/irc-logs / mozilla / #accessibility / 2012-01-03 / end

Options:

  1. # Session Start: Tue Jan 03 00:00:00 2012
  2. # Session Ident: #accessibility
  3. # [01:32] * Joins: mdcurran (mick@moz-C48DC1F2.static.tpgi.com.au)
  4. # [03:19] * khuey is now known as khuey|away
  5. # [04:08] * khuey|away is now known as khuey
  6. # [04:08] * khuey is now known as khuey|away
  7. # [09:18] * Joins: jhk (jiggy@8E6C34C1.A3F9767A.1C37C358.IP)
  8. # [09:43] * Quits: @hub (hub@moz-E2FCA694.figuiere.net) (Ping timeout)
  9. # [10:13] * Joins: sgimeno (santiago.g@A922400D.48362F18.BCC03455.IP)
  10. # [10:24] * Quits: jhk (jiggy@8E6C34C1.A3F9767A.1C37C358.IP) (Connection reset by peer)
  11. # [10:25] * Joins: jhk (jiggy@8E6C34C1.A3F9767A.1C37C358.IP)
  12. # [10:42] * Joins: MarcoZ (marco.zehe@moz-6969D20B.dip.t-dialin.net)
  13. # [10:42] * ChanServ sets mode: +o MarcoZ
  14. # [11:24] * Quits: jhk (jiggy@8E6C34C1.A3F9767A.1C37C358.IP) (Ping timeout)
  15. # [11:45] * Joins: jhk (jiggy@8E6C34C1.A3F9767A.1C37C358.IP)
  16. # [11:52] * Quits: jhk (jiggy@8E6C34C1.A3F9767A.1C37C358.IP) (Connection reset by peer)
  17. # [13:04] * khuey|away is now known as khuey
  18. # [14:04] * Quits: mdcurran (mick@moz-C48DC1F2.static.tpgi.com.au) (Connection reset by peer)
  19. # [14:38] * Joins: jhk (jiggy@8E6C34C1.A3F9767A.1C37C358.IP)
  20. # [15:03] * Joins: davidb (davidb@moz-671E60DE.eng.wind.ca)
  21. # [15:03] * ChanServ sets mode: +qo davidb davidb
  22. # [15:13] * Joins: clown (clown@67828CC7.C1A51174.9D42CF23.IP)
  23. # [15:21] * Joins: davidb_ (davidb@F2D29657.F60B0462.67AC9B1.IP)
  24. # [15:22] <davidb_> firebot: jprmc?
  25. # [15:22] <@firebot> davidb_: hmm... I think jprmc is director of platform engineering, and drinks tea
  26. # [15:22] <davidb_> true
  27. # [15:22] <davidb_> hi all!
  28. # [15:25] <khuey> lol
  29. # [15:31] * Quits: @davidb (davidb@moz-671E60DE.eng.wind.ca) (Quit: bye)
  30. # [15:31] * davidb_ is now known as davidb
  31. # [15:32] * Parts: davidb (davidb@F2D29657.F60B0462.67AC9B1.IP)
  32. # [15:32] * Joins: davidb (davidb@F2D29657.F60B0462.67AC9B1.IP)
  33. # [15:32] * ChanServ sets mode: +qo davidb davidb
  34. # [15:32] * Parts: @davidb (davidb@F2D29657.F60B0462.67AC9B1.IP)
  35. # [15:32] * Joins: davidb (davidb@F2D29657.F60B0462.67AC9B1.IP)
  36. # [15:32] * ChanServ sets mode: +qo davidb davidb
  37. # [15:46] * Joins: jprmc (jprmc@F2D29657.F60B0462.67AC9B1.IP)
  38. # [15:46] * ChanServ sets mode: +o jprmc
  39. # [15:48] * Parts: jhk (jiggy@8E6C34C1.A3F9767A.1C37C358.IP)
  40. # [16:02] <@MarcoZ> Hi davidb! :)
  41. # [16:02] <@davidb> Hi!
  42. # [16:04] <@MarcoZ> davidb: Happy new year!
  43. # [16:05] <@davidb> Happy new year!
  44. # [16:08] <@davidb> MarcoZ: I PMed you something similar earlier :)
  45. # [16:17] <clown> happy new year, davidb
  46. # [16:27] <@davidb> clown: happy new year!
  47. # [16:35] <@MarcoZ> davidb: Need to step out for a bit to take Hexe to the vet. Something's not right with her.
  48. # [16:36] <@davidb> oh dear
  49. # [16:44] * Quits: clown (clown@67828CC7.C1A51174.9D42CF23.IP) (Quit: Leaving.)
  50. # [16:45] * Joins: richardschwerdtfeger (richardsch@moz-130F1494.lightspeed.austtx.sbcglobal.net)
  51. # [16:47] * Joins: hub (hub@moz-E2FCA694.figuiere.net)
  52. # [16:47] * ChanServ sets mode: +o hub
  53. # [16:48] * Joins: hhillen (hhillen@moz-2A42E9B3.upc-d.chello.nl)
  54. # [16:49] * Joins: clown (clown@67828CC7.C1A51174.9D42CF23.IP)
  55. # [17:02] * Joins: ehsan (ehsan@F2D29657.F60B0462.67AC9B1.IP)
  56. # [17:03] <@davidb> tbsaunde: are we on in 2?
  57. # [17:08] * clown is now known as clown_mtg
  58. # [17:23] * clown_mtg is now known as clown
  59. # [17:26] <@tbsaunde> davidb: ugh, completely forgot
  60. # [17:26] * @tbsaunde just woke up
  61. # [17:27] <@davidb> np
  62. # [17:28] <@tbsaunde> I can go grab a phone now if you want
  63. # [17:29] <@davidb> tbsaunde: actually let's hold off if that's ok?
  64. # [17:29] <@tbsaunde> sure :)
  65. # [17:30] <@davidb> happy new year!
  66. # [17:31] <@tbsaunde> same to yu :)
  67. # [17:54] <@hub> happy new year
  68. # [18:04] <@tbsaunde> MarcoZ: can you reproduce bug 714579?
  69. # [18:04] <@firebot> Bug https://bugzilla.mozilla.org/show_bug.cgi?id=714579 cri, --, ---, nobody, NEW, crash nsINode::OwnerDoc
  70. # [18:05] <@tbsaunde> davidb heard anything from surkov?
  71. # [18:13] * @eeejay waves happy twenty twelve
  72. # [18:20] <@davidb> tbsaunde: no
  73. # [18:20] <@davidb> happy new year everyone!
  74. # [18:20] <@tbsaunde> ok :
  75. # [18:21] <@tbsaunde> \\
  76. # [18:24] * Quits: hhillen (hhillen@moz-2A42E9B3.upc-d.chello.nl) (Quit: hhillen)
  77. # [18:29] <@MarcoZ> davidb: She caught a cold. Nothing serious, just inconvenient. ;)
  78. # [18:29] <@davidb> phew
  79. # [18:30] <@MarcoZ> tbsaunde: No I can't. I just got the report from a community member.
  80. # [18:31] <@tbsaunde> MarcoZ: ok
  81. # [18:50] <@davidb> MarcoZ: what is our crasher situation? what is highest priority?
  82. # [18:52] <@hub> what Andrzej nickname on IRC?
  83. # [18:52] <@davidb> hub: dunno yet
  84. # [18:52] <@davidb> he's an IRC noob
  85. # [18:53] <@hub> oh
  86. # [18:53] <@hub> ok
  87. # [18:53] <@davidb> he should be on tomorrow
  88. # [18:53] <@davidb> my guess would be akuda
  89. # [18:54] <@tbsaunde> davidb: who is he?
  90. # [18:55] <@davidb> "the new guy" :)
  91. # [18:55] <@davidb> tbsaunde: I sent you a bunch of emails
  92. # [18:56] <@tbsaunde> davidb: yeah, I know "the new guy" was wondering if you knew more, your mails didn't say much as far as I know
  93. # [18:57] <@davidb> tbsaunde: best to ask him, when he's on. He's finishing his masters in Warsaw. Will be working part time until finished.
  94. # [18:57] <@davidb> this new wiki has come in handy, ramping him up: https://wiki.mozilla.org/Accessibility/Contribute
  95. # [18:58] <@tbsaunde> fair enough
  96. # [19:11] * Quits: @hub (hub@moz-E2FCA694.figuiere.net) (Ping timeout)
  97. # [19:28] * Joins: hub (hub@21B7B9F2.B87E9213.6E712CE2.IP)
  98. # [19:28] * ChanServ sets mode: +o hub
  99. # [19:28] * @tbsaunde finds himself wanting access to crash dumps again :(
  100. # [19:29] <@hub> back
  101. # [19:31] <@tbsaunde> davidb: any news on that?
  102. # [19:33] <@davidb> tbsaunde: i have access
  103. # [19:34] <@davidb> tbsaunde: let's backchannel on that
  104. # [19:34] <@tbsaunde> sure
  105. # [19:39] <@MarcoZ> davidb: I personally am not seeing any crashers, but am getting reports occasionally from particularly one guy from somewhere in the States who reports stuff to me.
  106. # [19:39] <@MarcoZ> However nobody else seems to get them.
  107. # [19:39] <@davidb> ok
  108. # [19:39] <@MarcoZ> Anyway I'm off for the evening. See you all tomorrow!
  109. # [19:40] * Quits: @MarcoZ (marco.zehe@moz-6969D20B.dip.t-dialin.net) (Quit: l8er)
  110. # [19:45] * tbsaunde is now known as tbsaunde|afk
  111. # [20:56] <@davidb> eeejay: ping
  112. # [20:56] <@davidb> (no rush)
  113. # [20:56] <@eeejay> davidb, yo
  114. # [20:57] * tbsaunde|afk is now known as tbsaunde
  115. # [20:57] <@davidb> eeejay: now or later?
  116. # [20:57] <@eeejay> davidb, now is good, lemme find a room
  117. # [20:59] <@eeejay> davidb, could we do skype? no vidyo gadgets with me now
  118. # [21:08] * Quits: victorporof (victorporo@3A53A763.F695A43D.79933D60.IP) (Quit: Leaving...)
  119. # [21:09] <@davidb> yes :)
  120. # [21:09] <@davidb> tbsaunde: can you cc me on bug 714679?
  121. # [21:09] <@firebot> Bug https://bugzilla.mozilla.org/show_bug.cgi?id=714679 is not accessible
  122. # [21:10] <khuey> davidb: your security clearance isn't high enough
  123. # [21:11] <khuey> need to know only
  124. # [21:11] <@tbsaunde> davidb: I've never heard of that bug before
  125. # [21:11] <@tbsaunde> I may have said bug 714579 though
  126. # [21:12] <@firebot> Bug https://bugzilla.mozilla.org/show_bug.cgi?id=714579 cri, --, ---, nobody, NEW, crash nsINode::OwnerDoc
  127. # [21:12] <@davidb> ok!
  128. # [21:13] <@davidb> oh he's running his own build
  129. # [21:13] <@tbsaunde> I'm looking for the other one, but my network connection is terrible today
  130. # [21:13] <@davidb> khuey: :P
  131. # [21:14] <khuey> :-D
  132. # [21:14] <@tbsaunde> davidb: ?
  133. # [21:14] <@tbsaunde> at the running own build thing
  134. # [21:17] <@davidb> tbsaunde: ignore that
  135. # [21:18] <@tbsaunde> ok
  136. # [21:20] <@eeejay> tbsaunde, hey, could i ask a q?
  137. # [21:21] <@eeejay> tbsaunde, you suggesting using nsTObserverArray<nsCOMPtr<nsIAccessiblePivitObserver> >
  138. # [21:22] <@tbsaunde> I think that makes sense
  139. # [21:24] <@eeejay> tbsaunde, yeah, nm. i just answered my own question
  140. # [21:25] <@eeejay> tbsaunde, it is definately not how other nsTObserverArray users do it, they don't own references which looks kinda dangerous
  141. # [21:26] <@tbsaunde> eeejay: ok, ask jonas about it maybe? I've never delt with implenting a observer thing before in mozilla
  142. # [21:27] <@davidb> or smaug
  143. # [21:27] <@davidb> jonas is crazy busy
  144. # [21:28] * geo is now known as geo|pto
  145. # [21:28] <@tbsaunde> yeah, smorg would be good there too
  146. # [21:28] <@eeejay> tbsaunde, maybe smaug should also be the person to do the super review if jonas is so super buzy
  147. # [21:33] <@davidb> you could sr? the wind
  148. # [21:33] <@davidb> neil maybe
  149. # [21:33] <@davidb> he's fast
  150. # [21:34] <@davidb> and careful
  151. # [21:34] <@davidb> i'm only one of those two at any given time
  152. # [21:35] <@tbsaunde> neil might work, though I don't know him particularly well, but Alex sseems to like im
  153. # [21:35] * @tbsaunde kills his connection
  154. # [21:50] <@tbsaunde> eeejay: so, I see you and Alex talking about naming the document interface, but I don't think I see you talking about why a new interface makes sense instead of adding to nsIAccessibleDocument
  155. # [21:51] <@eeejay> tbsaunde, that might not be in the bug, but i is something alex insisted on
  156. # [21:51] <@tbsaunde> oic
  157. # [21:51] * @tbsaunde has no good idea why
  158. # [21:52] * @tbsaunde not particularly thrilled about the idea of making nsDocAccessible 2*sizeof(void*) larger
  159. # [21:53] <@eeejay> tbsaunde, https://bugzilla.mozilla.org/show_bug.cgi?id=698823#c27
  160. # [21:53] <@firebot> Bug 698823 nor, --, mozilla10, eitan, NEW, Introduce virtual cursor/soft focus functionality to a11y API
  161. # [21:54] <@eeejay> comments 27 through 29
  162. # [21:55] * Joins: victorporof (victorporo@3A53A763.F695A43D.79933D60.IP)
  163. # [21:55] <@davidb> note potential confusion with name class of the COM interface SID_IAccessibleContentDocument
  164. # [21:55] <@davidb> class/clash
  165. # [21:55] <@davidb> http://mxr.mozilla.org/mozilla-central/source/accessible/src/msaa/nsAccessNodeWrap.cpp#139
  166. # [21:57] <@davidb> also, not an interface… just an ID
  167. # [21:59] <@tbsaunde> eeejay: I tink his confussion concern is a little silly, but I guess I'll live with it
  168. # [21:59] <@tbsaunde> so I guess the best option is a tear off
  169. # [22:00] <@tbsaunde> davidb: unless you have a better idea
  170. # [22:00] <@eeejay> tbsaunde, compromise is how we get stuff done
  171. # [22:00] <@eeejay> tbsaunde, what is a tearoff?
  172. # [22:00] <@davidb> new C++ thinger
  173. # [22:00] <@eeejay> davidb, we abandoned that name
  174. # [22:00] <@davidb> is there any harm to adding to nsIAccessibleDocument?
  175. # [22:01] <@eeejay> davidb, alex don't like it
  176. # [22:01] <@davidb> ok
  177. # [22:01] <@davidb> what about as a middle step?
  178. # [22:01] <@tbsaunde> davidb: alex thinks its confusing
  179. # [22:01] <@eeejay> davidb, a middle step to where?
  180. # [22:01] <@tbsaunde> eeejay: so, QueryInterface returns a pointer of the right type right?
  181. # [22:02] * @davidb thinks
  182. # [22:02] <@eeejay> tbsaunde, yeah?
  183. # [22:02] <@tbsaunde> now, usually you get that pointer by static_cast<IFoo*>(this)
  184. # [22:02] <@davidb> hmm
  185. # [22:03] <@tbsaunde> but, there's no rule that says you can't do say *aOutInterface = static_cast<IFoo*>(new baz())
  186. # [22:03] <@eeejay> tbsaunde, so implement it in QueryInterface?
  187. # [22:03] <@eeejay> ah.. i get it
  188. # [22:04] <@eeejay> gawd that is a headache for my feeble c++ brain
  189. # [22:04] <@tbsaunde> eeejay: not reall implemented there, you have some other class of whose type you have an object and return a pointer to
  190. # [22:05] <@eeejay> tbsaunde, but then you are returning a new object that you cannot QueryInterface to another supported interface of the first object
  191. # [22:05] <@eeejay> unless you hold a cyclical reference...
  192. # [22:07] <@tbsaunde> eeejay: do you have to be able to do that? I don't actually remember if that's required
  193. # [22:07] <@eeejay> tbsaunde, well, the expected behavior of QueryInterface is not to switch instances
  194. # [22:07] <khuey> that's only sort of true
  195. # [22:08] <@tbsaunde> practically though since this isn't ms com and its an implementation detail I don't think it matters
  196. # [22:08] <@eeejay> but to provide a different interface on the same instance
  197. # [22:08] <khuey> QueryInterface really only has to return the same pointer for nsISupports
  198. # [22:08] <khuey> for other things it can return different pointers each time
  199. # [22:08] <khuey> at least in theory
  200. # [22:08] <@eeejay> khuey, but it would mean that all of these distinct objects would need to hold references to the original nsisupports pointer
  201. # [22:09] <khuey> right
  202. # [22:09] <khuey> that's how tearoffs work
  203. # [22:09] <@eeejay> and the nsisupports instance would need to hold references to these special instances
  204. # [22:09] <khuey> no
  205. # [22:09] <khuey> it can just create a new one each time
  206. # [22:10] <@eeejay> khuey, when two are compared, they automatically compare the nsisupports instance?
  207. # [22:11] <khuey> no
  208. # [22:11] <khuey> it's technically illegal to compare two non-nsISupports pointers for equality
  209. # [22:11] <khuey> in practice we do this all over the place, and only implement a few interfaces on tearoffs where this is known to be ok
  210. # [22:12] <@davidb> interesting
  211. # [22:12] <@eeejay> khuey, could you point me at a tearoff example in the mozcode?
  212. # [22:13] <khuey> http://mxr.mozilla.org/mozilla-central/source/content/html/content/src/nsGenericHTMLElement.cpp#303
  213. # [22:14] <@eeejay> ooooh macros
  214. # [22:14] <khuey> yeah we like those
  215. # [22:14] <@eeejay> that suddenly makes it practical :)
  216. # [22:15] <@eeejay> ok, i am going to eat lunch and then implement this tearoff. thanks khuey!
  217. # [22:16] <@eeejay> tbsaunde, you sure the overhead of another interface on the document is so bad?
  218. # [22:16] <@tbsaunde> eeejay: note that I believe you need to cache the tearoff here
  219. # [22:16] <@eeejay> tbsaunde, seems like you could mostly count those instances on your fingers (and toes)
  220. # [22:17] <@tbsaunde> eeejay: I don't think the number is that small, but I doubt its huge
  221. # [22:18] <@tbsaunde> another concern is that eventually we'd really like nsAccessible to be a xpcom object
  222. # [22:18] <@eeejay> it is (top level)+(open tabs)+iframes
  223. # [22:18] <@tbsaunde> *not*
  224. # [22:18] <@eeejay> tbsaunde, how would that affect things?
  225. # [22:19] <@tbsaunde> eeejay: the object has to have space for the pointer and the vtable even if you can't QI to it
  226. # [22:20] <@eeejay> tbsaunde, which object?
  227. # [22:20] <@tbsaunde> btw does your test have different results than the one we use to decide if it should be a root accessible?
  228. # [22:20] <@tbsaunde> nsDocAccessible
  229. # [22:20] <@tbsaunde> which to be fair appears a lot larger than I thought it was
  230. # [22:21] <@eeejay> tbsaunde, don't understand your test question
  231. # [22:21] <@eeejay> tbsaunde, i understand that nsDocAccessible gets larger
  232. # [22:22] <@eeejay> but like i said earlier, in the overall scheme it is not that much
  233. # [22:22] <@tbsaunde> eeejay: I mean your check in nsDocAccessible to see if the doc should have a virtual curser / pivit
  234. # [22:22] <@eeejay> tbsaunde, yeah, we check that
  235. # [22:22] <@eeejay> tbsaunde, and have an iframe where we expect to not have one
  236. # [22:23] <@tbsaunde> mind explaining the context of that? I don't get it
  237. # [22:23] <@eeejay> sorry, i thought you meant the mochitests
  238. # [22:24] <@eeejay> tbsaunde, the test in nsDocAccessible is similar to the top-level test elsewhere, as far as i remember.
  239. # [22:24] <@tbsaunde> eeejay: so, would it perhaps make more sense for only nsRootAccessible to be able to QI to your interface
  240. # [22:25] <@tbsaunde> and only have the one type of check?
  241. # [22:26] <@eeejay> tbsaunde, it is not just root accessibles, but top-level tab documents as well
  242. # [22:27] <@tbsaunde> ok so I'm not sure what you mean by top evel tab document
  243. # [22:27] <@eeejay> tbsaunde, top-level html documents
  244. # [22:28] <@eeejay> tbsaunde, ie. all documents except for embedded frames
  245. # [22:29] <@eeejay> tbsaunde, i'm going to head out for lunch, be back in 20
  246. # [22:31] <@tbsaunde> oh, ok
  247. # [22:33] * Quits: @davidb (davidb@F2D29657.F60B0462.67AC9B1.IP) (Quit: davidb)
  248. # [22:50] <@eeejay> back!
  249. # [22:51] <@eeejay> tbsaunde, so you still think a tearoff is the way to go?
  250. # [22:52] <@tbsaunde> eeejay: well, I don't really like it, but I think I can live with adding an interface to nsDocAccessible for now
  251. # [22:52] <@eeejay> tbsaunde, cool
  252. # [22:53] * Parts: clown (clown@67828CC7.C1A51174.9D42CF23.IP)
  253. # [22:53] <@tbsaunde> mostly because I don't know how exactly the world without nsAccessibleFoo being a xpcom object will look like
  254. # [22:55] <@eeejay> tbsaunde, does this new world mean this stuff won't be scriptable?
  255. # [22:56] <@tbsaunde> eeejay: there'll be something implementing nsIAccessible etc
  256. # [22:56] <@tbsaunde> but it'll be a wrapper or something
  257. # [22:57] <@eeejay> gotcha, just not directly from nsAccessible
  258. # [22:57] <@tbsaunde> yeah
  259. # [22:57] <@tbsaunde> and if I'm really really lucky we can stop ref counting accessibles, and so kill IsDefunct()
  260. # [22:59] <@tbsaunde> eeejay: for sr? the usual thing to do is poke [s]r?er on irc or bug comment, but sr? neil would work too
  261. # [23:01] <@eeejay> tbsaunde, i don't mind. should i do the poking? maybe since smaug already looked at stuff he should do it
  262. # [23:04] <@eeejay> tbsaunde, which neil?
  263. # [23:04] <@tbsaunde> eeejay: yeah, you should poke
  264. # [23:04] <@eeejay> deakin?
  265. # [23:04] <@tbsaunde> rashbrook I think
  266. # [23:05] <@eeejay> tbsaunde, neal deakin?
  267. # [23:05] <@tbsaunde> but honestly I don't know the difference between them :p
  268. # [23:05] <@tbsaunde> eeejay: yeah, there's another neil rashbrook who Alex usually asks
  269. # [23:05] <@tbsaunde> which was who I meant
  270. # [23:06] <@tbsaunde> eeejay: smorg might be busy with cycle collector improvement, I'm not really sure
  271. # [23:07] <@eeejay> can't find this rashbrook dude in bz or in the phonebook
  272. # [23:08] <khuey> he's not an employee
  273. # [23:08] <@eeejay> khuey, can't find his bz email either
  274. # [23:08] <khuey> firebot: NeilAway?
  275. # [23:08] <@firebot> khuey: Neil Rashbrook is mailto:neil@parkwaycc.co.uk (bugmail only to neil@httl.net), foolishly trying to remove all demons of stupidity from Gecko code. If he's been disconnected from IRC by an NNTP alert then try MemoServ.
  276. # [23:09] <@eeejay> hah, thanks
  277. # [23:10] <@firebot> eitan@monotonous.org cancelled superreview?(jonas@sicking. cc) for attachment 585060 on bug 698823.
  278. # [23:10] <@firebot> eitan@monotonous.org requested review from trev.saunders@gmail .com for attachment 585550 on bug 698823.
  279. # [23:10] <@firebot> eitan@monotonous.org requested superreview from neil@httl.net for attachment 585550 on bug 698823.
  280. # [23:10] <@firebot> Bug https://bugzilla.mozilla.org/show_bug.cgi?id=698823 nor, --, mozilla10, eitan, NEW, Introduce virtual cursor/soft focus functionality to a11y API
  281. # [23:21] <@firebot> trev.saunders@gmail.com granted review for attachment 585092 on bug 698823.
  282. # [23:21] <@firebot> Bug https://bugzilla.mozilla.org/show_bug.cgi?id=698823 nor, --, mozilla10, eitan, NEW, Introduce virtual cursor/soft focus functionality to a11y API
  283. # [23:37] <@firebot> eitan@monotonous.org granted review for attachment 585559 on bug 698823.
  284. # [23:37] <@firebot> Bug https://bugzilla.mozilla.org/show_bug.cgi?id=698823 nor, --, mozilla10, eitan, NEW, Introduce virtual cursor/soft focus functionality to a11y API
  285. # [23:38] <@tbsaunde> eeejay: we don't usually set r+ like that
  286. # [23:39] <@eeejay> tbsaunde, how is it done?
  287. # [23:39] <@tbsaunde> eeejay: just upload a new patch without requesting review
  288. # [23:39] <@eeejay> gotcha
  289. # [23:39] <@tbsaunde> if someone cares who r+ed it they can look at all the attachments
  290. # [23:40] <@tbsaunde> hopefully Neil will take a look too
  291. # [23:47] <@tbsaunde> eeejay: making the DocAccessible a pivit observer means it will need to be cycle collected, which isn't really a good thing
  292. # [23:48] <@eeejay> tbsaunde, it is the main use of this whole observer pattern
  293. # [23:48] <@tbsaunde> is there a reason you can't either crawl the chain of documents from mRoot or keep a week pointer from construction?
  294. # [23:49] <@eeejay> tbsaunde, not sure what you mean
  295. # [23:50] <@eeejay> tbsaunde, we could have pivot's root optionally implement the observer interface, if it is, it is notified
  296. # [23:50] <@tbsaunde> I mena in addition to notifying the observers figure out what the top level document is and fire the event with FireDelayedEvent()
  297. # [23:50] <@eeejay> yeah, no need to figure anything out, the document is the pivot's root
  298. # [23:50] <@tbsaunde> oh, yeah, ugh your already having that effect :\
  299. # [23:51] <@eeejay> what effect?
  300. # [23:51] <@eeejay> oh, cycle collection.. yeah
  301. # [23:51] <@tbsaunde> the document needing to be cycle collected
  302. # [23:51] <@tbsaunde> though I guess thats sort of ok if we shut it down first
  303. # [23:51] <@eeejay> does it make a diff if it is referenced twice? once as root and once as observer?
  304. # [23:52] <@tbsaunde> not really
  305. # [23:52] <@tbsaunde> although its a little silly and extra work
  306. # [23:52] <@eeejay> not really
  307. # [23:55] <@tbsaunde> eeejay: there's an array mapping our events to windows ones in msaa/nsEventmap.h you need to update too
  308. # [23:58] <@eeejay> tbsaunde, added
  309. # [23:58] <@eeejay> it is a big problem when you add stuff to base and it affects platform code
  310. # [23:59] <@tbsaunde> what, why?
  311. # Session Close: Wed Jan 04 00:00:01 2012

The end :)