunde|afk
  420. # [15:31] * Quits: ehsan (ehsan@F0B20A8D.8458880F.57F33CED.IP) (Input/output error)
  421. # [15:32] * Joins: aaronlev (aaronlev@BCFC8969.9EBA4B5B.A641CD9D.IP)
  422. # [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.
  423. # [15:32] <@davidb> He had questions about performance and whether we can have modes.
  424. # [15:32] <@davidb> (A topic we've discussed before of course)
  425. # [15:33] <@surkov> davidb: what was a problem?
  426. # [15:33] <@davidb> Anyways wanted you to know he's active on it right now - we might need to move forward faster than GSoC
  427. # [15:33] <@davidb> surkov: no specific problem yet, just concern over the engine instantiation for all tablets etc.
  428. # [15:34] <@davidb> soft keyboard usage
  429. # [15:59] <@firebot> markcapella@twcny.rr.com requested feedback from surkov.alexander@ gmail.com for attachment 612525 on bug 739612.
  430. # [15:59] <@firebot> Bug https://bugzilla.mozilla.org/show_bug.cgi?id=739612 nor, --, ---, markcapella, ASSI, Cleanup A11y tests and test-suite organization
  431. # [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?
  432. # [16:06] * Joins: clown (clown@67828CC7.C1A51174.9D42CF23.IP)
  433. # [16:06] * clown is now known as clown_mtg
  434. # [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.
  435. # [16:07] <@davidb> no rush
  436. # [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.
  437. # [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
  438. # [16:08] <@MarcoZ> I myself haven't seen any crashes lately, either. The ones I saw were ones that had been filed before.
  439. # [16:08] * Joins: jhk (jigneshhk1@8E6C34C1.A3F9767A.1C37C358.IP)
  440. # [16:08] <@MarcoZ> (and I skipped last weekend where we had the short-lived crash regression).
  441. # [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.
  442. # [16:09] <@davidb> MarcoZ: ok cool, is the community member testing nightly?
  443. # [16:10] <@MarcoZ> No, he's testing Aurora.
  444. # [16:12] <@MarcoZ> And the patches we up-lifted were ones he was seeing or had reported initially, too.
  445. # [16:12] <@surkov> davidb: actually I would love if you are involved from the start
  446. # [16:13] * clown_mtg is now known as clown
  447. # [16:14] <@davidb> surkov: then I will follow the bug.
  448. # [16:15] <@tbsaunde|afk> davidb: I'm curious why are you so interested?
  449. # [16:16] <@davidb> tbsaunde|afk: It gets my neurons firing.
  450. # [16:16] <@davidb> for some reason.
  451. # [16:16] <@tbsaunde|afk> heh
  452. # [16:16] <@davidb> MarcoZ: ok cool.
  453. # [16:17] * Joins: jimm (jmathies@moz-7F164CA1.pn.at.cox.net)
  454. # [16:18] <@MarcoZ> Hi jimm, welcome!
  455. # [16:18] <jimm> Hey MarcoZ, how ya doing?
  456. # [16:19] <@surkov> davidb: ok, cool
  457. # [16:19] <@davidb> jimm: is there a FF windows 8 roadmap/plan?
  458. # [16:20] <jimm> davidb: most everything we have thus far is on https://wiki.mozilla.org/Firefox/Windows_8_Integration
  459. # [16:20] <jimm> no official schedule yet
  460. # [16:21] <jimm> Asa has been working on getting resources together / planning
  461. # [16:21] <@davidb> OK
  462. # [16:21] <jimm> davidb: question for you, does accessibility expect all the calls into it to be on the main thread?
  463. # [16:22] <@davidb> except for plugins I think so
  464. # [16:22] <@davidb> (windowed plugins)
  465. # [16:22] <@davidb> surkov, tbsaunde|afk ^
  466. # [16:22] <@davidb> jimm: it gets murky for me around how COM works under the hood
  467. # [16:23] <jimm> the reason I ask is I ran into an off main thread exception in the cycle collector collecting an accessible object
  468. # [16:23] <@surkov> jimm: we actively working with frame tree which requires us to be called from main thread, right?
  469. # [16:23] <jimm> metro uses multiple threads for different things. thankfully UI rendering/events are all on a single thread.
  470. # [16:23] <@davidb> jimm: do we have a wiki of all the threads FF uses?
  471. # [16:23] <jimm> surkov: not sure
  472. # [16:23] <@davidb> I recall networking using threads
  473. # [16:23] <jimm> just got things hooked up and am debugging.
  474. # [16:24] <@davidb> nice!
  475. # [16:24] <jimm> davidb: no not yet
  476. # [16:24] <@surkov> mm, ok, I don't have answer then
  477. # [16:24] <jimm> davidb: we don't know what those threads are yet but that's a good idea
  478. # [16:25] * Joins: ehsan (ehsan@F2D29657.F60B0462.67AC9B1.IP)
  479. # [16:25] <jimm> davidb: the real main thread is owned by metro, we never see it except on startup
  480. # [16:25] <@davidb> We get calls from in-process injected dlls, as well as OOP
  481. # [16:25] <@davidb> interesting
  482. # [16:25] <jimm> there is a worker thread for rendering and events
  483. # [16:25] <jimm> there is also a suspend/resume thread for certain key events related to app state
  484. # [16:25] <@davidb> This is a notion of a privileged (or synonym?) thread in metro i hear?
  485. # [16:26] <@surkov> a11y relies on WM_GETOBJECT event
  486. # [16:26] * @davidb digs for notes
  487. # [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.
  488. # [16:26] <@davidb> surkov, yeah, winproc is main thread right?
  489. # [16:26] <@surkov> jimm said they use worker thread for events
  490. # [16:26] <jimm> in metro there is no wm_getobject because there is no window/window procedure
  491. # [16:26] <jimm> no hwnds either
  492. # [16:26] <@surkov> oh, I see
  493. # [16:27] <jimm> metro owns the window
  494. # [16:27] <jimm> we are embedded basically
  495. # [16:27] <@surkov> jimm: is there COM there?
  496. # [16:27] <@davidb> jimm: does metro manage a system wide a11y thread?
  497. # [16:27] <@davidb> or process
  498. # [16:27] * @davidb is confused
  499. # [16:27] <jimm> surkov: it's 100% COM under the hood
  500. # [16:27] <@surkov> ok
  501. # [16:28] <jimm> davidb: trying to figure that out now by itemizing threads / careful placement of IsMainThread debug output
  502. # [16:28] <@davidb> Mick (NVDA) said that their IPC tricks are broken in metro.
  503. # [16:28] <@davidb> cool
  504. # [16:29] <@davidb> funny how reverse engineering always trumps documentation
  505. # [16:29] <jimm> btw, I did manage to get UiaProviderFromIAccessible working.
  506. # [16:29] <@davidb> nice
  507. # [16:29] <jimm> the problem was a lack of an HWND for accessiblity from nsWindow's GetNativeData
  508. # [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:
  509. # [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.
  510. # [16:34] * @MarcoZ nods.
  511. # [16:36] <@firebot> surkov.alexander@gmail.com requested review from marco.zehe@googlema il.com for attachment 612525 on bug 739612.
  512. # [16:36] <@firebot> Bug https://bugzilla.mozilla.org/show_bug.cgi?id=739612 nor, --, ---, markcapella, ASSI, Cleanup A11y tests and test-suite organization
  513. # [16:37] <@firebot> surkov.alexander@gmail.com changed the Target Milestone on bug 739889 from --- to mozilla14.
  514. # [16:37] <@firebot> Bug https://bugzilla.mozilla.org/show_bug.cgi?id=739889 nor, --, mozilla14, markcapella, ASSI, densify nsOuterDocAccessible
  515. # [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
  516. # [16:39] <@surkov> davidb: they asked it for
  517. # [16:39] <@davidb> hmm. ok.
  518. # [16:39] <@davidb> i guess it now means: please merge this in trunk for 14
  519. # [16:39] <@davidb> instead of: this was merged for 14
  520. # [16:40] <@MarcoZ> davidb: Yes, it's a recent change in policy.
  521. # [16:41] <@davidb> ok
  522. # [16:41] <@davidb> i did notice it happening but never bothered to ask
  523. # [16:42] <@davidb> surkov: so you want to block the merge on bug 733510?
  524. # [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
  525. # [16:43] <@surkov> davidb: if everybody agree then we can file new patch into inbound
  526. # [16:44] <@davidb> surkov: I wouldn't bother, but if you feel strongly that's fine.
  527. # [16:44] <@davidb> Note we will continue to get data for the old names on different trains
  528. # [16:44] <@surkov> davidb: I think I can live with that but not sure why you decided to go this way
  529. # [16:44] <@davidb> so there will be a clash
  530. # [16:44] <@surkov> davidb: but those data is not useful?
  531. # [16:45] <@davidb> surkov: comment 8
  532. # [16:45] <@surkov> you said you can remove that data?
  533. # [16:45] <@davidb> yeah but i was thinking at that time it would stop coming in
  534. # [16:45] <@davidb> based on the old way
  535. # [16:45] <@surkov> I see
  536. # [16:45] <@davidb> but the same named data will keep coming in from different trains - which sounded like a headache for the metrics guys
  537. # [16:46] <@surkov> ok, so we should live with these names
  538. # [16:46] <@davidb> this way we keep all the data and can eol the old data later
  539. # [16:46] <@davidb> ok cool
  540. # [16:46] <@davidb> surkov: want to flip it? or me?
  541. # [16:47] <@surkov> flip it? what does it mean?
  542. # [16:47] <@davidb> set the target mileston
  543. # [16:47] <@davidb> e
  544. # [16:47] <@davidb> (you removed it)
  545. # [16:48] <@firebot> dbolter@mozilla.com changed the Target Milestone on bug 733510 from --- to mozilla14.
  546. # [16:49] <@davidb> surkov: anyways. done
  547. # [16:49] <@surkov> k
  548. # [16:55] <@davidb> ok
  549. # [16:56] <@davidb> surkov: I want to understand what you're asking for on bug 729154
  550. # [16:56] <@firebot> Bug https://bugzilla.mozilla.org/show_bug.cgi?id=729154 nor, --, ---, dbolter, NEW, Telemetry for a11y instantiation by unknown cause.
  551. # [16:56] <@surkov> davidb: comment #?
  552. # [16:56] <@davidb> surkov: 7
  553. # [16:57] <@surkov> davidb: merge enums
  554. # [16:57] <@davidb> you want a combined structure but i thought we decided that was less clear
  555. # [16:57] * khuey|away is now known as khuey
  556. # [16:58] <@surkov> I don't recall this
  557. # [16:59] <@davidb> surkov: it complicates code. (bug 678965)
  558. # [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
  559. # [16:59] <@surkov> davidb: comment #?
  560. # [16:59] <@davidb> e.g. comment 27
  561. # [17:00] <@davidb> Nathan advised doing it this clear way.
  562. # [17:00] <@surkov> mm, sorry, don't see how it's related with having one enum idea
  563. # [17:00] <@davidb> then i must not be understanding
  564. # [17:01] <@davidb> ok so you want to get rid of the bitflag?
  565. # [17:01] <@davidb> ?
  566. # [17:01] <@davidb> surkov: ^
  567. # [17:01] * Joins: margle (margle@moz-9542A19B.dsl.mweb.co.za)
  568. # [17:01] <@davidb> that must be it
  569. # [17:01] <@surkov> davidb: yep, it seems it's enough to have one enum, iirc tbsaunde agree
  570. # [17:01] <@davidb> ok
  571. # [17:01] <@davidb> i misunderstood "get rid telemetry enum"
  572. # [17:02] * Joins: nhirata (nhirata.bu@moz-C03D0C61.vlan426.asr1.sfo1.gblx.net)
  573. # [17:02] <@davidb> i can agree with this too
  574. # [17:02] <@davidb> surkov: still i'm not sure what it buys us
  575. # [17:03] <@surkov> davidb: it means merge it to modes enum, have one enum for both cases
  576. # [17:03] <@davidb> it merges "modes" and "consumers" into one enum - which to me does not clarify things
  577. # [17:03] <@davidb> i realize there is overlap, but the non-overlap is the problem i have
  578. # [17:03] * Joins: fxa90id (fxa90id@moz-83BBC510.neoplus.adsl.tpnet.pl)
  579. # [17:04] <fxa90id> surkov, yes
  580. # [17:04] <fxa90id> sry
  581. # [17:04] <@davidb> also, our mode code usage of the bit flag is sort of nice
  582. # [17:04] * Quits: margle (margle@moz-9542A19B.dsl.mweb.co.za) (Quit: Computer has gone to sleep.)
  583. # [17:04] * Quits: aaronlev (aaronlev@BCFC8969.9EBA4B5B.A641CD9D.IP) (Quit: aaronlev)
  584. # [17:05] <@firebot> marco.zehe@googlemail.com cancelled review?(marco.zehe@googlema il.com) for attachment 612525 on bug 739612.
  585. # [17:05] <@firebot> Bug https://bugzilla.mozilla.org/show_bug.cgi?id=739612 nor, --, ---, markcapella, ASSI, Cleanup A11y tests and test-suite organization
  586. # [17:05] <@surkov> davidb: well, we have enum for running ATs, some of them have specific modes
  587. # [17:05] <@davidb> yeah
  588. # [17:05] <@surkov> it's clear enough, for me
  589. # [17:06] <@davidb> surkov: in bug 678965 you didn't want the mode data reported - is that still the case?
  590. # [17:06] <@davidb> because this is going to get unnecessarily complex i think
  591. # [17:06] <@surkov> davidb: what is mode data?
  592. # [17:06] <@davidb> comment 27
  593. # [17:06] <@surkov> IA2Off?
  594. # [17:06] <@davidb> yeah
  595. # [17:07] <@surkov> yeah, you need to have enum like JAWSNumberBlaBla and then have IA2Off mode { if JAWSNumberBlabla then true; }
  596. # [17:07] <@davidb> i don't understand that :)
  597. # [17:08] <@davidb> i'm thinking about the reporting part
  598. # [17:08] <@davidb> right now it is very simple.
  599. # [17:10] * clown is now known as clown_mtg
  600. # [17:13] <@davidb> surkov: do you know what I mean?
  601. # [17:14] <@surkov> not sure
  602. # [17:16] <@davidb> darn
  603. # [17:16] <@davidb> well i want to land this so i guess i'll try the merge
  604. # [17:17] * Joins: askalski (akuda@moz-6A36EC49.ip.abpl.pl)
  605. # [17:17] * ChanServ sets mode: +o askalski
  606. # [17:20] * Joins: hub (hub@moz-E2FCA694.figuiere.net)
  607. # [17:20] * ChanServ sets mode: +o hub
  608. # [17:21] <@davidb> surkov: you still don't like the idea of the modes to be reported right?
  609. # [17:21] <@davidb> i just want to be clear
  610. # [17:22] <@surkov> davidb: it doesn't make sense to report modes because modes are turned off and on deeding on running AT
  611. # [17:22] <@surkov> davidb: about landing, you still don't have review?
  612. # [17:22] <@davidb> surkov: i'm not sure i understand the concern unless you are worried about noise?
  613. # [17:23] <@davidb> I don't have a peer review.
  614. # [17:23] <@surkov> reporting modes means duping information
  615. # [17:23] <@surkov> you need to get one
  616. # [17:23] <@davidb> surkov: you cancelled review :)
  617. # [17:23] <@surkov> davidb: because I wan't happy with patch approach I guess
  618. # [17:23] <@davidb> right
  619. # [17:24] <@davidb> so that's why i need to try the merge right?
  620. # [17:24] <@surkov> I don't r- if the patch is correct but not nice
  621. # [17:24] <@surkov> yeah, that should be nice
  622. # [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 :)
  623. # [17:24] <@surkov> and easy :)
  624. # [17:25] <@surkov> well, you need to be agree if you think the approach is right
  625. # [17:25] <@surkov> otherwise you shouldn't :)
  626. # [17:25] <@davidb> surkov: i might put the unreported modes up in high numbers to keep them out of the way
  627. # [17:25] <@davidb> we can't always agree
  628. # [17:25] <@davidb> looks at {}'s ;)
  629. # [17:25] <@surkov> we can :)
  630. # [17:25] <@davidb> we might not always agree expediently :)
  631. # [17:25] <@surkov> ok
  632. # [17:28] * Joins: victorporof (victorporo@40A61E2F.70F4578F.79933D60.IP)
  633. # [17:29] * Quits: Mark_Capella (chatzilla@moz-DD0C7E4F.twcny.res.rr.com) (Quit: ChatZilla 0.9.88.1 [Firefox 13.0a2/20120402042010])
  634. # [17:34] * Joins: Mark_Capella (chatzilla@moz-DD0C7E4F.twcny.res.rr.com)
  635. # [17:34] <@davidb> oh actually the way telemetry works I'm not sure what big gaps will do
  636. # [17:35] * @davidb wanders over to metrics
  637. # [17:36] <@davidb> actually shouldn't matter really
  638. # [17:37] <jimm> davidb: this is the kind of threading issues I'm seeing:
  639. # [17:37] <jimm> mozilla::widget::winrt::UIAutomationWrap::get_ProviderOptions: IsMainThread:1 ThreadId:10C4
  640. # [17:37] <jimm> mozilla::widget::winrt::UIAutomationWrap::get_ProviderOptions: IsMainThread:0 ThreadId:FDC
  641. # [17:37] <jimm> mozilla::widget::winrt::UIAutomationWrap::Navigate: IsMainThread:0 ThreadId:FDC
  642. # [17:37] <jimm> ###!!! ASSERTION: nsAccessNode not thread-safe: '_mOwningThread.GetThread() == PR_GetCurrentThread()', file m:/Mozilla/elm/accessible/src/base/nsAccessNode.cpp, line 85
  643. # [17:38] <@davidb> this is all pretty new to me
  644. # [17:38] <jimm> different threads calling into our uia wrapper at different times
  645. # [17:38] <jimm> which end up making calls into our IAccessible interface
  646. # [17:38] <jimm> yeah me to :)
  647. # [17:38] <@davidb> i don't recall seeing that assertion
  648. # [17:39] <jimm> I think on desktop everything happens on the main thread
  649. # [17:39] <jimm> I can test to see
  650. # [17:39] <@davidb> that would be good to ensure
  651. # [17:41] <@MarcoZ> davidb, surkov: Isn't nsAccessNode only used for iSimpleDOM?
  652. # [17:42] <@surkov> MarcoZ: correct
  653. # [17:42] <@davidb> but doesn't nsAccessible derive from nsAccessNodeWrap that derives from nsAccessNode?
  654. # [17:43] <@surkov> davidb: correct
  655. # [17:43] <@davidb> i guess the question is what line 85 is doing
  656. # [17:43] <@davidb> or that part of the code
  657. # [17:43] <@surkov> davidb: addref/release?
  658. # [17:44] <@surkov> yes, cycle collecting addref/release
  659. # [17:44] <@davidb> ok
  660. # [17:44] <@surkov> cycle collection wants to be on main thread
  661. # [17:44] <@davidb> makes sense
  662. # [17:45] <@surkov> but we want to split msaa and internal (cycle collected) implementation
  663. # [17:45] <@surkov> it's long term though
  664. # [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.
  665. # [17:45] <@surkov> and big work
  666. # [17:45] <@davidb> surkov: should this big work be a priority?
  667. # [17:45] <@surkov> well, we have many priorities
  668. # [17:46] <@davidb> agreed
  669. # [17:46] <@surkov> and not much resources
  670. # [17:46] <@davidb> it is an optimization problem :)
  671. # [17:46] <@surkov> we are trying to put this work into good first bugs so we are getting permanent progress no it
  672. # [17:46] <@surkov> not super fast though
  673. # [17:46] <@davidb> (or constraint satisfaction)
  674. # [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.
  675. # [17:47] <@surkov> davidb: note optimization takes a time too, sometimes significant time, and nobody can confirm it will work better
  676. # [17:47] <@surkov> in the end
  677. # [17:47] <@davidb> surkov: I meant' optimizating our resources.. .but anyways… jimm is here! :)
  678. # [17:47] <@surkov> jimm: yes, COM and ISupports have shared ref counter
  679. # [17:47] <@surkov> davidb: that's what I meant too
  680. # [17:48] <@davidb> ok
  681. # [17:48] <@davidb> (let's talk about that later)
  682. # [17:48] * Joins: margle (margle@moz-9542A19B.dsl.mweb.co.za)
  683. # [17:48] <@surkov> davidb: you're thinking that human resources optimization is easier than code optimization? :)
  684. # [17:48] <@surkov> sure
  685. # [17:48] <@davidb> heheh
  686. # [17:49] <@davidb> surkov, jimm so who would be calling us?
  687. # [17:49] <@surkov> uiaautomation.dll I think
  688. # [17:49] <@surkov> or how is it callled
  689. # [17:49] <@davidb> close enough
  690. # [17:49] <@davidb> but that's in process
  691. # [17:50] <@surkov> because we have similar problems with some screen readers on winxp
  692. # [17:50] <@davidb> so what are the characteristics (thread wise) of injected dlls calling us
  693. # [17:50] <@surkov> that's why we started that work
  694. # [17:50] * @davidb nods
  695. # [17:50] <jimm> yes uiaautomation.dll
  696. # [17:51] <@davidb> ok
  697. # [17:51] <jimm> we setup the wrapper on the main thread, then uia at some point starts inspecting the document using a different thread
  698. # [17:51] <jimm> although from mxr it looks like those asserts are passive so everything should still be succeeding
  699. # [17:52] <jimm> I'll break point one of these..
  700. # [17:53] <jimm> yeah should be passive - http://mxr.mozilla.org/mozilla-central/source/xpcom/glue/nsISupportsImpl.h#337
  701. # [17:54] <jimm> I wonder why instect can't see the tree
  702. # [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.
  703. # [17:55] <jimm> *The -> They
  704. # [17:55] <@davidb> surkov knows that stuff deeply
  705. # [17:56] <@surkov> jimm: HWND is primary thing that AT relies on
  706. # [17:57] <jimm> 'AT'?
  707. # [17:57] <@surkov> jimm: like screen readers
  708. # [17:57] <jimm> davidb: note the same asserts don't happen on the desktop. it's all the same thread there.
  709. # [17:57] <@davidb> ok thanks
  710. # [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.
  711. # [17:59] <jimm> surkov: since metro uses uia to do things like bring up the soft keyboard when a text edit is tapped.
  712. # [17:59] <@surkov> well, uia is critical thing for metro
  713. # [18:00] <jimm> surkov: initially when I tried to create a uia wrapper of our IAccessible using the following api - UiaProviderFromIAccessible
  714. # [18:00] <jimm> http://msdn.microsoft.com/en-us/library/windows/desktop/hh437315(v=VS.85).aspx
  715. # [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)
  716. # [18:01] <@surkov> I wonder how do they find an accessible at all, since currently they use WM_GETOBJECT to get top accessible object
  717. # [18:02] <jimm> I hand it to them via UiaProviderFromIAccessible
  718. # [18:02] <jimm> they hand back IRawElementProviderSimple that wrpas our IAccessible
  719. # [18:02] <@surkov> how do they call into us?
  720. # [18:02] <@surkov> or what do they call
  721. # [18:02] <@surkov> I mean how do they start accessibility at all?
  722. # [18:02] <jimm> through IRawElementProviderSimple and the other uia interfaces their generated wrapper supports.
  723. # [18:03] <jimm> http://pastebin.mozilla.org/1554960
  724. # [18:03] <jimm> that might not make a lot of sense but..
  725. # [18:03] <jimm> basically it's our wrapper around their wrapper which metro talks to
  726. # [18:04] <@surkov> I see
  727. # [18:04] <jimm> their wrapper wraps our IAccessible. :)
  728. # [18:05] <@davidb> it is like christmas
  729. # [18:05] <@davidb> and IAccessible is the unwanted fruitcake
  730. # [18:06] <@surkov> :)
  731. # [18:06] <jimm> that's why they gave us UiaProviderFromIAccessible
  732. # [18:06] <jimm> they knew a lot of client apps still used IAccessible
  733. # [18:07] <@davidb> including office
  734. # [18:08] <@davidb> jimm: if we have questions i have a contact - i will pm you
  735. # [18:08] <jimm> cool
  736. # [18:08] <jimm> anyway, I'm thinking these dummy windows I'm using are still causing problems.
  737. # [18:09] <@davidb> perhaps the relationship between the windows
  738. # [18:10] * Joins: drexler (chatzilla@moz-6F22CAC7.e7jvsoe-gw.cust.sover.net)
  739. # [18:12] <@davidb> surkov: are you still awake? i can't recall our dummy window situation… if we needed a parent-child window relationship
  740. # [18:13] <@surkov> davidb: sort of awake, I'm still not sure what fails
  741. # [18:13] <@surkov> jimm: was you able to get our accessible object?
  742. # [18:13] <jimm> sure
  743. # [18:13] <@surkov> but when you try to wrap it then you fail?
  744. # [18:14] <jimm> that part is working too now that widget is handing back hwnds from GetNativeData
  745. # [18:14] <jimm> but those hwnds are fakes, so maybe that is still breaking something.
  746. # [18:14] <jimm> in our accessibility code
  747. # [18:14] <@davidb> do the IRawElementProviderSimple methods work?
  748. # [18:14] <@surkov> I see
  749. # [18:15] <jimm> davidb: yes, metro is quite busy calling those when a touch display is present.
  750. # [18:15] <@davidb> and good things are returned?
  751. # [18:15] <jimm> and those calls definitely call into our accessibility codee
  752. # [18:15] <@davidb> nice ok
  753. # [18:16] <@surkov> btw, we don't create accessible for window at all, wonder if that could be a problem
  754. # [18:16] <jimm> I need to check return resuts but it seems as though these calls are working
  755. # [18:16] <jimm> hmm
  756. # [18:16] <jimm> surkov: when I call UiaProviderFromIAccessible
  757. # [18:17] <@surkov> so on desktop Windows create an accessible for window, when we handle WM_GETOBJECT for client then we return document accessible
  758. # [18:17] <@surkov> we hack its parent to return window accessible created by Windows
  759. # [18:17] <jimm> surkov: what do I get from this
  760. # [18:17] <jimm> :
  761. # [18:17] <jimm> nsAccessible *rootAccessible = mWidget->GetRootAccessible();
  762. # [18:18] <jimm> IAccessible *msaaAccessible = NULL;
  763. # [18:18] <jimm> rootAccessible->GetNativeInterface((void**)&msaaAccessible);
  764. # [18:18] <jimm> document?
  765. # [18:18] <@surkov> jimm: root document accessible
  766. # [18:18] <jimm> ok
  767. # [18:18] <jimm> when I call UiaProviderFromIAccessible I pass OBJID_CLIENT as the child id
  768. # [18:18] <jimm> that seems correct to me.
  769. # [18:18] <@surkov> hierarchy should look like: Window accessible object (created by Windows for Firefox window) -> root document accessible
  770. # [18:19] <jimm> yeah that looks right. I thyink we're setup that way.
  771. # [18:19] <Mark_Capella> surkov: re:Bug 678429 - implement weak reference for accessible objects
  772. # [18:19] <Mark_Capella> That's one heck of a thread you've got me following :)
  773. # [18:19] <@firebot> Bug https://bugzilla.mozilla.org/show_bug.cgi?id=678429 nor, --, ---, markcapella, ASSI, implement weak reference for accessible objects
  774. # [18:20] <@surkov> jimm: do you pass OBJIID_CLIENT for any accessible object?
  775. # [18:20] <@surkov> for every
  776. # [18:20] <@surkov> Mark_Capella: sorry, tbsaunde is guilty ;)
  777. # [18:20] <Mark_Capella> Lets Get im!!!
  778. # [18:20] <@surkov> )
  779. # [18:21] <Mark_Capella> Well you said easy but interesting ... Very!
  780. # [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.
  781. # [18:21] <@surkov> ok
  782. # [18:21] <@surkov> so presumingly metro is interested in root document accessible object
  783. # [18:22] <@surkov> than it does all magic on its side when it traverses the tree
  784. # [18:22] <jimm> yes I believe so
  785. # [18:22] <@surkov> I assume
  786. # [18:29] * @MarcoZ decides he is very, very scared now. :D
  787. # [18:31] <jimm> surkov: is nsAccessibleWrap the document object?
  788. # [18:32] <@surkov> jimm: no, nsDocAccessible is document object
  789. # [18:32] <@surkov> root document object is nsRootAccessible
  790. # [18:32] <@surkov> of course they are both inherited from nsAccessibleWrap
  791. # [18:32] <jimm> ah ok
  792. # [18:32] <jimm> so from GetNativeInterface() I get nsRootAccessible?
  793. # [18:32] <jimm> (on that first call)
  794. # [18:33] <@surkov> GetNativeInterface I think returns IUnknown from nsISupports object
  795. # [18:33] <@surkov> but mWidget->GetRootAccessible(); return nsRootAccessible
  796. # [18:34] <@surkov> or it should at least in non e10s builds
  797. # [18:34] * Quits: fxa90id (fxa90id@moz-83BBC510.neoplus.adsl.tpnet.pl) (Ping timeout)
  798. # [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
  799. # [18:41] * Quits: jhk (jigneshhk1@8E6C34C1.A3F9767A.1C37C358.IP) (Connection reset by peer)
  800. # [18:41] <jimm> so I'm assuming that's the top level nsDocAccessible
  801. # [18:41] * Joins: jhk (jigneshhk1@8E6C34C1.A3F9767A.1C37C358.IP)
  802. # [18:41] <jimm> which is what I wrap with our uia wrapper
  803. # [18:44] <@firebot> surkov.alexander@gmail.com changed the Target Milestone on bug 741398 from --- to mozilla14.
  804. # [18:44] <@firebot> Bug https://bugzilla.mozilla.org/show_bug.cgi?id=741398 nor, --, mozilla14, surkov.alexander, ASSI, make ARIA state map extensible
  805. # [18:49] * clown_mtg is now known as clown
  806. # [18:50] <jimm> surkov: you raise an interesting point - do we know if this works with e10s / fennec XUL's out of process tabs?
  807. # [18:51] <jimm> I guess it works on desktop fennec so it should work in metro.
  808. # [18:52] <@surkov> the difference on e10s is mWidget->GetRootAccessible() return nsDocAccessibleWrap instead nsRootAccessibleWrap but that shouldn't be a problem
  809. # [18:53] <@surkov> we were somehow accessible on e10s build
  810. # [18:53] <@surkov> not perfectly though
  811. # [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.
  812. # [18:55] <@surkov> ok
  813. # [19:02] * Quits: @hub (hub@moz-E2FCA694.figuiere.net) (Ping timeout)
  814. # [19:05] * Joins: victorporo (victorporo@E25721DF.A00C134E.79933D60.IP)
  815. # [19:05] * @davidb steps out
  816. # [19:05] * Quits: victorporof (victorporo@40A61E2F.70F4578F.79933D60.IP) (Ping timeout)
  817. # [19:10] * khuey is now known as khuey|away
  818. # [19:14] * ctalbert|afk is now known as ctalbert
  819. # [19:21] * Quits: @surkov (surkov@B9538DF7.B7015738.33A1AC3C.IP) (Quit: surkov)
  820. # [19:25] * Quits: @MarcoZ (Daily@moz-C3875970.dip.t-dialin.net) (Quit: MarcoZ)
  821. # [19:26] * Joins: hub (hub@21B7B9F2.B87E9213.6E712CE2.IP)
  822. # [19:26] * ChanServ sets mode: +o hub
  823. # [19:32] * Quits: @askalski (akuda@moz-6A36EC49.ip.abpl.pl) (Ping timeout)
  824. # [19:39] <@davidb> hi hub
  825. # [19:42] * Joins: askalski (akuda@moz-6A36EC49.ip.abpl.pl)
  826. # [19:42] * ChanServ sets mode: +o askalski
  827. # [19:43] * jimm is now known as jimm-lunch
  828. # [19:46] <@hub> 3 people in the Yahoo! a11y team were laid off
  829. # [19:50] <@davidb> hub: is this on the twitters? i don't check during day
  830. # [19:50] <@davidb> hub: who?
  831. # [19:52] <@davidb> oh 2000 jobs
  832. # [19:52] * khuey|away is now known as khuey
  833. # [19:52] <@firebot> ehsan@mozilla.com changed the Status on bug 734566 from ASSIGNED to RESOLVED.
  834. # [19:52] <@firebot> ehsan@mozilla.com set the Resolution field on bug 734566 to FIXED.
  835. # [19:52] <@firebot> Bug https://bugzilla.mozilla.org/show_bug.cgi?id=734566 nor, --, mozilla14, michaljev, RESO FIXED, optimize layout of TextAttrsMgr
  836. # [19:53] <@hub> davidb: jenisson twitted about it
  837. # [19:53] <@firebot> ehsan@mozilla.com changed the Status on bug 559746 from NEW to RESOLVED.
  838. # [19:53] <@firebot> ehsan@mozilla.com set the Resolution field on bug 559746 to FIXED.
  839. # [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">
  840. # [19:53] <@hub> davidb: amongst the 2000 layoffs
  841. # [19:54] <@hub> davidb: https://twitter.com/#!/yahooaccess/status/187857369269665792
  842. # [19:55] <@davidb> thanks
  843. # [20:01] <@davidb> tbsaunde|afk: if you can answer my latest on bug 729154 feel free
  844. # [20:01] <@firebot> Bug https://bugzilla.mozilla.org/show_bug.cgi?id=729154 nor, --, ---, dbolter, NEW, Telemetry for a11y instantiation by unknown cause.
  845. # [20:24] * jimm-lunch is now known as jimm
  846. # [20:25] <@askalski> davidb, while reviewing a patch, should I recompile to test it?
  847. # [20:26] <@davidb> askalski: for contributors it can be a good idea
  848. # [20:27] <@davidb> or generally if you suspect a compile issue
  849. # [20:27] <@askalski> davidb, sure
  850. # [20:27] <@davidb> it isn't required though
  851. # [20:27] <@davidb> and you can always mention you didn't compile it if you like
  852. # [20:28] <@davidb> before landing we compile locally and/or on the try server depending on perceived risk
  853. # [20:30] <@askalski> davidb, I definitely need to learn about try servers
  854. # [20:30] <@askalski> I noticed dozens of memes about it
  855. # [20:31] <@askalski> right now I compile locally all the time, and do mochitests locally as well
  856. # [20:31] <@askalski> btw, the machine is very strong
  857. # [20:31] <@davidb> that's what i do
  858. # [20:31] <@davidb> good :)
  859. # [20:32] <@firebot> ehsan@mozilla.com changed the Status on bug 733510 from NEW to RESOLVED.
  860. # [20:32] <@firebot> ehsan@mozilla.com set the Resolution field on bug 733510 to FIXED.
  861. # [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
  862. # [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
  863. # [20:34] <@askalski> davidb, how did the meeting with microsoft go? do we imlement UIA?
  864. # [20:34] <@hub> *sigh* some big stuff landed on inbound between my last update and the end of my build...
  865. # [20:35] <@davidb> telemetry flags?
  866. # [20:35] <@davidb> or more wide ranging i gues
  867. # [20:43] <jimm> davidb: question for you if you have a sec
  868. # [20:43] <@davidb> jimm: yes.
  869. # [20:43] <jimm> trying to understand the signifigance of this event sending:
  870. # [20:43] <jimm> http://mxr.mozilla.org/mozilla-central/source/accessible/src/msaa/nsAccessibleWrap.cpp#1558
  871. # [20:44] <@davidb> ok i'm weakish on our msaa layer but am reading...
  872. # [20:44] <jimm> I think that's for screen readers and the like
  873. # [20:44] <@davidb> yeah
  874. # [20:44] <jimm> they might hook our window to get at those events
  875. # [20:44] <jimm> ok
  876. # [20:44] <jimm> but we (mozilla) don't hook for those afaict
  877. # [20:45] <jimm> we just send them
  878. # [20:45] <@davidb> our own? no i don't think we do
  879. # [20:45] <jimm> ok cool
  880. # [20:46] <@firebot> ehsan@mozilla.com changed the Status on bug 735645 from NEW to RESOLVED.
  881. # [20:46] <@firebot> ehsan@mozilla.com set the Resolution field on bug 735645 to FIXED.
  882. # [20:46] <@davidb> our engine has no instances of calling SetWinEventHook
  883. # [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
  884. # [20:48] <@firebot> ehsan@mozilla.com changed the Status on bug 670853 from NEW to RESOLVED.
  885. # [20:48] <@firebot> ehsan@mozilla.com set the Resolution field on bug 670853 to FIXED.
  886. # [20:48] <@firebot> Bug https://bugzilla.mozilla.org/show_bug.cgi?id=670853 nor, --, mozilla14, maxli, RESO FIXED, undetermined progressmeters should expose mixed state
  887. # [21:12] <@firebot> markcapella@twcny.rr.com cancelled feedback?(surkov.alexander@ gmail.com) for attachment 612525 on bug 739612.
  888. # [21:12] <@firebot> Bug https://bugzilla.mozilla.org/show_bug.cgi?id=739612 nor, --, ---, markcapella, ASSI, Cleanup A11y tests and test-suite organization
  889. # [21:29] <@askalski> davidb, is uiautomation part of UIA?
  890. # [21:29] <@davidb> it is UIA
  891. # [21:29] <@askalski> davidb, good. just checking the names
  892. # [21:30] <@davidb> yep
  893. # [21:30] <@davidb> one api two purposes
  894. # [21:30] <@davidb> testing and a11y
  895. # [21:30] <@askalski> yep, it's a great way to promote a11y in business, really
  896. # [21:30] <@davidb> yeah
  897. # [21:30] <@davidb> the testing side of it has made it really stable
  898. # [21:30] <@davidb> mostly it is MS using it though
  899. # [21:30] <@davidb> s
  900. # [21:30] <@davidb> so far
  901. # [21:31] <@askalski> getting these two things bound on most popular platform is a huge advantage
  902. # [21:31] <@davidb> LDTP uses atk/atspi similarly
  903. # [21:31] <@askalski> LDTP?
  904. # [21:31] <@davidb> linux desktop testing project
  905. # [21:31] <@davidb> http://ldtp.freedesktop.org/wiki/
  906. # [21:32] <@davidb> but I don't mean to distract you
  907. # [21:32] <@askalski> I was reviewing a no-brainer patch, it loks very much the same as referenced one
  908. # [21:33] <@askalski> so I just made sure the names match, now I will check for compilation errors (very unlikely) and do r+
  909. # [21:33] <@davidb> the telemetry one? yeah
  910. # [21:33] <@davidb> looks good
  911. # [21:33] <@askalski> I guess there is no point in debugging to see if it triggers
  912. # [21:33] <@askalski> since I have no setup ready
  913. # [21:33] <@davidb> then no
  914. # [21:34] <@askalski> I mean - I have build environment running etc, but I don't know how to run debugger for ff in VS
  915. # [21:36] * Quits: peteb-away (ptbrunet@moz-E9B02845.austin.res.rr.com) (Ping timeout)
  916. # [21:36] <@davidb> for this patch eye-balling is ok
  917. # [21:36] <@davidb> make sure it isn't destructive
  918. # [21:43] * Quits: jhk (jigneshhk1@8E6C34C1.A3F9767A.1C37C358.IP) (Connection reset by peer)
  919. # [21:43] * Joins: jhk (jigneshhk1@8E6C34C1.A3F9767A.1C37C358.IP)
  920. # [21:45] * Joins: icaaq1 (Adium@moz-7D09279.cust.bredbandsbolaget.se)
  921. # [21:45] * Quits: icaaq (Adium@moz-7D09279.cust.bredbandsbolaget.se) (Connection reset by peer)
  922. # [21:49] * Quits: victorporo (victorporo@E25721DF.A00C134E.79933D60.IP) (Quit: Linkinus - http://linkinus.com)
  923. # [21:52] * Joins: peteb-away (ptbrunet@moz-E9B02845.austin.res.rr.com)
  924. # [21:59] * Quits: drexler (chatzilla@moz-6F22CAC7.e7jvsoe-gw.cust.sover.net) (Quit: ChatZilla 0.9.88-rdmsoft [XULRunner 1.9.0.17/2009122204])
  925. # [22:23] <@askalski> davidb, once I spoke with the guy from google on his testing experience
  926. # [22:23] <@askalski> he told me, that he started his adventure with getting a lot of historical data from test slaves
  927. # [22:23] <@davidb> makes sense
  928. # [22:24] <@askalski> does a11y has it's own test harness running on testbots, or am I building the first one to be deployed
  929. # [22:24] <@askalski> and mochitests is all we had so far
  930. # [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 :)
  931. # [22:27] <@davidb> yep i haven't forgotten that script
  932. # [22:27] <@davidb> in general we have bugs for failures that get comments with a freq relative to their happenings
  933. # [22:27] <@davidb> i'm on a phone meeting atm
  934. # [22:28] <@askalski> no problem
  935. # [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 :)
  936. # [22:34] <@davidb> good to know :)
  937. # [22:34] <@davidb> askalski: are you off friday and monday?
  938. # [22:35] <@askalski> hmm, not sure yet. I have to write something for Uni and help family with holiday preparation
  939. # [22:35] <@askalski> if I find any free time I will check in :)
  940. # [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 :)
  941. # [22:37] <@askalski> *weekends
  942. # [22:37] <@askalski> as long as I do not have to synchronize of course
  943. # [22:37] <@davidb> sure
  944. # [22:37] <@askalski> and am not blocked
  945. # [22:40] <@askalski> you know what I love about working at Mozilla?
  946. # [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
  947. # [22:42] <@askalski> this is a huuuge advantage, the company should advertise itself as "human corporation"
  948. # [22:42] <@askalski> or something like this
  949. # [22:42] <@askalski> (I hope it's not an idiom or something like "speclenium")
  950. # [22:44] <@davidb> :)
  951. # [22:44] * Quits: margle (margle@moz-9542A19B.dsl.mweb.co.za) (Quit: Textual IRC Client: http://www.textualapp.com/)
  952. # [22:44] <@davidb> it is all about 'getting it done'
  953. # [22:47] * Quits: jhk (jigneshhk1@8E6C34C1.A3F9767A.1C37C358.IP) (Ping timeout)
  954. # [22:49] * Joins: mwcampbell (matt@moz-94197894.ks.ok.cox.net)
  955. # [22:50] <mwcampbell> Is there an explanation somewhere of how a screen reader with a virtual buffer should treat the "application" ARIA role?
  956. # [22:50] <mwcampbell> What exactly is the role for?
  957. # [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?
  958. # [22:57] <mwcampbell> This is apparently the behavior of NVDA in Firefox; I just want to make sure it's correct.
  959. # [23:02] * Joins: jhk (jigneshhk1@8E6C34C1.A3F9767A.1C37C358.IP)
  960. # [23:03] <clown> happy easter, davidb!
  961. # [23:03] * Joins: fxa90id (fxa90id@moz-83BBC510.neoplus.adsl.tpnet.pl)
  962. # [23:03] <@davidb> clown: you too!
  963. # [23:04] <@davidb> and to A
  964. # [23:04] <clown> ttfn
  965. # [23:04] <@davidb> ttgn
  966. # [23:04] <@davidb> errr g/f
  967. # [23:04] <clown> and to J
  968. # [23:04] <clown> too
  969. # [23:04] <@davidb> ok
  970. # [23:04] * clown forgets all davidb's initials.
  971. # [23:04] <@davidb> oh yes your J of course
  972. # [23:04] <clown> tah!
  973. # [23:04] <@davidb> N G and R :)
  974. # [23:05] <@firebot> markcapella@twcny.rr.com changed the Status on bug 740758 from NEW to ASSIGNED.
  975. # [23:05] <@firebot> Bug https://bugzilla.mozilla.org/show_bug.cgi?id=740758 nor, --, ---, markcapella, ASSI, dexpcom nsAccessible::GetValue
  976. # [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).
  977. # [23:05] <@davidb> lol
  978. # [23:06] * Quits: clown (clown@67828CC7.C1A51174.9D42CF23.IP) (Quit: Leaving.)
  979. # [23:06] <@askalski> clown, "use regular expressions, Luke!"
  980. # [23:07] <@davidb> ok phone meeting done, time to spin in circles and turn into a short order cook!
  981. # [23:07] <@davidb> shazam!
  982. # [23:07] * Quits: @davidb (davidb@471D72E.2257F909.F30C9E9E.IP) (Quit: davidb)
  983. # [23:15] * Quits: mwcampbell (matt@moz-94197894.ks.ok.cox.net) (Quit: Leaving)
  984. # [23:16] * Quits: jhk (jigneshhk1@8E6C34C1.A3F9767A.1C37C358.IP) (Ping timeout)
  985. # [23:19] <@firebot> askalski@mozilla.com granted review for attachment 612503 on bug 741699.
  986. # [23:19] <@firebot> Bug https://bugzilla.mozilla.org/show_bug.cgi?id=741699 nor, --, ---, maxli, NEW, telemetry for uiautomation.dll
  987. # [23:20] * Quits: ehsan (ehsan@F2D29657.F60B0462.67AC9B1.IP) (Input/output error)
  988. # [23:21] * Quits: @askalski (akuda@moz-6A36EC49.ip.abpl.pl) (Quit: Wychodzi)
  989. # [23:23] * Joins: jhk (jigneshhk1@8E6C34C1.A3F9767A.1C37C358.IP)
  990. # [23:25] * khuey is now known as khuey|away
  991. # [23:55] <@firebot> hub@mozilla.com changed the Target Milestone on bug 740696 from --- to mozilla14.
  992. # [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.
  993. # [23:58] * Quits: peteb-away (ptbrunet@moz-E9B02845.austin.res.rr.com) (Ping timeout)
  994. # Session Close: Fri Apr 06 00:00:01 2012

The end :)