/irc-logs / mozilla / #accessibility / 2012-02-21 / end
Options:
- # Session Start: Tue Feb 21 00:00:00 2012
- # Session Ident: #accessibility
- # [00:06] <@firebot> andrew.quartey@gmail.com cancelled review?(surkov.alexander@gm ail.com) for attachment 598573 on bug 691267.
- # [00:06] <@firebot> andrew.quartey@gmail.com requested review from surkov.alexander@gm ail.com for attachment 598980 on bug 691267.
- # [00:06] <@firebot> Bug https://bugzilla.mozilla.org/show_bug.cgi?id=691267 nor, --, ---, andrew.quartey, NEW, HTML input button and HTML 4 button should share the same class
- # [01:01] * khuey is now known as khuey|away
- # [01:02] * Joins: Jamie (jamie@moz-CA26021.jantrid.net)
- # [01:52] * Quits: ehsan (ehsan@F2D29657.F60B0462.67AC9B1.IP) (Input/output error)
- # [02:17] * Quits: Jamie (jamie@moz-CA26021.jantrid.net) (Ping timeout)
- # [02:21] * Joins: Jamie (jamie@moz-CA26021.jantrid.net)
- # [02:36] * shorlander is now known as shorlander-away
- # [02:37] * Quits: shorlander-away (shorlander@moz-853043D6.dhcp.insightbb.com) (Quit: Quit)
- # [02:46] * Joins: tty234 (telex@moz-F9058B8A.net)
- # [02:49] * khuey|away is now known as khuey
- # [03:11] * Joins: surkov (surkov@703D4713.1344B664.5D3F4C44.IP)
- # [03:11] * ChanServ sets mode: +o surkov
- # [03:19] * Quits: Jamie (jamie@moz-CA26021.jantrid.net) (Ping timeout)
- # [03:19] * Quits: @hub (hub@21B7B9F2.B87E9213.6E712CE2.IP) (Ping timeout)
- # [03:24] * Joins: Jamie (jamie@moz-CA26021.jantrid.net)
- # [03:33] <@firebot> surkov.alexander@gmail.com granted review for attachment 598929 on bug 726097.
- # [03:33] <@firebot> Bug https://bugzilla.mozilla.org/show_bug.cgi?id=726097 blo, --, ---, hub, NEW, Hit testing broken
- # [04:15] * Quits: Jamie (jamie@moz-CA26021.jantrid.net) (Quit: leaving)
- # [04:38] * Quits: jlin (Jonathan@moz-619C4F64.x.rootbsd.net) (Quit: ZNC - http://znc.in)
- # [04:45] * Joins: hub (hub@moz-E2FCA694.figuiere.net)
- # [04:45] * ChanServ sets mode: +o hub
- # [04:54] * Joins: shorlander (shorlander@moz-853043D6.dhcp.insightbb.com)
- # [05:32] * Quits: drexler (chatzilla@moz-BD26ECCD.hsd1.vt.comcast.net) (Quit: ChatZilla 0.9.88 [Firefox 10.0.2/20120215223356])
- # [05:41] * Joins: Jamie (jamie@moz-CA26021.jantrid.net)
- # [06:16] * khuey is now known as khuey|away
- # [06:26] <@firebot> hub@mozilla.com changed the Target Milestone on bug 726097 from --- to mozilla13.
- # [06:26] <@firebot> Bug https://bugzilla.mozilla.org/show_bug.cgi?id=726097 blo, --, mozilla13, hub, NEW, Hit testing broken
- # [06:29] <@firebot> New Core - Disability Access APIs bug 729006 filed by hub@mozilla.com.
- # [06:29] <@firebot> Bug https://bugzilla.mozilla.org/show_bug.cgi?id=729006 nor, --, ---, nobody, NEW, Cleanup hit test
- # [08:18] * Quits: Jamie (jamie@moz-CA26021.jantrid.net) (Quit: leaving)
- # [08:42] * Joins: jhk (jiggy@8E6C34C1.A3F9767A.1C37C358.IP)
- # [08:53] * Quits: jhk (jiggy@8E6C34C1.A3F9767A.1C37C358.IP) (Quit: Leaving.)
- # [08:53] * Joins: jhk (jiggy@8E6C34C1.A3F9767A.1C37C358.IP)
- # [08:58] * Joins: MarcoZ (marco.zehe@moz-72E20CAA.dip.t-dialin.net)
- # [08:59] * ChanServ sets mode: +o MarcoZ
- # [09:04] * Quits: jhk (jiggy@8E6C34C1.A3F9767A.1C37C358.IP) (Ping timeout)
- # [09:05] * Joins: jhk (jiggy@8E6C34C1.A3F9767A.1C37C358.IP)
- # [09:06] * Joins: Mark_Capella (chatzilla@moz-DD0C7E4F.twcny.res.rr.com)
- # [09:32] * Joins: sgimeno (santiago.g@A922400D.48362F18.BCC03455.IP)
- # [09:48] * Quits: Mark_Capella (chatzilla@moz-DD0C7E4F.twcny.res.rr.com) (Quit: ChatZilla 0.9.88 [Firefox 13.0a1/20120220104313])
- # [09:48] * Quits: jhk (jiggy@8E6C34C1.A3F9767A.1C37C358.IP) (Connection timed out)
- # [10:08] * Quits: @hub (hub@moz-E2FCA694.figuiere.net) (Ping timeout)
- # [10:21] <@firebot> surkov.alexander@gmail.com granted review for attachment 598980 on bug 691267.
- # [10:21] <@firebot> Bug https://bugzilla.mozilla.org/show_bug.cgi?id=691267 nor, --, ---, andrew.quartey, NEW, HTML input button and HTML 4 button should share the same class
- # [10:25] * Joins: askalski (akuda@moz-104CC309.mv.mozilla.com)
- # [10:25] * ChanServ sets mode: +o askalski
- # [10:26] <@askalski> morning everyone
- # [10:27] <@tbsaunde> late evening askalski
- # [10:28] <@askalski> tbsaunde, :D
- # [10:29] <@askalski> tbsaunde, we definitely should rent a boat and take a cruise through carribean together. We would arrange shifts without changing our habits
- # [10:30] <@tbsaunde> lol
- # [10:33] * Quits: @askalski (akuda@moz-104CC309.mv.mozilla.com) (Ping timeout)
- # [10:35] * Joins: askalski (akuda@moz-104CC309.mv.mozilla.com)
- # [10:35] * ChanServ sets mode: +o askalski
- # [10:47] * Quits: @askalski (akuda@moz-104CC309.mv.mozilla.com) (Ping timeout)
- # [10:47] * Joins: askalski (akuda@moz-104CC309.mv.mozilla.com)
- # [10:47] * ChanServ sets mode: +o askalski
- # [10:53] <@MarcoZ> Hi askalski, tbsaunde!
- # [10:54] * Quits: @askalski (akuda@moz-104CC309.mv.mozilla.com) (Ping timeout)
- # [11:48] * Joins: jhk (jiggy@8E6C34C1.A3F9767A.1C37C358.IP)
- # [11:48] * Joins: askalski (akuda@moz-104CC309.mv.mozilla.com)
- # [11:48] * ChanServ sets mode: +o askalski
- # [11:57] <@askalski> yes! I googled the bypass to Ukkonen!
- # [11:58] <@askalski> yeehaww! looks like several people on the planet were looking for the stuff too, it's pretty recent
- # [11:58] <@askalski> oh irony, I did it by following links from blog of a guy, who made some tremendous effort to recruit me to microsoft bing
- # [11:59] <@askalski> (for a huge pile of money I actually miss nowadays...)
- # [12:00] * Quits: jhk (jiggy@8E6C34C1.A3F9767A.1C37C358.IP) (Ping timeout)
- # [12:01] * Joins: jhk (jiggy@8E6C34C1.A3F9767A.1C37C358.IP)
- # [12:03] <@askalski> and now I am reading algorithms from 2010........
- # [12:36] <@askalski> surkov, I have updated on bug
- # [12:36] <@askalski> surkov, Ukonnen is a dead end, but I googled the alternative, I am on it noe
- # [12:36] <@askalski> surkov, ok?
- # [12:36] <@surkov> yep
- # [12:39] <@surkov> askalski: do I understand right you work on O(n + d^d) now?
- # [12:39] <@askalski> surkov, yes. you can look into code and see the exact single while loop I am trying to wrap into o(1) time
- # [12:40] <@askalski> surkov, now I know 2-3 ways to do it using suffix arrays
- # [12:40] <@askalski> now I learn how to build suffix array in O(n) time
- # [12:40] <@surkov> and you have O(md)?
- # [12:40] <@askalski> instead of suffix tree
- # [12:40] <@askalski> yes
- # [12:41] <@surkov> is m << d? so md < d^d?
- # [12:41] <@surkov> what is m btw :)
- # [12:41] <@askalski> m = max(m,n)
- # [12:41] <@askalski> d = edit distance
- # [12:41] <@surkov> sounds reqursive
- # [12:41] <@askalski> reqursive?
- # [12:41] <@surkov> m = max(m, n)
- # [12:42] <@surkov> recursive
- # [12:42] <@askalski> well, I stated at the very beginning, this algorithm is very fast if edit distance is small
- # [12:42] <@askalski> but slow in case of totally different strings
- # [12:43] <@askalski> I guess tbsaunde told me to try with that
- # [12:43] <@surkov> mm, can we have both and run depending on string length?
- # [12:43] <@askalski> anyway, the key argument I guess is that that's how diff is mostly implemented
- # [12:43] <@askalski> surkov, no problem
- # [12:43] <@surkov> ok
- # [12:43] <@askalski> I implemented both just to test correctness
- # [12:43] <@surkov> great
- # [12:44] <@askalski> there is also #3 variant, that breaks as soon as certain difference is reached
- # [12:44] <@askalski> for decision problem
- # [12:44] <@askalski> I can write it as well
- # [12:44] <@surkov> is it faster then string competence like we do now?
- # [12:44] <@surkov> comparison
- # [12:45] <@surkov> now we trim coinciding string parts and do a difference for different parts
- # [12:45] <@askalski> empirically, diff based on this approach is faster than based on alg that we have right now
- # [12:46] <@surkov> ok
- # [12:46] <@askalski> can you further explain "now we..." ?
- # [12:46] <@askalski> because I approached bug as "optimise algorithm", and maybe, big picture, further improvements are possible
- # [12:46] <@surkov> we get two strings, then we cut common parts from beginning and the end and then run alg
- # [12:47] <@askalski> in the one I wrote as p-o-c it's not necessary
- # [12:47] <@surkov> what is p-o-c?
- # [12:47] <@askalski> proof-of-concept
- # [12:48] <@askalski> I posted in in bug
- # [12:48] <@askalski> bugzilla
- # [12:48] <@askalski> it's not necessary, because the alg will jump over similar sections immediately
- # [12:48] <@surkov> so this thing is faster than running through string chars?
- # [12:48] <@surkov> so it should be the same
- # [12:48] <@askalski> no, it's the same fast as trimming prefixes and suffixes, it's just not necessary step in this alg
- # [12:49] <@surkov> extra memory, no?
- # [12:49] <@askalski> less.
- # [12:49] <@askalski> linear
- # [12:49] <@askalski> O(min(n,m))
- # [12:50] <@askalski> now it's not yet implemented, but can be done easily with some indexing magic I already know how to write
- # [12:50] <@surkov> mm, I wonder how we can spend less memory if we just run through string chars
- # [12:50] <@askalski> embedding what I have now will take me several hours anyway
- # [12:50] <@askalski> less than linear?
- # [12:51] <@surkov> yeah :)
- # [12:51] <@askalski> but God, why!? It'll kill the performance
- # [12:51] <@surkov> in cycle we run by string chars, If I get right then you create some structure
- # [12:52] <@askalski> you need to write more clearly so I can understand. what cycle?
- # [12:52] <@askalski> ah, right, slide() function?
- # [12:52] <@surkov> nah, I'm not about your algorithm
- # [12:52] <@surkov> I talk about strings cutting :)
- # [12:53] <@surkov> I mean I don't if that's a slide() function because I didn't look at the code yet
- # [12:53] <@askalski> ah. well, as I said, it's not necessary step, but if you want to, I can keep it
- # [12:53] <@askalski> but I really think
- # [12:53] <@askalski> that it's not necessary
- # [12:53] <@surkov> yeah, right and I asked about memory :)
- # [12:53] <@surkov> anyway, let's see, if it's not necessary then we can remove it
- # [12:53] <@askalski> because I told you max memory usage
- # [12:53] <@askalski> I guess, if that's the case, when common prefixes and suffixes are long
- # [12:54] <@askalski> the alg will stop early
- # [12:54] <@surkov> I think in most cases strings are the same
- # [12:54] <@surkov> maybe in 90%
- # [12:54] <@askalski> then alg will run linearly
- # [12:54] <@askalski> it's linear as long as d <= sqrt(max(m,n))
- # [12:55] <@askalski> and memory usage is really cut compared to what is done now
- # [12:55] <@surkov> you talk about main alg I assume
- # [12:55] <@askalski> yes
- # [12:55] <@surkov> and that's great, that's what we wanted
- # [12:55] <@surkov> speed and memory I mean
- # [12:55] <@askalski> ok. so can i spend another day on this O(mn) - > O(n+d^2) transition?
- # [12:56] <@askalski> it's still a long way until it's a patch, even if I run with what I have now
- # [12:56] <@askalski> O(dn), not O(mn).
- # [12:56] <@askalski> (computation time all)
- # [12:56] <@surkov> askalski: if you feel hot on this then go ahead
- # [12:57] <@surkov> but since O(n+d^2) is extra alg, then we could split the work on parts
- # [12:57] <@askalski> surkov, true.
- # [12:58] <@askalski> surkov, although I guess I'm more focused now
- # [12:58] <@surkov> yeah, that makes sense, it's always bad to interrupt the wokr
- # [12:58] <@surkov> so up to you
- # [12:58] <@askalski> I'd stick with it for today
- # [12:58] <@surkov> ok
- # [12:58] <@askalski> and then ask david
- # [12:59] <@surkov> yep
- # [13:07] <@surkov> askalski: that'd be nice if you put some comments into the code helping to understand the alg
- # [13:07] <@surkov> in English :)
- # [13:07] <@askalski> surkov, yes, in patch
- # [13:07] <@surkov> ok
- # [13:08] <@askalski> surkov, also, I make some effort so the names mimic these in paper
- # [13:08] <@askalski> surkov, because explaining this algorithm in comments is pointless
- # [13:08] <@surkov> ok, if names are close
- # [13:38] * Joins: jlin (Jonathan@moz-619C4F64.x.rootbsd.net)
- # [13:55] * Quits: @askalski (akuda@moz-104CC309.mv.mozilla.com) (Ping timeout)
- # [14:00] * Quits: jhk (jiggy@8E6C34C1.A3F9767A.1C37C358.IP) (Connection reset by peer)
- # [14:05] * Quits: richardschwerdtfeger (richardsch@moz-130F1494.lightspeed.austtx.sbcglobal.net) (Quit: richardschwerdtfeger)
- # [14:10] * Joins: askalski (akuda@moz-104CC309.mv.mozilla.com)
- # [14:10] * ChanServ sets mode: +o askalski
- # [15:04] <@askalski> surkov, how important is this function? How do you think?
- # [15:04] <@askalski> this doTextUpdate()
- # [15:05] <@askalski> I am seriously thinking I am overthinking this
- # [15:06] <@surkov> askalski: we don't really have measuring, but you should keep in mind that this part of code is called for text frames when you resize the window, it can be pretty heavvy
- # [15:07] <@askalski> surkov, OK. I'll work on it further
- # [15:08] * Quits: @askalski (akuda@moz-104CC309.mv.mozilla.com) (Quit: Wychodzi)
- # [15:13] * Joins: davidb (davidb@F2D29657.F60B0462.67AC9B1.IP)
- # [15:13] * ChanServ sets mode: +qo davidb davidb
- # [15:13] <@davidb> hi all!
- # [15:16] <@davidb> nice to see bz on the canvas solution
- # [15:27] <@firebot> surkov.alexander@gmail.com requested review from marco.zehe@googlema il.com for attachment 599142 on bug 650241.
- # [15:27] <@firebot> Bug https://bugzilla.mozilla.org/show_bug.cgi?id=650241 nor, --, ---, surkov.alexander, NEW, Location returned by accessibles sometimes incorrect when page zoomed
- # [15:28] <@surkov> hi, davidb
- # [15:32] <@davidb> hi surkov
- # [15:32] * Joins: jprmc (jprmc@F2D29657.F60B0462.67AC9B1.IP)
- # [15:32] * ChanServ sets mode: +o jprmc
- # [15:32] * Joins: peteb-away (ptbrunet@moz-E9B02845.austin.res.rr.com)
- # [15:32] <@davidb> surkov: nice post, I hope Foliot isn't distracting you too much :)
- # [15:32] <@davidb> (he has that way about him)
- # [15:33] <@surkov> davidb: thanks, that depends from what :)
- # [15:33] <@surkov> I'd say he doesn't distract he involves
- # [15:33] <@davidb> ok
- # [15:35] * Joins: jhk (jiggy@8E6C34C1.A3F9767A.1C37C358.IP)
- # [15:42] <@davidb> MarcoZ: would you like to attend that CSUN thing about a professional society of a11y twits?
- # [15:42] <@davidb> err twits/people
- # [15:42] <@davidb> it might be interesting for one of us to go and report back
- # [15:43] <@MarcoZ> davidb: Oh, what's that about?
- # [15:43] <@davidb> MarcoZ: Derek Featherstone has been tweeting about it
- # [15:44] <@MarcoZ> davidb: Don't even know who that is. ;)
- # [15:45] <@MarcoZ> davidb: Should I?
- # [15:47] <@davidb> hmm
- # [15:53] <@MarcoZ> davidb: Are you referring to http://simplyaccessible.com/ There are articles by a Derek Featherstone there, but tbh this is really the first I've heard of these guys.
- # [16:03] <@MarcoZ> surkov: Question: If your proposal from the blog post were implemented, what would then be the difference between role="presentation" and aria-hidden="true"? If I read your blog post correctly, there would be none. (esp the last paragraph)
- # [16:04] <@surkov> MarcoZ: aria-hidden is inherited, no difference at all
- # [16:04] <@surkov> I meant no other difference
- # [16:04] <@MarcoZ> surkov: Ah, but role="presentation" would need to be specified for *each* element, would have no effect on subtrees.
- # [16:05] <@surkov> MarcoZ: sorry, not sure I understood, why would we need to specify it for each element?
- # [16:07] <@MarcoZ> surkov: role="presentation" does not hide subtrees, aria-hidden="true" does, right?
- # [16:07] <@surkov> MarcoZ: not in Firefox
- # [16:07] <@surkov> aria-hidden hides nothing
- # [16:08] <@surkov> per spec it's allowed to hide but not required
- # [16:08] <@MarcoZ> surkov: It does expose the object attribute, then.
- # [16:08] <@surkov> yes
- # [16:08] <@MarcoZ> surkov: Thanks!
- # [16:08] <@MarcoZ> surkov: Reviewing your patch.
- # [16:08] <@surkov> thanks! :)
- # [16:09] <@surkov> MarcoZ: yeah, I see the last paragraph is ambiguous
- # [16:13] <@firebot> marco.zehe@googlemail.com granted review for attachment 599142 on bug 650241.
- # [16:13] <@firebot> Bug https://bugzilla.mozilla.org/show_bug.cgi?id=650241 nor, --, ---, surkov.alexander, NEW, Location returned by accessibles sometimes incorrect when page zoomed
- # [16:13] * Joins: askalski (akuda@moz-104CC309.mv.mozilla.com)
- # [16:13] * ChanServ sets mode: +o askalski
- # [16:14] <@askalski> davidb, hi!
- # [16:18] * khuey|away is now known as khuey
- # [16:26] * khuey is now known as khuey|away
- # [16:27] * Quits: @askalski (akuda@moz-104CC309.mv.mozilla.com) (Ping timeout)
- # [16:37] <@davidb> MarcoZ: sorry was in a meeting
- # [16:37] * Quits: nhirata|afk (nhirata.bu@moz-2A9C9106.hsd1.ca.comcast.net) (Quit: nhirata|afk)
- # [16:41] <@davidb> MarcoZ: http://www.atia.org/i4a/pages/index.cfm?pageid=4161
- # [16:42] <@davidb> $300
- # [16:43] <@davidb> a lot of people are abstaining on principle
- # [16:43] <@davidb> it should be a free event.
- # [16:44] <@davidb> it should not be "pay to play"
- # [16:45] <@davidb> MarcoZ, surkov, tbsaunde what questions would you like me to ask Taras about telemetry
- # [16:46] <@davidb> for example "How often is a11y instantiated not by a known screen reader?"
- # [16:47] <@surkov> davidb: I guess this question you can address yourself by adding Other const to Compatibility.h
- # [16:47] <@surkov> davidb: can somebody (not we) implement an interface to collect strings so we can use it from accessibility, I'd like to see NVDA not 0
- # [16:48] * Joins: hub (hub@moz-E2FCA694.figuiere.net)
- # [16:48] * ChanServ sets mode: +o hub
- # [16:52] <@davidb> surkov: it feels low priority to me, but maybe?
- # [16:52] <@davidb> regarding the new const - it makes sense to me
- # [16:53] <@davidb> I answered those points out of order
- # [16:53] <@davidb> hey hub
- # [16:53] <@surkov> davidb: we're not unique customers of this behavior
- # [16:53] <@davidb> yeah
- # [16:53] <@davidb> but does it feel important/urgent?
- # [16:54] <@davidb> if it bothers you enough - that's fair :)
- # [16:54] <@davidb> Oh I see some old jaws users now
- # [16:54] <@davidb> wow
- # [16:58] * Joins: hhillen (hhillen@moz-81563C98.ip.telfort.nl)
- # [17:10] <@davidb> surkov: I've created a patch to collect for 'unknown' AT telemetry
- # [17:10] * @davidb compiles
- # [17:11] * @davidb goes back to TODO list
- # [17:11] <@surkov> davidb: nice
- # [17:11] <@tbsaunde> davidb: ask him why we don't have ui to get unknown from the data we already have?
- # [17:11] <@davidb> yeah, but i sort of wish i had left '0' for this enum
- # [17:11] <@davidb> tbsaunde: lol
- # [17:12] <@tbsaunde> davidb: well, the unknown consant is a bit dumb...
- # [17:12] * Joins: nhirata (nhirata.bu@moz-BBE3ABD.mv.mozilla.com)
- # [17:12] <@davidb> tbsaunde: do tell.
- # [17:12] <@davidb> tbsaunde: you mean since we should be able to deduce it?
- # [17:13] <@tbsaunde> davidb: yes
- # [17:13] <@davidb> yeah
- # [17:13] <@davidb> still.
- # [17:13] <@tbsaunde> what?
- # [17:14] <@davidb> i think having the constant gets us what we need faster
- # [17:14] <@tbsaunde> davidb: is that really ture? we have to wait for you patch to make a release instead of analizing data we already have
- # [17:14] <@davidb> hmm that is a good point
- # [17:14] <@davidb> I'll ask Taras either way.
- # [17:15] <@MarcoZ> davidb: Just read the link. ATIA and CSUN are competing shows. ATIA is struggling to gain as much interest as CSUN does, esp in international audience, which is completely ignoring that show.
- # [17:15] <@MarcoZ> davidb: And $300 is just hideously expensive for a one-day event.
- # [17:15] <@davidb> yeah
- # [17:16] <@davidb> it feels sort of incestuous to me
- # [17:16] <@tbsaunde> davidb: I suppose you can't do something clever like run some js in one of those dev tools things against the page?
- # [17:16] <@MarcoZ> It is. Besides, I'm only arriving on the 28th at 7 PM, so wouldn't be able to make this event anyway. And no, I don't feel it is the right thing to do, esp for that price.
- # [17:16] <@davidb> tbsaunde: I would like to
- # [17:16] <@davidb> BTW I wrote an add-on on the weekend
- # [17:17] <@MarcoZ> davidb: Cool, what does it do?
- # [17:17] <@davidb> it is very lightweight, and uses workers - it detects aria and role attributes in web pages
- # [17:17] <@MarcoZ> Nice!
- # [17:17] <@davidb> provides the counts
- # [17:18] <@davidb> in related news i kind of like xpath regex
- # [17:18] <@tbsaunde> davidb: provides the counts to who?
- # [17:18] <@davidb> tbsaunde: the user of the add-on.
- # [17:18] <@tbsaunde> ok
- # [17:18] <@davidb> counts # elements
- # [17:18] <@tbsaunde> isn't the DOM single threaded? so what are workers for?
- # [17:18] <@davidb> so we don't slow down browsing
- # [17:19] <@MarcoZ> Sponsored by Microsoft, IBM and AISquared? Humm....
- # [17:19] <@davidb> MarcoZ: heheh
- # [17:19] * Joins: ehsan (ehsan@F2D29657.F60B0462.67AC9B1.IP)
- # [17:19] <@tbsaunde> davidb: isn't that just updating a hash table?
- # [17:20] * @tbsaunde would expect the overhead of sending a message to a different thread is higher than the work itself
- # [17:20] <@davidb> tbsaunde: the beauty is that with this set up i can add more complex criteria.
- # [17:20] <@davidb> right now there are two tests each in their own worker
- # [17:21] <@davidb> also this is the blessed way
- # [17:21] <@davidb> we want all add-on devs to do this
- # [17:21] <@tbsaunde> two tests?
- # [17:22] * Joins: victorporof (victorporo@1ADEBCE.63BE4D17.4873E54D.IP)
- # [17:23] <@davidb> tbsaunde: i look for roles in one worker using get elements*, and aria- in another worker using document.evaluate xpath magic.
- # [17:24] <@davidb> some pages only have roles
- # [17:24] <@davidb> and i have separate indicators for each
- # [17:24] <@hub> hi
- # [17:24] <@davidb> heyo
- # [17:24] <@tbsaunde> davidb: I see, doesn't that mean your touching the dom from non main thread?
- # [17:24] <@davidb> tbsaunde: that's magic to me.
- # [17:24] <@davidb> but yes
- # [17:25] <@tbsaunde> interesting
- # [17:25] <@davidb> presumably that's optimized or will be
- # [17:25] <@davidb> i don't think i actually run until the doc is loaded
- # [17:25] <@davidb> and actually, i have 2 workers for each doc
- # [17:25] <@davidb> so one 'page' can have 2n workers
- # [17:25] <@davidb> if is has n docs
- # [17:26] <@davidb> i'm not sure that is user friendly
- # [17:26] <@davidb> since i can end up with several indicators on the status bar
- # [17:26] <@tbsaunde> lol
- # [17:26] <@davidb> heh
- # [17:27] <@tbsaunde> its also a bit heavy on the kthreads imho
- # [17:27] <@tbsaunde> consider say having 100 tabs open :)
- # [17:28] <@tbsaunde> davidb: what does the usage on linux look like btw
- # [17:28] * Quits: jhk (jiggy@8E6C34C1.A3F9767A.1C37C358.IP) (Connection reset by peer)
- # [17:28] <@davidb> well they run to completion after page load
- # [17:28] <@davidb> or doc load i should say
- # [17:28] <@davidb> tbsaunde: a11y usage?
- # [17:28] <@tbsaunde> davidb: oh, you don't handle java script adding aria or roles?
- # [17:29] <@tbsaunde> davidb: yes
- # [17:30] <@davidb> tbsaunde: it usually does that before i run i think
- # [17:30] <@davidb> (seems to)
- # [17:30] <@davidb> maybe i should have them run on demand
- # [17:30] <@tbsaunde> davidb: yeah, or just use event handlers?
- # [17:31] <@davidb> ew
- # [17:32] <@davidb> well i want this to be as non invasive as possible
- # [17:32] <@davidb> a lot of sites use roles btw
- # [17:32] * Joins: jhk (jiggy@8E6C34C1.A3F9767A.1C37C358.IP)
- # [17:32] <@davidb> especially since they are used in js libraries
- # [17:32] <@davidb> hi jhk
- # [17:33] <@tbsaunde> davidb: btw if you dislike gcc's errors you might find these ones from sml/nj fun http://paste.debian.net/157065/
- # [17:33] <@davidb> i do not
- # [17:33] <@davidb> :)
- # [17:34] <@tbsaunde> hehe
- # [17:35] <@tbsaunde> you mean expected ?.t got ?.t isn't useful ... ;)
- # [17:35] <@MarcoZ> LOL
- # [17:35] <@davidb> i can't say i enjoyed that part especially
- # [17:35] <@davidb> llvm FTW
- # [17:36] <@davidb> tbsaunde: have you looked at rust?
- # [17:36] <@davidb> i was going to look at rust or write an add-on on the weekend, i chose both :)
- # [17:37] <@davidb> but only read about rust so far
- # [17:42] <@firebot> surkov.alexander@gmail.com requested review from trev.saunders@gmail .com for attachment 599193 on bug 728907.
- # [17:42] <@firebot> surkov.alexander@gmail.com changed the Status on bug 728907 from NEW to ASSIGNED.
- # [17:42] <@firebot> Bug https://bugzilla.mozilla.org/show_bug.cgi?id=728907 nor, --, ---, nobody, NEW, don't use GetComputedStyle for text attributes
- # [17:54] * Quits: sgimeno (santiago.g@A922400D.48362F18.BCC03455.IP) (Quit: KVIrc 4.0.2 Insomnia http://www.kvirc.net/)
- # [17:54] * Joins: sgimeno (santiago.g@A922400D.48362F18.BCC03455.IP)
- # [18:00] <@firebot> bmo@edmorley.co.uk changed the Status on bug 714579 from REOPENED to RESOLVED.
- # [18:00] <@firebot> bmo@edmorley.co.uk set the Resolution field on bug 714579 to FIXED.
- # [18:00] <@firebot> Bug https://bugzilla.mozilla.org/show_bug.cgi?id=714579 cri, --, mozilla13, surkov.alexander, RESO FIXED, crash nsINode::OwnerDoc
- # [18:01] <jhk> hi davidb
- # [18:01] <@davidb> hey hey
- # [18:01] <jhk> :)
- # [18:04] * Quits: geo (geo@moz-BBE3ABD.mv.mozilla.com) (Ping timeout)
- # [18:05] * Quits: peteb-away (ptbrunet@moz-E9B02845.austin.res.rr.com) (Ping timeout)
- # [18:17] * Quits: hhillen (hhillen@moz-81563C98.ip.telfort.nl) (Ping timeout)
- # [18:19] <@tbsaunde> davidb: no, haven't really looked at rust
- # [18:19] * @davidb nods
- # [18:19] <@tbsaunde> I've been meaning to, but I think haskel miht be a little higher on the list
- # [18:20] * Joins: hhillen (hhillen@moz-81563C98.ip.telfort.nl)
- # [18:20] * Quits: hhillen (hhillen@moz-81563C98.ip.telfort.nl) (Quit: hhillen)
- # [18:24] <@firebot> New Core - Disability Access APIs bug 729154 filed by bolterbugz@gmail.com.
- # [18:24] <@firebot> Bug https://bugzilla.mozilla.org/show_bug.cgi?id=729154 nor, --, ---, nobody, NEW, Telemetry for a11y instantiation by unknown cause.
- # [18:25] <@firebot> bolterbugz@gmail.com requested review from trev.saunders@gmail .com for attachment 599217 on bug 729154.
- # [18:26] <@firebot> surkov.alexander@gmail.com cancelled review?(trev.saunders@gmail .com) for attachment 599193 on bug 728907.
- # [18:26] <@firebot> surkov.alexander@gmail.com requested review from trev.saunders@gmail .com for attachment 599218 on bug 728907.
- # [18:26] <@firebot> Bug https://bugzilla.mozilla.org/show_bug.cgi?id=728907 nor, --, ---, surkov.alexander, ASSI, don't use GetComputedStyle for text attributes
- # [18:27] <@davidb> yeah haskell is interesting
- # [18:27] <@davidb> rust is closer to work related though
- # [18:28] <@davidb> tbsaunde: ignore that patch
- # [18:28] <@davidb> gotta refresh
- # [18:28] <@firebot> bolterbugz@gmail.com cancelled review?(trev.saunders@gmail .com) for attachment 599217 on bug 729154.
- # [18:32] <@tbsaunde> davidb: maybe
- # [18:32] <@tbsaunde> I won't do reviews until probbably tomorrow :/
- # [18:33] <@firebot> bolterbugz@gmail.com requested review from trev.saunders@gmail .com for attachment 599220 on bug 729154.
- # [18:34] <@davidb> np
- # [18:34] * Joins: askalski (akuda@moz-4C8A107E.pool85-48-91.dynamic.orange.es)
- # [18:34] * ChanServ sets mode: +o askalski
- # [18:34] <@askalski> davidb,
- # [18:34] <@askalski> hi
- # [18:35] <@askalski> can we talk for just a moment? need 2-3 quick answets
- # [18:35] <@davidb> askalski: yes! how? skype?
- # [18:35] <@askalski> ok
- # [18:35] <@davidb> 1sec
- # [18:38] * Joins: drexler (chatzilla@moz-BD26ECCD.hsd1.vt.comcast.net)
- # [18:42] * Joins: joanie (joanmarie@moz-456EDD85.net)
- # [18:44] * Quits: joanie (joanmarie@moz-456EDD85.net) (Quit: Changing server)
- # [18:55] <@davidb> tbsaunde: did you catch the PMs
- # [18:58] * Quits: @askalski (akuda@moz-4C8A107E.pool85-48-91.dynamic.orange.es) (Quit: Wychodzi)
- # [19:08] <@tbsaunde> davidb: I think I responded to you?
- # [19:21] * Quits: @surkov (surkov@703D4713.1344B664.5D3F4C44.IP) (Quit: surkov)
- # [19:21] * Quits: @hub (hub@moz-E2FCA694.figuiere.net) (Ping timeout)
- # [19:22] * Quits: sgimeno (santiago.g@A922400D.48362F18.BCC03455.IP) (Quit: KVIrc 4.0.2 Insomnia http://www.kvirc.net/)
- # [19:28] * Quits: jhk (jiggy@8E6C34C1.A3F9767A.1C37C358.IP) (Ping timeout)
- # [19:35] * khuey|away is now known as khuey
- # [19:37] * Joins: iLoSt (iLoSt@7F98D9B0.BB48A6FB.5302C6B7.IP)
- # [19:38] * Quits: iLoSt (iLoSt@7F98D9B0.BB48A6FB.5302C6B7.IP) (Client exited)
- # [19:39] * Joins: hub (hub@21B7B9F2.B87E9213.6E712CE2.IP)
- # [19:39] * ChanServ sets mode: +o hub
- # [19:52] * Quits: @MarcoZ (marco.zehe@moz-72E20CAA.dip.t-dialin.net) (Quit: night!)
- # [20:03] * Parts: eeejay (eitan@moz-EE2D6C10.monotonous.org) (Ex-Chat)
- # [20:03] * Joins: eeejay (eitan@moz-EE2D6C10.monotonous.org)
- # [20:03] * ChanServ sets mode: +o eeejay
- # [20:06] <@davidb> eeejay: seems ok to me
- # [20:06] <@eeejay> cool, i'll post
- # [20:06] <@davidb> might be epic ;)
- # [20:06] <@eeejay> hope so!
- # [20:06] <@eeejay> epic in a good way, right?
- # [20:07] <@eeejay> i feel like i was fair to google, they wouldn't deny focusing on per-task solutions
- # [20:07] <@davidb> yeah
- # [20:09] <@eeejay> sent
- # [20:09] <@davidb> eeejay: this link works right? http://groups.google.com/group/mozilla.dev.platform/browse_thread/thread/c93b67726f79cd83#
- # [20:10] <@eeejay> davidb, yeah
- # [20:10] <@eeejay> davidb, the bottom line for you to repeat is "expect paches in your modules", specifically in the chrome js
- # [20:10] * Quits: nhirata (nhirata.bu@moz-BBE3ABD.mv.mozilla.com) (Ping timeout)
- # [20:11] <@davidb> eeejay: please confirm https://wiki.mozilla.org/Platform/2012-02-21#Accessibility
- # [20:12] <@davidb> look right?
- # [20:13] <@davidb> eeejay: i tweaked the second bullet https://wiki.mozilla.org/Platform/2012-02-21#Accessibility
- # [20:13] <@eeejay> davidb, looks great
- # [20:13] <@davidb> ok
- # [20:14] <@davidb> eeejay: would you prefer to do the audible yourself?
- # [20:14] <@eeejay> davidb, nah
- # [20:14] <@davidb> ok
- # [20:14] <@eeejay> davidb, could hardly get vidyo to work
- # [20:14] * @davidb nods solemnly
- # [20:16] * Joins: nhirata (nhirata.bu@moz-BBE3ABD.mv.mozilla.com)
- # [20:21] * Joins: jhk (jiggy@8E6C34C1.A3F9767A.1C37C358.IP)
- # [20:25] <@eeejay> *phew* no questions
- # [20:25] <@davidb> heheh
- # [20:32] * Quits: nhirata (nhirata.bu@moz-BBE3ABD.mv.mozilla.com) (Connection reset by peer)
- # [20:32] * Joins: nhirata (nhirata.bu@moz-BBE3ABD.mv.mozilla.com)
- # [20:34] <@firebot> akeybl@mozilla.com granted approval-mozilla-aurora for attachment 598222 on bug 677154.
- # [20:34] <@firebot> Bug https://bugzilla.mozilla.org/show_bug.cgi?id=677154 maj, P1, mozilla13, surkov.alexander, RESO FIXED, Detached document accessibility tree
- # [20:39] * Quits: jhk (jiggy@8E6C34C1.A3F9767A.1C37C358.IP) (Ping timeout)
- # [20:40] * Joins: jhk (jiggy@8E6C34C1.A3F9767A.1C37C358.IP)
- # [20:46] <@firebot> akeybl@mozilla.com granted approval-mozilla-beta for attachment 598719 on bug 718237.
- # [20:46] <@firebot> akeybl@mozilla.com granted approval-mozilla-aurora for attachment 598719 on bug 718237.
- # [20:46] <@firebot> Bug https://bugzilla.mozilla.org/show_bug.cgi?id=718237 maj, --, ---, nobody, NEW, [SeaMonkey] "accessible/events/test_focus_autocomplete.xul | Test timed out." (which also causes lot
- # [21:06] <@davidb> nice
- # [21:20] * Quits: nhirata (nhirata.bu@moz-BBE3ABD.mv.mozilla.com) (Connection reset by peer)
- # [21:20] * Joins: nhirata (nhirata.bu@moz-BBE3ABD.mv.mozilla.com)
- # [21:28] <@eeejay> *lunch*
- # [21:34] * Joins: satdav (satdav@moz-1ECB932B.cable.virginmedia.com)
- # [21:59] * Joins: askalski (akuda@moz-4C8A107E.pool85-48-91.dynamic.orange.es)
- # [21:59] * ChanServ sets mode: +o askalski
- # [21:59] <@firebot> andrew.quartey@gmail.com requested review from surkov.alexander@gm ail.com for attachment 599311 on bug 691267.
- # [21:59] <@firebot> Bug https://bugzilla.mozilla.org/show_bug.cgi?id=691267 nor, --, ---, andrew.quartey, NEW, HTML input button and HTML 4 button should share the same class
- # [21:59] <@askalski> hi
- # [22:00] * Joins: geo (geo@moz-BBE3ABD.mv.mozilla.com)
- # [22:00] <@askalski> davidb, I came up with 2 ideas
- # [22:00] <@askalski> davidb, science-related
- # [22:00] <@davidb> hey
- # [22:00] <@davidb> juggling tasks ATM
- # [22:00] * @eeejay loves science
- # [22:00] <@askalski> davidb, ok. I can mail you so you can answer some other time
- # [22:01] <@davidb> sure
- # [22:01] <@davidb> cc people who might help
- # [22:01] <@davidb> if poss
- # [22:04] * Quits: jhk (jiggy@8E6C34C1.A3F9767A.1C37C358.IP) (Connection reset by peer)
- # [22:19] * Quits: webben (benjamin@moz-BA505DDD.static.cloud-ips.com) (Ping timeout)
- # [22:37] * @tbsaunde considers writing a message filter for bugs containing "discussion forum" or maybe even "mozilla.accessibility"
- # [22:43] * Joins: peteb-away (ptbrunet@moz-E9B02845.austin.res.rr.com)
- # [22:44] <@davidb> tbsaunde: spam?
- # [22:44] * davidb is now known as davidb|mth
- # [22:44] * davidb|mth is now known as davidb|mtg
- # [22:47] * Quits: @askalski (akuda@moz-4C8A107E.pool85-48-91.dynamic.orange.es) (Quit: Wychodzi)
- # [22:49] <@tbsaunde> davidb|mtg: yeah, its sort of spammy bugmail
- # [22:50] * Joins: webben (benjamin@moz-BA505DDD.static.cloud-ips.com)
- # [23:00] <@firebot> leonard.camacho@gmail.com requested feedback from neil@httl.net for attachment 599342 on bug 291082.
- # [23:00] <@firebot> Bug https://bugzilla.mozilla.org/show_bug.cgi?id=291082 nor, --, ---, nobody, NEW, preventDefault doesn't block keyboard navigation in select-one drop-down lists
- # [23:05] <satdav> does anyone know does http://en.wikipedia.org/wiki/JAWS_%28screen_reader%29 work with firefox
- # [23:06] <@hub> on windows, supposedly
- # [23:12] <@davidb|mtg> ciao
- # [23:12] * Quits: @davidb|mtg (davidb@F2D29657.F60B0462.67AC9B1.IP) (Quit: davidb|mtg)
- # [23:13] <satdav> never used it before
- # [23:17] * Joins: Mark_Capella (chatzilla@moz-DD0C7E4F.twcny.res.rr.com)
- # [23:18] <@firebot> alex_mayorga@yahoo.com changed the Resolution on bug 249735 from FIXED to INCOMPLETE.
- # [23:18] <@firebot> Bug https://bugzilla.mozilla.org/show_bug.cgi?id=249735 nor, P1, Firefox 12, dao, RESO INCOMPLETE, F6 goes to the first focusable element instead of the address bar (does not focus on location bar)
- # [23:19] <@firebot> enndeakin@gmail.com changed the Resolution on bug 249735 from INCOMPLETE to FIXED.
- # [23:42] * Quits: Mark_Capella (chatzilla@moz-DD0C7E4F.twcny.res.rr.com) (Quit: ChatZilla 0.9.88 [Firefox 12.0a2/20120212042014])
- # Session Close: Wed Feb 22 00:00:00 2012
The end :)