/irc-logs / mozilla / #accessibility / 2012-02-17 / end
Options:
- # Session Start: Fri Feb 17 00:00:00 2012
- # Session Ident: #accessibility
- # [00:00] * Joins: davidb_ (davidb@moz-6F9F653A.dsl.bell.ca)
- # [00:34] * Quits: davidb_ (davidb@moz-6F9F653A.dsl.bell.ca) (Quit: davidb_)
- # [00:40] <@eeejay> davidb, http://monotonous.org/files/t.webm
- # [00:40] <@eeejay> davidb, http://monotonous.org/files/t.webm
- # [00:41] <davidb> \o/
- # [00:41] <@eeejay> davidb, a wip. i will add another segment showing generally how android a11y works
- # [00:41] <@eeejay> davidb, should i add some slides and explain the whole approach?
- # [00:42] <@eeejay> it is kind of underwhelming
- # [00:43] <davidb> this is decent
- # [00:44] <davidb> I'll leave slides/explanation up to you.
- # [00:47] <davidb> heh. sausagefests
- # [01:09] * Quits: victorporof (victorporo@C092FEB2.1C233438.79933D60.IP) (Connection reset by peer)
- # [01:09] * Joins: victorporof (victorporo@C092FEB2.1C233438.79933D60.IP)
- # [01:16] * Quits: jamesn (jamesn@moz-34C566A2.oracle.com) (Ping timeout)
- # [01:23] <@tbsaunde> eeejay: what is that movie supposed to be?
- # [01:24] <@eeejay> it is me navigating twitter with android
- # [01:24] <@eeejay> tbsaunde, should be self-explanatory
- # [01:25] <@tbsaunde> eeejay: ok, I got that it was android on a web page, but it wasn't really clear that was fenic you where using
- # [01:29] * Quits: satdav (satdav@moz-1ECB932B.cable.virginmedia.com) (Quit: Leaving)
- # [01:34] * Quits: victorporof (victorporo@C092FEB2.1C233438.79933D60.IP) (Quit: Linkinus - http://linkinus.com)
- # [01:37] * Joins: peteb-away (ptbrunet@moz-E9B02845.austin.res.rr.com)
- # [01:37] * Joins: Jamie (jamie@moz-CA26021.jantrid.net)
- # [01:39] <davidb> I'm not great with the eyes free virtual dpad yet.
- # [01:45] <@hub> that awesome
- # [01:50] * Jamie groans
- # [01:50] <Jamie> there shouldn't even be a need for that damned thing
- # [01:51] <Jamie> Heh. I'm sorry. I can't help it. Thinking about Android a11y just makes me angry and bitter.
- # [01:54] <Jamie> I'm not normally one to slam someone for trying, but I dislike half-baked attempts that go mainstream, are claimed as acceptable, don't learn anything from past mistakes and totally fragment all hope for polished a11y on for the entire platform.
- # [01:58] <Jamie> davidb: Hey, is there any way for the public ot see telemetry stats? no huge need, just curious
- # [02:00] <davidb> yep. hard to help right now. on mobile.
- # [02:00] <Jamie> no problem
- # [02:01] <davidb> search mozilla metrics pentaho. needs a11y improvements.
- # [02:02] * davidb rests phone to charge
- # [02:04] <Jamie> ah, not accessible to public. no matter./
- # [02:06] <davidb> oh
- # [02:10] <Jamie> arrg. I really wish I had the time to get involved in some of the w3c groups
- # [02:20] <@hub> Jamie: your grief is about Android and its a11y support?
- # [02:21] <@hub> we should really make sure b2g is accessible
- # [02:31] <davidb> yes.
- # [02:39] * Quits: ehsan (ehsan@F2D29657.F60B0462.67AC9B1.IP) (Input/output error)
- # [02:46] * Joins: davidb_ (davidb@moz-6F9F653A.dsl.bell.ca)
- # [02:47] <@eeejay> Jamie, thinking about android a11y makes me think how great we could be just by not sucking
- # [02:53] * khuey is now known as khuey|away
- # [02:53] <davidb_> tbsaunde: ping (backchannel)
- # [02:54] <davidb_> eeejay: lol
- # [02:57] * Joins: surkov (surkov@74A597FA.C2B307F0.34044A7F.IP)
- # [02:57] * ChanServ sets mode: +o surkov
- # [02:57] <davidb_> hi surkov
- # [02:57] <@surkov> hi, davidb_
- # [02:58] * Quits: davidb (davidb@moz-5E2C9899.eng.wind.ca) (Quit: Yaaic - Yet another Android IRC client - http://www.yaaic.org)
- # [02:58] * davidb_ is now known as davidb
- # [03:07] * Quits: nhirata (nhirata.bu@moz-BBE3ABD.mv.mozilla.com) (Quit: nhirata)
- # [03:09] * Joins: ehsan (ehsan@F0B20A8D.8458880F.57F33CED.IP)
- # [03:22] <@firebot> New Core - Keyboard: Navigation bug 728103 filed by masayuki@d-toybox.com.
- # [03:22] <@firebot> Bug https://bugzilla.mozilla.org/show_bug.cgi?id=728103 nor, --, ---, masayuki, NEW, Shouldn't we change modifier for HTML accesskey from Control to Control + Option
- # [03:25] * Quits: @hub (hub@21B7B9F2.B87E9213.6E712CE2.IP) (Input/output error)
- # [03:27] * Quits: @jprmc (jprmc@F2D29657.F60B0462.67AC9B1.IP) (Ping timeout)
- # [03:41] <Jamie> eeejay: it's inherently "bad", even if it works well, because it isn't a unified experience
- # [03:42] <Jamie> eeejay: Mozilla has no choice, I get that. That doesn't mean that good Gecko a11y on Android makes Android a more accessible platform
- # [03:42] <Jamie> Mozilla is having to fragment a11y on Android even further than it is already fragmented
- # [03:42] <Jamie> entirely because Android gave Mozilla no choice
- # [03:44] <Jamie> It'd be like saying Firevox is the correct solution for Windows accessibility of Firefox. We both know that isn't true.
- # [03:46] <@tbsaunde> Jamie: I'm not sure I see what we're doing at this point that makes anything worse
- # [03:47] <Jamie> tbsaunde: Mm, misphrase on my part. It's not making anything worse. It's just continuing an already bad trend.
- # [03:48] <Jamie> tbsaunde: and I must emphasise: I'm not saying this has anything to do with Mozilla. It's entirely the fault of Android.
- # [03:48] <@tbsaunde> Jamie: which trend is that?
- # [03:48] <@tbsaunde> Jamie: yeah, I got that part
- # [03:50] <Jamie> tbsaunde: Well, Gecko isn't quite as bad here; it's not doing the self voicing thing. However, it still has to implement its own touch screen access and navigation
- # [03:50] * Joins: jprmc (jprmc@8F761026.B5EF4AF6.E96CA9D8.IP)
- # [03:50] * ChanServ sets mode: +o jprmc
- # [03:51] <@tbsaunde> Jamie: true
- # [03:51] <@tbsaunde> Jamie: on the other hand we do caret nav mode on desktop
- # [03:52] <Jamie> tbsaunde: This leads to a fragmented user experience, potentially. Even if everyone does the same thing, that's a lot of pointless code duplication
- # [03:52] <@tbsaunde> and firefox handles navigation for non-screen reader users
- # [03:53] <@tbsaunde> Jamie: sure, but that is true for sighted users naviagting different apps ui's too
- # [03:53] <Jamie> tbsaunde: that's true. however, screen reader gestures are somewhat more specialised than that.
- # [03:54] <Jamie> tbsaunde: Mm, true, but most apps try to keep a certain degree of consistency with the native UI
- # [03:55] <@tbsaunde> Jamie: true
- # [03:55] <Jamie> tbsaunde: to use another example, doing it this way makes it very difficult for, say, braille display access
- # [03:55] <Jamie> tbsaunde: Gecko would then have to support braille display commands directly
- # [03:55] <@tbsaunde> Jamie: I think my point is that the virtual buffer setup in windows screen readers is kind of a weird way to organize things
- # [03:56] <Jamie> tbsaunde: It is
- # [03:56] <Jamie> but at least the exposure from the browser can be standardised using, say, IA2
- # [03:56] <@tbsaunde> Jamie: yeah, braille on android is I suspect fubard but I don't really know anything about it
- # [03:57] <davidb> this is so much better than twitter for conversations.
- # [03:57] <@tbsaunde> I'm not sure how much it matters though since using a brailleddisplay with a pone sounds ultra anoying
- # [03:57] <Jamie> tbsaunde: virtual buffers only really exist for two reasons: 1. out-of-process calls are slow and there's no collections API in IA2. 2. we believe that we render things in a better, more reliable form than browser caret navigation
- # [03:58] <@tbsaunde> davidb: haha
- # [03:58] <Jamie> tbsaunde: android has no support for braille displays, so right now, it's academic
- # [03:58] <Jamie> Using a braille display with an iPhone is actually surprisingly nice
- # [03:58] <davidb> tbsaunde: see what i did there
- # [03:59] <@tbsaunde> davidb: I did ;)
- # [03:59] <Jamie> tbsaunde: In short, virtual buffers are a big nasty hack necessary to solve a problem that no one has the resources to solve properly :)
- # [04:00] <@tbsaunde> Jamie: I'm not sure I see how #1 is a reason for virtual buffers, but with the current state of caret naviagtion in firefox I'd probably agree with #2 but that's because we should fix it
- # [04:00] <Jamie> tbsaunde: even if Gecko fixed it, the other browsers probably wouldn't
- # [04:00] <Jamie> and then we'd end up with an inconsistent UX
- # [04:00] <@tbsaunde> Jamie: true
- # [04:00] <Jamie> tbsaunde: 1 is a reason because of quick navigation and the like
- # [04:01] <@tbsaunde> Jamie: but couldn't you dll inject something that just changes the focus
- # [04:01] <Jamie> tbsaunde: example: I press a key to jump to the next link. with IA2, we'd have to walk through every element to find the next link
- # [04:02] <Jamie> tbsaunde: If that were the only factor, you're right. except there's also: you have a paragraph consisting of lots of text interspursed with 100 links
- # [04:02] <@tbsaunde> Jamie: I imagine caret nav in webkit doesn't work on windows? I understand it works fairly well on linux
- # [04:02] <Jamie> tbsaunde: now, we want ot read that paragraph as one block. we have to crawl through that paragraph bit by bit, walking through the objects, retrieving data
- # [04:02] <@tbsaunde> Jamie: ok
- # [04:02] <Jamie> remember, each fetch of an object is multiple cross-process calls
- # [04:03] <Jamie> tbsaunde: I'm not sure, because WebKit a11y in windows is so broken that it's hardly worth bothering with
- # [04:03] <Jamie> and I'm not being overly negative here
- # [04:03] <Jamie> it's simple fact :)
- # [04:03] <Jamie> the only place it really works is iTunes and I'm not sure you can enable caret nav in iTunes, so I can't test it
- # [04:04] * Jamie still thinks b2g has really exciting a11y potential
- # [04:04] <davidb> Jamie: what about chrome a11y?
- # [04:04] <Jamie> davidb: not sure if chrome has caret nav/whether it uses webkit caret nav
- # [04:05] <Jamie> their a11y implementation is entirely separate though
- # [04:05] <davidb> they probably use webkit caret
- # [04:05] <davidb> how is their a11y impl coming along?
- # [04:05] <@tbsaunde> Jamie: yeah, I gues a better way to put it is couldn't you do all the other stuff you want to do with an injected dll but not have a virtual buffer?
- # [04:05] <Jamie> tbsaunde: forgetting about caret nav? probably, yes
- # [04:06] <Jamie> tbsaunde: although it's worth noting that communicating objects cross-process is actually really damned hard
- # [04:06] <Jamie> tbsaunde: so if it didn't involve moving focus (and NVDA has a separate review cursor, so this is a valid use case), it'd be very hard
- # [04:08] <Jamie> tbsaunde: using the paragraph with 50 links example, I'm still not sure how we'd do that. communicating raw speech stuff back to the screen reader isn't an option, so you'd have to try to communicate object "data"
- # [04:09] <@tbsaunde> Jamie: why not?
- # [04:09] <Jamie> we'd probably end up using something similar to our virtual buffer code, just without the pre-rendering
- # [04:09] <Jamie> tbsaunde: two reasons: 1. we'd have to duplicate the presentation logic between screen reader and in-process dll. 2. that doesn't cover braille
- # [04:12] <@tbsaunde> Jamie: can't you link the same code into two things? and couldn't you do something similar like sending stuff directly as braille
- # [04:14] <davidb> Jamie: what is foliot talking about escalating?
- # [04:16] <davidb> oh never mind actually
- # [04:16] <davidb> catch y'all later
- # [04:16] * Quits: davidb (davidb@moz-6F9F653A.dsl.bell.ca) (Quit: davidb)
- # [04:22] <Jamie> tbsaunde: NVDA is written in Python. Our in-proc code is written in C++
- # [04:22] <@tbsaunde> Jamie: oic, that would make it a little more tricky
- # [04:22] <Jamie> :)
- # [04:23] <Jamie> I guess the only way would be to rewrite the presentation code in C++. eeeewwwww.
- # [04:23] <@tbsaunde> unless you called out to c++ from python
- # [04:23] <Jamie> hmm?
- # [04:23] <@tbsaunde> I sort of hate saying it, but honestly I dislike dynically typed langauages enough I sort of prefer C++ sometimes
- # [04:24] <@tbsaunde> Jamie: same thing you said, so nvm
- # [04:24] <Jamie> we don't really abuse dynamic typing much, but it's not just the dynamic typing that makes Python easier
- # [04:25] <Jamie> there's just a lot less stuff to worry about
- # [04:25] <Jamie> With Python, I don't tend to have to think about the language much; I think mroe algorithmically. With C++, I have to worry about how I code it
- # [04:26] <@tbsaunde> Jamie: its not so much the abuse of dynamic typeing as not having a compiler to catch stupid crap for me
- # [04:26] <Jamie> NVDA's architecture is also a bit insane in terms of object orientation. We mutate classes after construction in certain cases to allow for overriding of specific behaviour
- # [04:27] <Jamie> tbsaunde: oh, sure. there are times when I miss compilers :)
- # [04:27] <Jamie> the number of times I've had to restart NVDA 4 or 5 times because I am having a bad coding day
- # [04:27] <@tbsaunde> Jamie: you mena change the methods of a class at runtime by fiddling with the byte code?
- # [04:27] <Jamie> no
- # [04:28] <Jamie> in brief, because we have to deal with so many different APIs, NVDA abstracts all accessibles as NVDAObjects. there are IAccessible NVDAObjects, UIA NVDAObjects, Java NVDAObjects, etc.
- # [04:28] <@tbsaunde> oic
- # [04:29] <Jamie> so let's say we're dealing with Gecko. we get an IAccessible NVDAObject and that's all fine. so we construct it so that we can query properties on the object; e.g. role, states, name, etc.
- # [04:30] * Joins: jhk (jiggy@8E6C34C1.A3F9767A.1C37C358.IP)
- # [04:30] <Jamie> then we have another level which checks for "overlay" classes. so we query the windowClassName of our NVDAObject (obj.windowClassName) and we see that it's MozillaWindowClass. then we see that it has ROLE_DOCUMENT, so we mutate the instance so that the mozilla.Document class is now part of the inheritance
- # [04:30] <@tbsaunde> fun
- # [04:31] <Jamie> there are also abstract behaviors such as Dialog. these are mix-in classes which can be applied to any API class. so they can be added as overlays as well
- # [04:31] <Jamie> If we didn't do it this way, we'd have to talk raw APIs a lot more and we'd have to have a lot more hard-coded classes; e.g. IAccessibleDialog, UIADialog, etc. This way, it's all handled without code duplication
- # [04:32] <Jamie> Also, app specific code and global plugins are given the chance ot add their own overlays, so users can customise object behaviour
- # [04:32] <Jamie> this is probably a hell of a lot more than you really wanted to know :)
- # [04:33] <@tbsaunde> Jamie: its fine, its kind of interesting
- # [04:34] * khuey|away is now known as khuey
- # [04:47] <Jamie> urg, I just tried to use ed commands in a wysiwyg editor
- # [04:55] <@tbsaunde> haha
- # [04:56] <@tbsaunde> I haven't used one of those in a long time
- # [05:24] <Jamie> ouch. lots of firefox crashes this morning
- # [05:28] <Jamie> ah, bug 714579
- # [05:28] <@firebot> Bug https://bugzilla.mozilla.org/show_bug.cgi?id=714579 cri, --, mozilla13, surkov.alexander, REOP, crash nsINode::OwnerDoc
- # [05:29] <Jamie> this is the only problem with using nightlies as my main browser :)
- # [05:29] <@surkov> yeah, I hope you don't get them too often
- # [05:29] <@surkov> Jamie: don't you check defunct state before you operate on object?
- # [05:30] <@surkov> or would you like object disconnected error when you call the method on defunct object?
- # [05:30] <@surkov> i.e. bug 539683
- # [05:30] <@firebot> Bug https://bugzilla.mozilla.org/show_bug.cgi?id=539683 nor, --, ---, nobody, NEW, defunct object methods should return CO_E_OBJECTNOTCONNECTED
- # [05:31] <Jamie> surkov: we don't check defunct state, no, but even if we did, there's still a possible race condition; i.e. we check defunct, it isn't, but then it goes defunct just as we call accName or similar
- # [05:32] <Jamie> it's also a major perf hit
- # [05:32] <@surkov> sure
- # [05:33] <Jamie> I don't really care whether you throw an error or not
- # [05:33] <@surkov> Jamie: did you said the object goes defunct when you call accName?
- # [05:33] <Jamie> well, NVDA doesn't care
- # [05:33] <Jamie> surkov: mm, is this related to the crash?
- # [05:33] <@surkov> yes
- # [05:34] <@surkov> we didn't check defunct either
- # [05:34] <Jamie> I'm not sure what triggers the crash to be honest
- # [05:34] <@surkov> that bug 539683 is going to fix all problems, but until it's fixed from time to time we introduce new ones
- # [05:34] * Jamie nods
- # [05:35] <@surkov> null-check makes a crash
- # [05:35] <Jamie> surkov: I mean taht checking defunct state on our side wouldn't help because of race conditions
- # [05:35] <@surkov> what is a race conditions?
- # [05:36] <Jamie> surkov: well, in order for us to avoid this, we'd have ot check for defunct state before every call we make
- # [05:36] * @tbsaunde tries to figure out how to explain race conditions gives up and looks at wikipedia
- # [05:37] <Jamie> surkov: scenario: 1. NVDA wants to speak an object, so it wants name, role, states and value.
- # [05:37] <Jamie> surkov: 2. NVDA calls states to check for defunct state.
- # [05:37] <Jamie> surkov: 3. object goes defunct after we call states
- # [05:37] <Jamie> surkov: 4. NVDA calls accName
- # [05:37] <Jamie> crash
- # [05:37] <Jamie> (accName is just an example; I don't know which one is causing the crash)
- # [05:38] <@surkov> Jamie: technically it shouldn't go defunct if call a method
- # [05:38] <@surkov> do you observe this behavior?
- # [05:38] <Jamie> no, I'm just pointing out that it wouldn't fix the problem
- # [05:38] <Jamie> surkov: an object can go defunct at any time, no?
- # [05:38] <@surkov> but object should never go defunct when you call a method
- # [05:38] <@surkov> no
- # [05:38] <Jamie> so when can they go defunct?
- # [05:38] <@surkov> if that happens then this is something we should address I think
- # [05:39] <@surkov> we are running loop from time to time and invalidate our tree, at this point object can go defunct
- # [05:39] <@tbsaunde> surkov: en.wikipedia.org/wiki/Race_condition\#Computing
- # [05:40] <@surkov> basically from your presspective object goes defunct after you received hide event
- # [05:40] <Jamie> surkov: there can still be a few ms between when we call states and when we call, say, name
- # [05:40] <Jamie> surkov: what if your loop invalidates the object between those two calls?
- # [05:40] <@surkov> you do that async?
- # [05:40] <@surkov> or in own process?
- # [05:40] <Jamie> everything is async cross-process
- # [05:40] <@surkov> ok
- # [05:41] <@surkov> you're right then
- # [05:41] <Jamie> sorry, I should have pointed that out in the first place :)
- # [05:41] <@surkov> I should have remember that :)
- # [05:41] <Jamie> surkov: I think this is our out-proc code, not in-proc. we don't check defunct in-proc either, though
- # [05:42] <Jamie> calling states before every call is fairly expensive
- # [05:42] <@surkov> agree
- # [05:42] <@surkov> we should handle that on our side
- # [05:43] <Jamie> surkov: yup, definitely our out-proc code, since VBufBackend_gecko_ia2.dll isn't in the stack trace
- # [05:43] <@surkov> I see
- # [05:43] <Jamie> I'd say we're probably querying the old object just after it loses focus (before we get the new focus event)
- # [05:46] <@tbsaunde> surkov: so, do you want to land what you have for bug 714579 or do you want to wait for new stlye api
- # [05:46] <@firebot> Bug https://bugzilla.mozilla.org/show_bug.cgi?id=714579 cri, --, mozilla13, surkov.alexander, REOP, crash nsINode::OwnerDoc
- # [05:47] <@surkov> tbsaunde: we might need to wait too long :)
- # [06:02] * Joins: hub (hub@83874EA1.EB7C1AF9.6F478678.IP)
- # [06:02] * ChanServ sets mode: +o hub
- # [06:14] * Quits: @surkov (surkov@74A597FA.C2B307F0.34044A7F.IP) (Ping timeout)
- # [06:18] * Joins: surkov (surkov@1A667863.26ED71FF.34044A7F.IP)
- # [06:18] * ChanServ sets mode: +o surkov
- # [06:22] * Quits: jhk (jiggy@8E6C34C1.A3F9767A.1C37C358.IP) (Connection reset by peer)
- # [06:34] * Quits: peteb-away (ptbrunet@moz-E9B02845.austin.res.rr.com) (Ping timeout)
- # [06:54] * khuey is now known as khuey|away
- # [07:04] <@firebot> New Core - Disability Access APIs bug 728127 filed by surkov.alexander@gmail.com.
- # [07:04] <@firebot> Bug https://bugzilla.mozilla.org/show_bug.cgi?id=728127 nor, --, ---, surkov.alexander, ASSI, use GetComputedStyle carefully
- # [07:41] * Joins: jhk (jiggy@8E6C34C1.A3F9767A.1C37C358.IP)
- # [07:57] <@firebot> surkov.alexander@gmail.com changed the Status on bug 694818 from NEW to RESOLVED.
- # [07:57] <@firebot> surkov.alexander@gmail.com set the Resolution field on bug 694818 to DUPLICATE of bug 714579.
- # [07:57] <@firebot> Bug https://bugzilla.mozilla.org/show_bug.cgi?id=694818 cri, --, ---, nobody, RESO DUPLICATE, crash mozilla::a11y::FocusManager::IsFocused
- # [07:57] <@firebot> Bug https://bugzilla.mozilla.org/show_bug.cgi?id=714579 cri, --, mozilla13, surkov.alexander, REOP, crash nsINode::OwnerDoc
- # [08:11] <@firebot> surkov.alexander@gmail.com changed the Status on bug 634200 from ASSIGNED to RESOLVED.
- # [08:11] <@firebot> surkov.alexander@gmail.com set the Resolution field on bug 634200 to WORKSFORME.
- # [08:11] <@firebot> Bug https://bugzilla.mozilla.org/show_bug.cgi?id=634200 cri, --, ---, surkov.alexander, RESO WORKSFORME, crash [@ nsIFrame::GetStyleVisibility() ]
- # [08:15] * Quits: jhk (jiggy@8E6C34C1.A3F9767A.1C37C358.IP) (Connection reset by peer)
- # [08:16] <@firebot> surkov.alexander@gmail.com changed the Status on bug 646355 from NEW to RESOLVED.
- # [08:16] <@firebot> surkov.alexander@gmail.com set the Resolution field on bug 646355 to WORKSFORME.
- # [08:16] <@firebot> Bug https://bugzilla.mozilla.org/show_bug.cgi?id=646355 cri, --, ---, nobody, RESO WORKSFORME, crash [@ @0x0 | nsAccessibleWrap::UpdateSystemCaret()]
- # [08:18] * Joins: jhk (jiggy@8E6C34C1.A3F9767A.1C37C358.IP)
- # [09:16] * Quits: Jamie (jamie@moz-CA26021.jantrid.net) (Quit: leaving)
- # [10:19] * Quits: jhk (jiggy@8E6C34C1.A3F9767A.1C37C358.IP) (Connection reset by peer)
- # [10:19] * Joins: victorporof (victorporo@C092FEB2.1C233438.79933D60.IP)
- # [10:21] * Quits: @hub (hub@83874EA1.EB7C1AF9.6F478678.IP) (Ping timeout)
- # [10:26] * Quits: victorporof (victorporo@C092FEB2.1C233438.79933D60.IP) (Quit: Linkinus - http://linkinus.com)
- # [10:26] * Joins: jhk (jiggy@8E6C34C1.A3F9767A.1C37C358.IP)
- # [10:31] * Quits: @surkov (surkov@1A667863.26ED71FF.34044A7F.IP) (Quit: surkov)
- # [10:37] * Joins: victorporof (victorporo@C092FEB2.1C233438.79933D60.IP)
- # [12:00] * Joins: MarcoZ (marco.zehe@moz-DF186A16.dip.t-dialin.net)
- # [12:00] * ChanServ sets mode: +o MarcoZ
- # [12:23] * Quits: jhk (jiggy@8E6C34C1.A3F9767A.1C37C358.IP) (Ping timeout)
- # [12:25] * Joins: askalski (akuda@moz-4C8A107E.pool85-48-91.dynamic.orange.es)
- # [12:25] * ChanServ sets mode: +o askalski
- # [12:26] <@askalski> morning!
- # [12:29] * Joins: jhk (jiggy@8E6C34C1.A3F9767A.1C37C358.IP)
- # [12:39] * Joins: surkov (surkov@1A667863.26ED71FF.34044A7F.IP)
- # [12:39] * ChanServ sets mode: +o surkov
- # [12:56] <@askalski> surkov, how to appy a patch? I get abort: patch bug727163-0.patch not in series
- # [12:56] <@firebot> Bug https://bugzilla.mozilla.org/show_bug.cgi?id=727163 nor, --, ---, nfroyd, ASSI, fix compiler warnings in accessible/
- # [12:56] <@surkov> askalski: you should do hg qimport patchname; hg qpush
- # [12:57] <@askalski> surkov, ok, thanks
- # [12:57] <@surkov> yw
- # [13:10] <@askalski> surkov, how to rebuild just a11y ? and can I pass "warnings as errors" to make?
- # [13:11] <@askalski> MarcoZ, hi
- # [13:11] <@surkov> askalski: cd objdir/accessible; make; cd ../toolkit/library; make
- # [13:11] <@askalski> surkov, thanks
- # [13:11] <@askalski> surkov, would make -Werror do the trick?
- # [13:12] <@MarcoZ> Hi askalski!
- # [13:12] <@MarcoZ> Hi surkov!
- # [13:18] <@askalski> MarcoZ, do you know the way
- # [13:18] <@askalski> MarcoZ, to add some compiler flags to a single build or to Makefile
- # [13:19] <@askalski> I want to build a11y with -Werror flag
- # [13:20] <@MarcoZ> askalski: Sorry, not my field of expertise. Ask in #developers maybe?
- # [13:20] <@MarcoZ> askalski: There are MOZCONFIG options to add compiler options IIRC. Look on MDC in the "Configuring build options" sections.
- # [13:22] * Quits: ehsan (ehsan@F0B20A8D.8458880F.57F33CED.IP) (Input/output error)
- # [13:27] <@surkov> hi, MarcoZ
- # [13:50] * Quits: victorporof (victorporo@C092FEB2.1C233438.79933D60.IP) (Quit: Leaving...)
- # [14:23] * Quits: @jprmc (jprmc@8F761026.B5EF4AF6.E96CA9D8.IP) (Ping timeout)
- # [14:25] <@askalski> surkov, I was thinking about taking a break from text algoritms, and doing 606080
- # [14:26] <@askalski> surkov, I am looking for someone experienced to tell me "yeah, go with range tree, write it"
- # [14:26] <@surkov> askalski: sure
- # [14:26] <@askalski> surkov, cause honestly, it might not necessairly do the improvement
- # [14:26] <@surkov> askalski: ping neil (neil@httl.net)
- # [14:26] <@askalski> surkov, but until I try I never know
- # [14:26] <@surkov> that's all about investigations
- # [14:27] <@surkov> sometimes you should just try, if theory says it's faster then it's worth to try and see how it works in real life
- # [14:28] <@surkov> for 32bit unique ids I think you need some help with interfaces to measure it
- # [14:28] * Quits: jhk (jiggy@8E6C34C1.A3F9767A.1C37C358.IP) (Connection reset by peer)
- # [14:28] <@askalski> surkov, ok, so I should write to Neil an e-mail, right?
- # [14:28] <@surkov> askalski: just cc him and ask for advice
- # [14:29] <@askalski> cc in bug, right?
- # [14:29] <@surkov> yes
- # [14:31] <@askalski> ok
- # [14:32] * Joins: ehsan (ehsan@F0B20A8D.8458880F.57F33CED.IP)
- # [14:32] <@tbsaunde> surkov: askalski mkae -C foo/bar/baz is a bit nicer than cd foo/bar/baz/ && mkae fwiw
- # [14:33] <@tbsaunde> askalski: CFLAGS="blah" make might work or you could just redirect stdout and see if gcc spews any warnings ;)
- # [14:33] <@surkov> tbsaunde: ok
- # [14:36] <@askalski> tbsaunde, CFLAGS works, but Werror=something was defined, so another Werror is ignored. and routing cstdout doesn't work :(
- # [14:36] <@askalski> tried these two
- # [14:37] <@tbsaunde> askalski: what is the problem with redirecting stdout? that is the general way look for warnings
- # [14:37] <@askalski> tbsaunde, do you know how to route stderr?
- # [14:38] * Joins: jprmc (jprmc@F2D29657.F60B0462.67AC9B1.IP)
- # [14:38] * ChanServ sets mode: +o jprmc
- # [14:40] <@tbsaunde> askalski: same way usually do, make >/dev/null
- # [14:40] * Quits: ehsan (ehsan@F0B20A8D.8458880F.57F33CED.IP) (Input/output error)
- # [14:40] <@tbsaunde> askalski: why do you need to change stderr? what's wrong with it still going to the terminal?
- # [14:42] <@askalski> tbsaunde, I don't know. I do make clean && make | grep 'warning' and I get all the "make[3] ..." stuff
- # [14:43] <@tbsaunde> askalski: I'm not sure why, but that is different from what I am suggesting doing so...
- # [14:44] <@askalski> ha!
- # [14:44] <@askalski> ha! make 2>&1 1> /dev/null | grep 'warning' that's the magic I was looking for
- # [14:46] <@askalski> tbsaunde, ah, you mean that if I redirect stdout all I am left with are warnings
- # [14:46] <@askalski> neat
- # [14:46] <@askalski> I didn't get you
- # [14:46] <@askalski> I was looking for warnings in output...
- # [14:46] <@tbsaunde> askalski: uh, no, I'm pretty sure that doesn't do what you want because you end up redirecting both stdout and stderr to /dev/null
- # [14:47] <@askalski> tbsaunde, but it does return the warnings from grep
- # [14:47] <@tbsaunde> askalski: weird, I'm not sure why
- # [14:48] <@tbsaunde> I don't understand redirection well enough to completely get why that works
- # [14:49] <@tbsaunde> but seriously make > /dev/null is what I generally use and it works fiarly well
- # [14:51] <@askalski> tbsaunde, thanks, it's good to get the tools right
- # [14:52] <@askalski> tbsaunde, my apologies for the lags in my answer, I thought I got enlighted on text algorithms, and I was looking for a counterexample
- # [14:52] <@askalski> guess I wont get nobel prize this year
- # [14:53] * @askalski goes for lunch
- # [14:53] * Joins: davidb (davidb@F2D29657.F60B0462.67AC9B1.IP)
- # [14:53] * ChanServ sets mode: +qo davidb davidb
- # [15:04] <@MarcoZ> Hi davidb, tbsaunde!
- # [15:04] <@davidb> hi MarcoZ! all!
- # [15:10] <@davidb> ok surkov you're out windows 8 a11y lead https://wiki.mozilla.org/Windows8
- # [15:11] * khuey|away is now known as khuey
- # [15:11] <@surkov> davidb: really out?
- # [15:11] <@surkov> lost in translation
- # [15:12] <@davidb> typo
- # [15:12] <@davidb> sorry
- # [15:12] <@davidb> "our"
- # [15:12] <@davidb> :)
- # [15:12] <@surkov> :)
- # [15:14] <@askalski> hi davidb !
- # [15:14] <@davidb> hi askalski!
- # [15:15] <@firebot> bolterbugz@gmail.com changed the Status on bug 626642 from NEW to RESOLVED.
- # [15:16] <@firebot> bolterbugz@gmail.com set the Resolution field on bug 626642 to INCOMPLETE.
- # [15:16] <@firebot> Bug https://bugzilla.mozilla.org/show_bug.cgi?id=626642 nor, --, ---, nobody, RESO INCOMPLETE, Map aria-hidden true/false to specific events
- # [15:18] <@davidb> surkov: are you worried a non-flushing version would be too incorrect or something (bug 714579)?
- # [15:18] <@firebot> Bug https://bugzilla.mozilla.org/show_bug.cgi?id=714579 cri, --, mozilla13, surkov.alexander, REOP, crash nsINode::OwnerDoc
- # [15:19] <@surkov> davidb: no more incorrect than our accessible tree :)
- # [15:19] <@davidb> that's true
- # [15:19] <@davidb> and a good argument not to flush
- # [15:19] <@surkov> we don't guarantee they get updated information, we should show them what user sees on screen
- # [15:19] <@davidb> right
- # [15:19] <@davidb> ok
- # [15:20] <@davidb> surkov: why doesn't the non-flushing version live in layout>?
- # [15:20] <@surkov> if frame doesn't know about new styles then I bet user doesn't see that too
- # [15:20] <@davidb> actually -- i need to read the patch closely
- # [15:20] <@surkov> davidb: it's not done I guess yet
- # [15:21] <@surkov> I think inside layout they don't need string representation of styles, they just use styles
- # [15:21] * Joins: victorporof (victorporo@717ECBD0.6465CCD9.79933D60.IP)
- # [15:21] <@surkov> os I filed bug to add this capability, hope somebody will come with a fix
- # [15:22] <@davidb> I wonder if anyone else gets bitten by this flush
- # [15:24] <@tbsaunde> davidb: I'd tend doubt that many people outside of layout have a reason to want the information
- # [15:24] <@davidb> tbsaunde: I'm curious about the callers of getcomputedstyle.
- # [15:24] <@tbsaunde> davidb: well, you can grep ;)
- # [15:25] <@davidb> indeed
- # [15:25] <@davidb> want to get through bug mail first :)
- # [15:25] <@davidb> dxr is useful when it works
- # [15:25] <@askalski> can someone tell me what different boxes in patch review mean?
- # [15:25] <@davidb> (our static analysis charged code search)
- # [15:25] <@davidb> askalski: during "review"
- # [15:25] <@davidb> ?
- # [15:26] <@askalski> davidb, Alex added me to review a patch
- # [15:26] <@davidb> if you click in the patch you can annotate a comment.
- # [15:26] <@davidb> patch source code
- # [15:26] <@davidb> ok describe "boxes" :)
- # [15:26] <@askalski> davidb, I clicked "splinter review" and filled the draft. I dont know what flags means
- # [15:27] <@davidb> oh so wherever your email is, you can change it to + or - or empty
- # [15:27] <@askalski> there is my name as requestee, so I am not sure if I'm in the right place
- # [15:27] <@davidb> that's correct
- # [15:27] <@davidb> surkov is the requester
- # [15:27] <@askalski> if I say "good start, but incomplete" I should set the "-", right?
- # [15:27] <@askalski> or "empty"
- # [15:27] <@davidb> empty is preferred.
- # [15:27] <@askalski> ok
- # [15:28] <@tbsaunde> davidb: yeah, unless you mean excuse me while I find a garbage can to puke in now that I'm done with this :/
- # [15:28] <@askalski> davidb, once i set empty, my e-mail disappears
- # [15:28] <@askalski> is that ok?
- # [15:28] <@tbsaunde> davidb: it suffers from being a web app
- # [15:28] <@davidb> tbsaunde: clearly you've seen my patches
- # [15:29] <@davidb> askalski: yes
- # [15:29] * Joins: clown (clown@67828CC7.C1A51174.9D42CF23.IP)
- # [15:30] * Joins: peteb-away (ptbrunet@moz-E9B02845.austin.res.rr.com)
- # [15:30] <@askalski> davidb, done
- # [15:30] <@firebot> askalski@mozilla.com cancelled review?(askalski@mozilla.co m) for attachment 597093 on bug 727163.
- # [15:30] <@firebot> Bug https://bugzilla.mozilla.org/show_bug.cgi?id=727163 nor, --, ---, nfroyd, ASSI, fix compiler warnings in accessible/
- # [15:31] <@davidb> surkov, tbsaunde, eeejay, if you would like to get email updates from other team members, you can sign up to http://benjamin.smedbergs.us/weekly-updates.fcgi/
- # [15:31] <@davidb> and add yourself to the accessibility team
- # [15:31] <@davidb> and of course you can post updates :)
- # [15:31] <@surkov> ok
- # [15:32] <@surkov> is anybody posting, i.e. doesn't make sense to cc there?
- # [15:32] <@surkov> I like spamming :)
- # [15:32] <@surkov> bugzilla is not enough for me :)
- # [15:34] <@firebot> surkov.alexander@gmail.com requested review from marco.zehe@googlema il.com for attachment 598222 on bug 677154.
- # [15:34] <@firebot> Bug https://bugzilla.mozilla.org/show_bug.cgi?id=677154 maj, P1, ---, surkov.alexander, ASSI, Detached document accessibility tree
- # [15:34] <@surkov> MarcoZ: are you dancing? :)
- # [15:35] <@surkov> so small fix, so much pain to find a problem
- # [15:37] <@davidb> i bet
- # [15:37] * Quits: sgimeno (santiago.g@A922400D.48362F18.BCC03455.IP) (Quit: KVIrc 4.0.2 Insomnia http://www.kvirc.net/)
- # [15:37] <@davidb> surkov: what derives from nsHTMLScrollFrame?
- # [15:38] <@surkov> davidb: layout guys knows :)
- # [15:38] <@surkov> or I miss your questions
- # [15:38] <@davidb> http://dxr.lanedo.com/search.cgi?tree=mozilla-central&request_time=1329489081007&derived=nsHTMLScrollFrame
- # [15:38] <@davidb> that's opposite what i want lol
- # [15:38] <@davidb> oh no, it is right
- # [15:39] <@davidb> but i can't believe it
- # [15:39] <@surkov> yeah, that's what you asked
- # [15:39] <@surkov> why not?
- # [15:39] <@surkov> list is scrollable thing
- # [15:39] <@davidb> only nsListControlFrame?
- # [15:39] <@surkov> by default
- # [15:39] <@surkov> ah
- # [15:39] * @davidb shrugs
- # [15:40] <@davidb> aha http://dxr.lanedo.com/search.cgi?tree=mozilla-central&request_time=1329489178718&callers=getcomputedstyle
- # [15:42] * @tbsaunde suprised people think using a web app is easier than grep -ri nsHTMLScrollFrame layout/ | grpe public
- # [15:42] <@MarcoZ> surkov: Just done! :D
- # [15:42] <@MarcoZ> dancing, that is.
- # [15:42] <@surkov> :)
- # [15:42] <@davidb> tbsaunde: less false hits
- # [15:43] <@davidb> surkov: really nice find.
- # [15:43] <@MarcoZ> surkov: Let me look!
- # [15:43] <@surkov> thx
- # [15:43] <@tbsaunde> davidb: in theory probably, but other than the declaration of nsHTMLScrollFrame how would you have a false hit
- # [15:43] <@davidb> (also this rings a bell for some reason - scrollable frames and focus)
- # [15:44] <@tbsaunde> I supose it doesn't handle the case of people using non-public inheritance, but that's crazy
- # [15:44] <@askalski> davidb, I asked Alexander whether to go with the ranged trees, and he told me to ask one person more, so I wait for answer on 606080
- # [15:44] <@davidb> tbsaunde: i'd have to make stuff up, like nsHTMLScrollFrame(int aRepublic)
- # [15:45] <@davidb> tbsaunde: but I am a grep fan
- # [15:45] <@davidb> askalski: gorse, have you asked tbsaunde?
- # [15:45] <@davidb> gorse?
- # [15:45] <@askalski> what gorse means?
- # [15:45] <@davidb> silly autocorrect
- # [15:45] <@davidb> "gosh"
- # [15:45] * @davidb looks
- # [15:46] <@tbsaunde> davidb: I don't follow the make stuff up thing
- # [15:47] <@askalski> davidb, I guess tb told me to go with it. anyway, let me know what to do. options: 1) back on LED, writing Ukkonen 2) writing trees 3) look for some other bug
- # [15:47] <@davidb> tbsaunde: i had "public" as part of an arg name
- # [15:47] <@tbsaunde> surkov: what about the case Style::foo() returns empty string?
- # [15:47] <@davidb> askalski: gut reaction is 2.
- # [15:47] * Joins: ehsan (ehsan@F2D29657.F60B0462.67AC9B1.IP)
- # [15:47] <@davidb> but I'm not sure how I can make a reasoned choice.
- # [15:47] <@tbsaunde> davidb: oh, yeah, mind just autofiltered :)
- # [15:48] <@davidb> askalski: did you hear back about LED?
- # [15:48] * @davidb tries to recall
- # [15:49] * @tbsaunde will go a bit of his way to avoid web apps
- # [15:49] <@askalski> davidb, not from the professor. some algorithm guys either say it's impossible without suffix trees, or state that I'll be fine with Ukkonen
- # [15:49] <@davidb> askalski: Ukkonen is complex right?
- # [15:49] <@askalski> davidb, not as much as as what I wrote already
- # [15:49] <@davidb> oh
- # [15:50] <@askalski> but the first part took me like 20 hours, more than I expected
- # [15:50] <@davidb> it happens
- # [15:50] <@askalski> I guess 1) is OK too.
- # [15:50] <@davidb> yeah
- # [15:50] <@askalski> I'm on the subject, I don't expect anyone to break it other way
- # [15:50] <@askalski> and I have sources gathered
- # [15:50] <@davidb> guys did we reach out to mozilla developers about a pool algorithm?
- # [15:51] <@askalski> davidb, I have one question though. How to check if these algorithms are not patented?
- # [15:51] <@davidb> oh man
- # [15:51] <@davidb> no idea
- # [15:51] <@askalski> that is something I thought just in the morning
- # [15:51] <@davidb> askalski: you might ask jrmuizel
- # [15:51] <@askalski> in EU we don't think about it, it's illegal to patent an algorithm
- # [15:51] <@askalski> but in US it's an issue, right?
- # [15:51] <@surkov> tbsaunde: does it return?
- # [15:51] <@davidb> i think so
- # [15:51] <@askalski> where I will find jrmuizel?
- # [15:52] <@davidb> he is usually on IRC
- # [15:52] <@davidb> he sits near me, i'll go poke him
- # [15:52] <@askalski> because for example parts of Habanero are patented in USA
- # [15:52] <@askalski> but as long as we do our stuff in EU we don't care
- # [15:53] <@askalski> surkov, take a look on my review and tell me your thoughts
- # [15:53] <@askalski> surkov, it's my first one :), thanks
- # [15:54] <@tbsaunde> askalski: so, if you use stuff that there is prior art for in the publicc domain I'm pretty sure your safe
- # [15:54] <@tbsaunde> beyound that not really sure (arg i hate IP law)
- # [15:54] <@surkov> askalski: honestly I'm not sure what is review status
- # [15:54] <@askalski> tbsaunde, well, I am not a lawyer, I guess that the best it would be that if I just mailed someone in Legal Dep. and tell them "we added algorithms X and Y to our code"
- # [15:55] <@askalski> surkov, review status?
- # [15:55] <@tbsaunde> askalski: the thing is I believe people have gotten pattents on things like a fancy linked list
- # [15:55] <@surkov> askalski: is r+ or r- and why do you talk about other warning :)
- # [15:55] <@askalski> tbsaunde, yes, I've seend that :D
- # [15:56] <@tbsaunde> askalski: and we don't tell legal department every time we add any code so I'm not really sure
- # [15:56] <@askalski> surkov, oh, I must have misunderstood the bug title
- # [15:56] <@askalski> surkov, I thought "all warnings"
- # [15:56] <@surkov> I don't think so, at least patch doesn't fix all of them
- # [15:56] <@askalski> surkov, and David adviced me to set "empty"
- # [15:57] <@surkov> that should be bad summary
- # [15:57] <@surkov> askalski: what does "empty" means?
- # [15:57] <@askalski> surkov, so my reply is bad, or "bad summary" is something else
- # [15:57] <@askalski> surkov, I had options -, + and "empty"
- # [15:57] <@surkov> sure
- # [15:57] <@surkov> but are you fine with patch
- # [15:57] <@askalski> surkov, my conversation with David is up there
- # [15:58] <@surkov> if it fixes warnings then ok it works I say
- # [15:58] <@askalski> surkov, no, I thought that it should fix all
- # [15:58] <@askalski> surkov, and it's a good start
- # [15:58] <@firebot> marco.zehe@googlemail.com granted review for attachment 598222 on bug 677154.
- # [15:58] <@firebot> Bug https://bugzilla.mozilla.org/show_bug.cgi?id=677154 maj, P1, ---, surkov.alexander, ASSI, Detached document accessibility tree
- # [15:58] <@askalski> surkov, but now that I see it, I must have got it wrong
- # [15:58] <@davidb> askalski: when we are aware of a patent, we respect it.
- # [15:58] <@surkov> I think we can take this patch since it makes improvements and fix bug summary and say it's fixed
- # [15:59] <@askalski> surkov, ok, should I change it?
- # [15:59] <@surkov> please, add a comment
- # [15:59] <@askalski> davidb, yes, but as tbsaunde said, pretty much everything is patented
- # [15:59] <@surkov> askalski: for your findings please file another bug(bugs)
- # [15:59] <@askalski> ok
- # [16:00] <@surkov> askalski: yeah, you should look at bug description: "Just a couple of signed-vs-unsigned comparisons."
- # [16:00] <@firebot> askalski@mozilla.com granted review for attachment 597093 on bug 727163.
- # [16:00] <@firebot> Bug https://bugzilla.mozilla.org/show_bug.cgi?id=727163 nor, --, ---, nfroyd, ASSI, fix compiler warnings in accessible/
- # [16:00] <@askalski> surkov, sorry, my mistake
- # [16:00] <@surkov> he was just too lazy to provide descriptive summary
- # [16:01] <@surkov> sometimes people does it :)
- # [16:01] <@surkov> when you know the problem and have a patch then you don't spend a time to describe it well
- # [16:01] <@tbsaunde> askalski: davidb the even better one is the pattent on a stick ;)
- # [16:01] <@askalski> tbsaunde, no, linked list is my favorite
- # [16:03] <@davidb> askalski: we (engineers) don't spend any time looking for patents.
- # [16:03] <@askalski> davidb, I understand, that's why I came up with idea of just reporting to Legal, so they can run it
- # [16:03] <@askalski> davidb, "idealistic" of course
- # [16:04] <@davidb> i probably wouldn't bother
- # [16:04] <@davidb> do you know there is a patent?
- # [16:04] <@davidb> i'm not asking you to find one
- # [16:04] <@davidb> just whether you know already there is one
- # [16:04] <@askalski> davidb, no
- # [16:05] <@askalski> davidb, I don't
- # [16:05] <@davidb> ok
- # [16:05] <@davidb> if you hear there is one definitely follow up to legal.
- # [16:05] <@askalski> davidb, sure. I wouldn't work on it if I knew about some kind of pattent
- # [16:06] <@davidb> ok great
- # [16:06] <@askalski> I know some kind of trees in computational geometry are patented (I remember from classes)
- # [16:06] <@askalski> but these text algorithms are usually from 70s and 80s, before the whole patent epidemy
- # [16:06] <@davidb> good
- # [16:07] <@davidb> i feel like this channel needs a shower, after all this patent talk :)
- # [16:07] <@askalski> yes :D
- # [16:09] <@tbsaunde> surkov: not ure what you mean can it return
- # [16:10] <@surkov> tbsaunde: about Style()->Foo()
- # [16:10] <@surkov> style always has a value
- # [16:11] <@tbsaunde> surkov: yeah, I gues that makes sense
- # [16:15] <@firebot> trev.saunders@gmail.com requested review from marco.zehe@googlemail.com for attachment 597861 on bug 714579.
- # [16:15] <@firebot> Bug https://bugzilla.mozilla.org/show_bug.cgi?id=714579 cri, --, mozilla13, surkov.alexander, REOP, crash nsINode::OwnerDoc
- # [16:15] * @tbsaunde lazy
- # [16:22] <@askalski> going for a gym break, be back evening (or at least so I hope
- # [16:26] * Quits: @askalski (akuda@moz-4C8A107E.pool85-48-91.dynamic.orange.es) (Ping timeout)
- # [16:28] <@firebot> trev.saunders@gmail.com granted review for attachment 597707 on bug 727735.
- # [16:28] <@firebot> Bug https://bugzilla.mozilla.org/show_bug.cgi?id=727735 nor, --, ---, surkov.alexander, NEW, don't use computed style for width detection in IsProbablyForLayout
- # [16:30] <@firebot> marco.zehe@googlemail.com granted review for attachment 597861 on bug 714579.
- # [16:30] <@firebot> Bug https://bugzilla.mozilla.org/show_bug.cgi?id=714579 cri, --, mozilla13, surkov.alexander, REOP, crash nsINode::OwnerDoc
- # [16:33] <@tbsaunde> surkov: where do you think the include should added in the patch ofr bug 726283?
- # [16:33] <@firebot> Bug https://bugzilla.mozilla.org/show_bug.cgi?id=726283 nor, --, ---, jigneshhk1992, ASSI, consider using NS_GetContentList() directly in nsHTMLRadioButtonAccessible::GetPositionAndSizeIntern
- # [16:35] <@surkov> tbsaunde: technically we have two blocks: a11y block and other block, they have alphabetical order
- # [16:35] <@surkov> so to the 2nd list in your case
- # [16:36] <@tbsaunde> surkov: yeah, was wondering if you wanted to keep other interfaces and normal headers eperate
- # [16:37] <@surkov> I realized that's too complicated :)
- # [16:37] <@surkov> a11y and non a11y makes sense
- # [16:37] <@tbsaunde> ok
- # [16:38] <@davidb> surkov: they are discussing pdf.js a11y in #devtools in case you are interested
- # [16:38] <@davidb> et. al.
- # [16:38] * Joins: jhk (jiggy@8E6C34C1.A3F9767A.1C37C358.IP)
- # [16:39] <@surkov> thx
- # [16:39] <@davidb> np
- # [16:43] <@MarcoZ> surkov: Glad to report that the first patch in bug 714579 fixes the crashes for me. Today's nightly FINALLY came out.
- # [16:43] <@firebot> Bug https://bugzilla.mozilla.org/show_bug.cgi?id=714579 cri, --, mozilla13, surkov.alexander, REOP, crash nsINode::OwnerDoc
- # [16:43] <@surkov> nice
- # [16:43] <@davidb> phew
- # [16:44] <@firebot> trev.saunders@gmail.com cancelled review?(trev.saunders@gmail .com) for attachment 597326 on bug 726283.
- # [16:44] <@firebot> Bug https://bugzilla.mozilla.org/show_bug.cgi?id=726283 nor, --, ---, jigneshhk1992, ASSI, consider using NS_GetContentList() directly in nsHTMLRadioButtonAccessible::GetPositionAndSizeIntern
- # [16:53] <@davidb> (surkov it moved to #pdfjs)
- # [16:53] * Joins: hub (hub@moz-E2FCA694.figuiere.net)
- # [16:53] * ChanServ sets mode: +o hub
- # [16:53] <@surkov> :)
- # [16:56] <@davidb> MarcoZ: can you try out pdf.js and report at bug 727819?
- # [16:56] <@firebot> Bug https://bugzilla.mozilla.org/show_bug.cgi?id=727819 nor, --, ---, nobody, NEW, Make PDF.js accessible
- # [16:56] <@davidb> http://mozilla.github.com/pdf.js/web/viewer.html
- # [17:01] * Quits: peteb-away (ptbrunet@moz-E9B02845.austin.res.rr.com) (Ping timeout)
- # [17:04] <@firebot> surkov.alexander@gmail.com requested review from marco.zehe@googlema il.com for attachment 598242 on bug 727940.
- # [17:04] <@firebot> surkov.alexander@gmail.com changed the Status on bug 727940 from NEW to ASSIGNED.
- # [17:04] <@firebot> Bug https://bugzilla.mozilla.org/show_bug.cgi?id=727940 nor, --, ---, surkov.alexander, ASSI, Windows app can't read Firefox setting for hardware acceleration (D2D)
- # [17:09] <@firebot> marco.zehe@googlemail.com granted review for attachment 598242 on bug 727940.
- # [17:10] <@tbsaunde> surkov: shouldn't you make that code run only on windows?
- # [17:11] <@surkov> tbsaunde: nsIGFXInfo is available on windows only, now. But they said nothing about its future on other platforms :)
- # [17:11] <@surkov> I could ifdef it but not sure if it's a big deal
- # [17:12] <@surkov> we just rely on interface
- # [17:14] <@tbsaunde> surkov: so we don't get a service at all on !windows?
- # [17:15] <@MarcoZ> davidb: Done.
- # [17:16] <@surkov> tbsaunde: I guess so
- # [17:17] <@davidb> MarcoZ: thanks
- # [17:17] <@surkov> I saw that somewhere in the code
- # [17:17] <@surkov> but maybe I'm wrong
- # [17:18] <@tbsaunde> surkov: ok, I'll go dig through gfx land later and make sure
- # [17:18] <@surkov> tbsaunde: http://mxr.mozilla.org/mozilla-central/source/toolkit/content/aboutSupport.js#251
- # [17:18] <@surkov> thank you!
- # [17:27] * Joins: richardschwerdtfeger (richardsch@moz-130F1494.lightspeed.austtx.sbcglobal.net)
- # [17:40] * Quits: victorporof (victorporo@717ECBD0.6465CCD9.79933D60.IP) (Quit: Linkinus - http://linkinus.com)
- # [17:43] * Joins: victorporof (victorporo@717ECBD0.6465CCD9.79933D60.IP)
- # [17:56] * Quits: victorporof (victorporo@717ECBD0.6465CCD9.79933D60.IP) (Quit: Linkinus - http://linkinus.com)
- # [17:58] * Joins: jamesn (jamesn@moz-34C566A2.oracle.com)
- # [18:00] <@surkov> davidb, marcoz, it sounds like Firefox users prefers NVDA
- # [18:01] <@MarcoZ> Yeah!
- # [18:01] <@davidb> certainly early adopters (nightly/aurora users)
- # [18:02] <@surkov> rihgt
- # [18:02] <@surkov> stable versions are on JAWS
- # [18:03] <@davidb> we'll see
- # [18:03] * Joins: drexler (chatzilla@moz-BD26ECCD.hsd1.vt.comcast.net)
- # [18:04] * Quits: jamesn (jamesn@moz-34C566A2.oracle.com) (Quit: Leaving)
- # [18:04] <@davidb> surkov: did you see we have IAccessibleTableUsage?
- # [18:04] <@davidb> err
- # [18:04] <@davidb> IACCESSIBLE_TABLE_USAGE
- # [18:05] <@surkov> davidb: who gave this name? it should be started from a11y prefix
- # [18:06] <@davidb> possibly tbsaunde
- # [18:06] <@tbsaunde> surkov: I think I remember me with r=you ;)
- # [18:06] <@tbsaunde> git log actually knows though ;)
- # [18:06] <@davidb> I'd be okay with changing that
- # [18:06] <@surkov> tbsaunde: damn, you should have been said me that I shouldn't do that :)
- # [18:06] <@davidb> heheh
- # [18:06] <@surkov> tbsaunde: please change
- # [18:07] <@tbsaunde> surkov: ok, I can file a bug soon or you can :)
- # [18:07] <@davidb> I'm wondering if we want these to be int
- # [18:07] <@davidb> I think the current data is not ideal
- # [18:07] <@tbsaunde> davidb: I'm not sure I follow
- # [18:07] <@davidb> let me start over
- # [18:08] <@davidb> for each firefox session, or each telemetry ping (which is 1-4 times a day - i forget)...
- # [18:08] <@davidb> I want one data point for each probe.
- # [18:09] <@surkov> tbsaunde: please file :) otherwise I'm starting to afraid that one day all you will set up spam filter against me :)
- # [18:09] <@tbsaunde> surkov: ;p
- # [18:09] <@davidb> So for … table interface usage… I want to know, for the case where our engine is invoked, was it used or not. so i want to see a bar for non-usage as well.
- # [18:09] <@tbsaunde> davidb: I gues that makes sense probably shouldn't be hard, didn't you already do it for one thing?
- # [18:09] <@davidb> yes
- # [18:09] <@davidb> you r+'ed :)
- # [18:10] <@surkov> btw, can we fix 0, 1, 2 on NVDA, JAWS and etc somehow
- # [18:10] <@davidb> knowing a function was called 100,000 times is less useful to me until we can get better database access
- # [18:10] <@davidb> surkov: i sort of like this minor level of privacy
- # [18:10] <@tbsaunde> davidb: wmakes sense
- # [18:10] <@davidb> surkov: but i could right you a bookmarklet
- # [18:11] <@davidb> write
- # [18:11] <@tbsaunde> davidb: can't you dirive no from the accessibility instantiation number though if we only call used() once per session?
- # [18:11] <@davidb> not easily
- # [18:11] <@surkov> :)
- # [18:11] <@davidb> at least i haven't been told clearly how to do that
- # [18:11] <@davidb> but!
- # [18:12] <@davidb> we can ask taras (tglek) questions and he can give us answers
- # [18:12] <@tbsaunde> davidb: privacy in that someone just has to match 1 with our code
- # [18:12] <@davidb> tbsaunde: right :)
- # [18:12] <@surkov> davidb, maybe file a bug, it's not comfortable to read source code to read stats
- # [18:12] <@davidb> really?
- # [18:12] <@surkov> I am :)
- # [18:12] <@davidb> i've already memorized the ones with hits :)
- # [18:12] <@tbsaunde> surkov: I'm starting to think I think the hardware accel stuff should live in msaa/
- # [18:13] <@davidb> tbsaunde: i think it should live outside /accessible
- # [18:13] <@tbsaunde> davidb: depends what problem your trying to solve I think
- # [18:13] <@davidb> I want to avoid non accessibility consumers.
- # [18:13] <@tbsaunde> if you want to solve arbitrary code in our process wants to know then I agree
- # [18:13] <@surkov> tbsaunde: if they provide our cross platform interface then we can use it in cross platform code :) I don't know. We could move it to MSAA code, but no testing and no XPCOM
- # [18:14] <@tbsaunde> surkov: xpcom sort of makes sense
- # [18:14] <@davidb> tbsaunde: I also want our engine to have no noticeable perf impact, but that has proven harder.
- # [18:15] <@tbsaunde> surkov: but I think the little bit of testing it makes sense for us to do would be a lot better as platform layer testing
- # [18:15] <@surkov> if we would have it
- # [18:15] <@tbsaunde> surkov: its not really cross platform though
- # [18:15] <@davidb> surkov: we should ask jost if they have automated tests with FF
- # [18:15] <@tbsaunde> surkov: true
- # [18:15] <@surkov> this interface is not implemented on other platforms so we shouldn't care about that really?
- # [18:16] <@tbsaunde> davidb: I don't really see why this particular thing would be a particular big reason for someone to start using us
- # [18:16] <@surkov> davidb: that might be interesting, but doesn't mean we shouldn't test
- # [18:16] <@davidb> surkov: agreed
- # [18:16] <@davidb> tbsaunde: it is hard to guess.
- # [18:16] <@tbsaunde> surkov: there really isn't much for us to test
- # [18:16] * @tbsaunde has to run cya
- # [18:17] * tbsaunde is now known as tbsaunde|afk
- # [18:17] <@davidb> tbsaunde: see bug 490794 for pain
- # [18:17] <@firebot> Bug https://bugzilla.mozilla.org/show_bug.cgi?id=490794 nor, --, ---, nobody, NEW, Kaspersky antivirus hooks into our process and enables the accessibility module
- # [18:18] <@davidb> We've had d2d security bugs, where users screens have been captured.
- # [18:19] <@davidb> it is really hard to predict how people will use the info a11y provides.
- # [18:19] <@davidb> and this feels more generic than a11y-only-tool to me.
- # [18:26] * khuey is now known as khuey|away
- # [18:27] <@surkov> davidb: should we cc somebody from privacy team? it doesn't sound like security problem
- # [18:27] <@davidb> surkov: yes exactly right, privacy.
- # [18:28] <@surkov> davidb, will you cc someone?
- # [18:28] <@davidb> ok
- # [18:28] <@surkov> thank you
- # [18:29] <@surkov> if there are privacy concerns then we need to end up with solution asking the user if he wants to provide this info to improve his expierence
- # [18:32] <@davidb> could be
- # [18:32] * @davidb looks for emails
- # [18:34] <@surkov> davidb: but the problem is we can't tell to the user who is asking for this information
- # [18:34] <@davidb> surkov: Zoomtext can :)
- # [18:35] <@surkov> davidb: yeah, but it can't it enable or disable
- # [18:35] <@surkov> it can say I'm going to ask FIrefox to enable this feature, Firefox is going to ask you about that please enable it
- # [18:35] <@surkov> that's sort of funny
- # [18:36] <@davidb> yeah, all it could do is say "you are using Firefox and we don't know if you are using D2D can tou..."
- # [18:36] <@davidb> tou/you
- # [18:36] <@davidb> yeah
- # [18:36] <@davidb> well, let's not get ahead of ourselvess
- # [18:36] <@surkov> :) just ask somebody for feedback
- # [18:36] <@davidb> I cc'ed geekboy
- # [18:37] <@MarcoZ> OK, I'm headed off for the weekend. Anybody here on Monday?
- # [18:37] <@davidb> MarcoZ: have a great one.
- # [18:37] <@davidb> 10 days until CSUN
- # [18:37] <@hub> I'll be here
- # [18:37] <@hub> no holiday in BC
- # [18:38] * @davidb won't
- # [18:38] <@MarcoZ> Yup! And haven't heard anything from Steve and richardschwerdtfeger on the preso. Guess I'll just prepare my HTML5 and WAI-ARIA forms combo stuff. ;)
- # [18:38] <@davidb> MarcoZ: it is worth pinging them
- # [18:38] <@davidb> oh
- # [18:38] <@davidb> you just did :)
- # [18:38] * @davidb steps away
- # [18:39] <@MarcoZ> Yeah, but am off now. richardschwerdtfeger: If you have anything, please send mail, or look at http://www.marco-zehe.de/examples (the English ones). I'd be showing off those in a little hands-down demo.
- # [18:39] * Quits: @MarcoZ (marco.zehe@moz-DF186A16.dip.t-dialin.net) (Quit: Bye everyone! Weekend!)
- # [18:40] * Quits: jhk (jiggy@8E6C34C1.A3F9767A.1C37C358.IP) (Connection reset by peer)
- # [18:42] * Joins: jhk (jiggy@8E6C34C1.A3F9767A.1C37C358.IP)
- # [18:56] * Joins: victorporof (victorporo@717ECBD0.6465CCD9.79933D60.IP)
- # [18:57] <@firebot> jigneshhk1992@gmail.com requested feedback from trev.saunders@gma il.com for attachment 598267 on bug 726283.
- # [18:57] <@firebot> Bug https://bugzilla.mozilla.org/show_bug.cgi?id=726283 nor, --, ---, jigneshhk1992, ASSI, consider using NS_GetContentList() directly in nsHTMLRadioButtonAccessible::GetPositionAndSizeIntern
- # [19:02] * Joins: sgimeno (santiago.g@moz-3F77BCB2.red-79-152-74.dynamicip.rima-tde.net)
- # [19:02] * Quits: @surkov (surkov@1A667863.26ED71FF.34044A7F.IP) (Quit: surkov)
- # [19:02] <@davidb> hi eeejay
- # [19:03] <@eeejay> davidb, yo!
- # [19:03] <@eeejay> davidb, writing up an outline of mobile a11y for wider distribution
- # [19:03] <@davidb> that's great to hear
- # [19:03] * khuey|away is now known as khuey
- # [19:09] * Joins: surkov (surkov@1A667863.26ED71FF.34044A7F.IP)
- # [19:09] * ChanServ sets mode: +o surkov
- # [19:10] <@tbsaunde|afk> davidb: so, I would sort of agree, I think us reporting hardware accel state is a little bit of a layering violation, but that wasn't a battle I thought was worth fighting
- # [19:11] <@tbsaunde|afk> davidb: I don't see how it can be a privacy issue, since whoever can call IAccessible2::get_Attributes() could also inject code into our process and grovel through memory to find out themselves
- # [19:11] <@tbsaunde|afk> or link against libxul and use the same api we do
- # [19:11] <@davidb> It isn't really for me to decide.
- # [19:12] * tbsaunde|afk is now known as tbsaunde
- # [19:12] <@davidb> I'm not familiar with possible vs difficult arguments around privacy either, but I'm sure they happen.
- # [19:12] <@tbsaunde> davidb: true, my point is more I don't think there is anything special here we need to ask about for privacy
- # [19:12] <@davidb> I was on the fence until surkov inquired.
- # [19:13] <@davidb> I do hear you though.
- # [19:13] <@tbsaunde> well, if you link agaainst libxul it isn't exactly hard though you suffer the need to recompile stuff for binary addons
- # [19:14] <@tbsaunde> I'd also say other processes like zoom text being able to take screen shots isn't a security / privacy problem for the same reason
- # [19:15] <@davidb> I'm the wrong guy to defend privacy audits.
- # [19:21] * Joins: nhirata (nhirata.bu@moz-BBE3ABD.mv.mozilla.com)
- # [19:24] <@tbsaunde> davidb: I'm not saying privacy auditing stuff is always a bad idea just auditing stuff that's alreayd available
- # [19:24] <@davidb> oh
- # [19:24] <@davidb> yeah, that's interesting.
- # [19:28] <@tbsaunde> davidb: ?
- # [19:28] <@davidb> figuring out what info you can already easily get.
- # [19:29] <@tbsaunde> davidb: yeah, its really a question of how you get it more than making it available
- # [19:29] <@tbsaunde> if your willing to write an addon getting if hardware accel on is trivial
- # [19:29] <@davidb> true
- # [19:30] <@davidb> we don't want third parties installing add ons though
- # [19:31] <@tbsaunde> davidb: sure, but they can and there isn't a lot we can do to stop them
- # [19:32] <@davidb> I can feel myself getting into a conversation I don't want to have right now :)
- # [19:32] <@tbsaunde> heh
- # [19:33] <@davidb> tbsaunde: do you happen to know what the mutation observer stuff that landed in chrome recently was about?
- # [19:33] <@davidb> it has been a while since i followed the DomAttrModified problem
- # [19:36] <@tbsaunde> davidb: no, I don't know much about mutation events, ask sicking or smorg?
- # [19:36] <@davidb> yeah just finished chat with smaug
- # [19:37] <@davidb> bug 641821
- # [19:37] <@firebot> Bug https://bugzilla.mozilla.org/show_bug.cgi?id=641821 nor, --, ---, bugs, NEW, Implement mutation events replacement (sync approach) (using moz prefix)
- # [19:37] <@tbsaunde> davidb: interesting
- # [19:38] <@tbsaunde> davidb: smorg have anything else interesting?
- # [19:38] * @tbsaunde reads bug
- # [19:40] * @surkov thinks sometimes Russian localization enjoys - https://bugzilla.mozilla.org/show_bug.cgi?id=728303
- # [19:40] <@firebot> Bug 728303 nor, --, ---, nobody, NEW, please find a better translation for 'test pilot'
- # [19:42] * Quits: @surkov (surkov@1A667863.26ED71FF.34044A7F.IP) (Quit: surkov)
- # [19:44] <@davidb> LOL
- # [19:46] <@tbsaunde> yeah
- # [19:47] <@tbsaunde> I'm sort of tempted to try and understand why espeak isn't doing something better than reading the unicdoe for the russian there
- # [19:52] * Quits: sgimeno (santiago.g@moz-3F77BCB2.red-79-152-74.dynamicip.rima-tde.net) (Ping timeout)
- # [19:59] <@tbsaunde> davidb: hm, reading that bug makes me wonder if nsIMutationObserver would be useful to us
- # [20:00] <@davidb> we use it
- # [20:00] <@davidb> nsDocAccessible
- # [20:00] <@davidb> or am I misunderstanding
- # [20:00] <@davidb> your point
- # [20:01] <@tbsaunde> davidb: oic, I hadn't caught that nsIDocumentObserver inherits from it
- # [20:01] <@davidb> yeah
- # [20:01] <@davidb> but clearly your intuition it would be useful is correct :)
- # [20:05] <@tbsaunde> davidb: yeah, particularly it'd be really nice if we could use it for text updates :/
- # [20:05] <@davidb> boy howdy
- # [20:06] <@tbsaunde> davidb: ?
- # [20:06] * Quits: @hub (hub@moz-E2FCA694.figuiere.net) (Ping timeout)
- # [20:08] <@davidb> i agree
- # [20:08] <@davidb> i think
- # [20:21] * Quits: drexler (chatzilla@moz-BD26ECCD.hsd1.vt.comcast.net) (Client exited)
- # [20:32] * Joins: hub (hub@21B7B9F2.B87E9213.6E712CE2.IP)
- # [20:32] * ChanServ sets mode: +o hub
- # [20:43] * Quits: jhk (jiggy@8E6C34C1.A3F9767A.1C37C358.IP) (Connection reset by peer)
- # [20:46] * Joins: jhk (jiggy@8E6C34C1.A3F9767A.1C37C358.IP)
- # [21:08] * Joins: peteb-away (ptbrunet@moz-E9B02845.austin.res.rr.com)
- # [21:13] * @eeejay is writing an analysis of android a11y
- # [21:13] <@eeejay> it is making me twitch
- # [21:14] <@davidb> you might need to hose down after
- # [21:15] <@eeejay> yep
- # [21:16] <@hub> use high pressure water jets
- # [21:20] <@davidb> eeejay: I feel like you are a navy seal with a butterfly net.
- # [21:20] <@eeejay> the mental image is amusing
- # [21:20] <@eeejay> maybe with a snorkel too
- # [21:21] <@davidb> or an entomologist with an RPG
- # [21:21] <@eeejay> heh
- # [21:21] <@davidb> and instead of a round, you have a suckling pig in the RPG tube
- # [21:21] <@davidb> and it is on acid
- # [21:22] <@davidb> there. I've written it for you.
- # [21:22] <@eeejay> davidb, thanks, i'll forward that with my doc
- # [21:23] <@davidb> heh
- # [21:23] <@davidb> eeejay: who is this for?
- # [21:23] * khuey is now known as khuey|away
- # [21:23] <@eeejay> davidb, want to send it to dev-planning and dev-moble
- # [21:23] <@eeejay> mobile
- # [21:23] <@eeejay> and the a11y list
- # [21:23] <@davidb> alrighty
- # [21:24] <@eeejay> may dev platform? dunno the diff
- # [21:24] <@davidb> planning is more for coordination
- # [21:24] <@eeejay> gotcha
- # [21:24] <@davidb> platform is for dev discussion
- # [21:33] <@hub> I have an Android tablet
- # [21:33] <@hub> we had spares here
- # [21:33] <@eeejay> hub, nice.. if there is another one, send it my way
- # [21:34] <@davidb> yes.
- # [21:34] <@davidb> eeejay: failing that we'll get you one at work week.
- # [21:34] * @eeejay high fives no one in particular
- # [21:36] <@hub> eeejay: there is some samsung junk
- # [21:36] <@eeejay> junk?
- # [21:36] <@hub> but then the tranformer show it is designed by idiots
- # [21:36] <@hub> "sorry can use your wifi, use an insecure one"
- # [21:36] <@eeejay> bloggers seem to be in love with it
- # [21:37] <@hub> eeejay: those that bought it or those that got it for free from Samsung?
- # [21:38] <@eeejay> hub, probably those that salivate over the pictures
- # [21:38] <@hub> eeejay maybe davidb get get you one
- # [21:38] <@davidb> yep
- # [21:39] <@eeejay> i don't mind either way. didn't you guys get some kind of tab 10?
- # [21:39] <@hub> I got one because karen has spares
- # [21:39] <@davidb> i have the samsung
- # [21:40] <@davidb> got it in san jose (all hands)
- # [21:43] <@davidb> eeejay: do you want that nexus one mailed or wait and pickup at work week?
- # [21:44] <@eeejay> davidb, maybe mailed? might be useful in csun
- # [21:45] <@davidb> ok!
- # [21:45] <@davidb> eeejay: can you give jlin your address in that bug?
- # [21:45] <@davidb> eeejay: what is your high end android?
- # [21:45] <@eeejay> davidb, yep
- # [21:45] <@eeejay> nexus s
- # [21:47] * Quits: jhk (jiggy@8E6C34C1.A3F9767A.1C37C358.IP) (Ping timeout)
- # [21:47] <@davidb> eeejay: seems to me you should also have ICS on something
- # [21:47] <@eeejay> davidb, it could run ics
- # [21:47] <@davidb> really?
- # [21:47] <@eeejay> davidb, not upgrading yet before i get something lower end
- # [21:47] <@davidb> i heard the nexus s2 can't so i assumed...
- # [21:47] <@davidb> ok
- # [21:47] <@eeejay> nah, it could
- # [21:47] <@davidb> i recall that converseation now
- # [21:47] <@eeejay> i just upgraded my personal phone to ics
- # [21:47] <@eeejay> samsung galaxy s 2
- # [21:48] <@davidb> ok
- # [21:48] <@eeejay> yesterday i handed the nexus s to a blind user. it was really bad
- # [21:48] <@davidb> tbsaunde: what do you have?
- # [21:48] <@davidb> :(
- # [21:48] <@eeejay> there are no error margins
- # [21:48] <@davidb> eeejay: is there an android device you could have handed over that would be better?
- # [21:49] <@eeejay> the 4 buttons on the bottom are touch and not tactile
- # [21:49] <@davidb> oh
- # [21:49] <@eeejay> so they keep hitting back and home by mistake while looking for the screen
- # [21:50] <@eeejay> tbsaunde, has a samsung captivate, afair
- # [21:51] <@davidb> that's a galaxy s i tink
- # [21:51] <@eeejay> cool
- # [21:51] <@davidb> where is dpad?
- # [21:51] <@davidb> or ball
- # [21:51] <@eeejay> on the captivate? none
- # [21:51] <@davidb> huh
- # [21:51] <@eeejay> they stopped making dpad/ball devices
- # [21:52] <@davidb> must be fun to use - not
- # [21:52] <@eeejay> it seems
- # [21:52] <@eeejay> the have arrows on the kb
- # [21:52] <@davidb> oh
- # [21:52] <@davidb> a slide out?
- # [21:52] <@eeejay> yeah
- # [21:52] <@davidb> ah
- # [21:52] <@davidb> ok
- # [21:52] <@davidb> i must be looking at the wrong thing
- # [21:53] <@eeejay> url?
- # [21:53] <@davidb> http://www.samsung.com/us/mobile/cell-phones/SGH-I897ZKAATT-gallery
- # [21:53] <@eeejay> yeah, i guess it is not the captivate
- # [21:53] * @eeejay looks at his notes
- # [21:54] <@eeejay> Stratosphere
- # [21:54] <@eeejay> samsung stratosphere
- # [21:54] <@davidb> ah ok
- # [21:54] <@davidb> good good
- # [21:55] <@eeejay> http://www.samsung.com/us/mobile/cell-phones/SCH-I405LKAVZW
- # [21:57] <@hub> https://bugzilla.mozilla.org/show_bug.cgi?id=726097#c17 <- can't we just get that fix in and have a new bug for this enhanced tests?
- # [21:57] <@firebot> Bug 726097 blo, --, ---, hub, NEW, Hit testing broken
- # [22:00] <@hub> sometime I wonder if we can't avoid ADD. I mean I wrote a test for the breakage and fixed the breakage.
- # [22:02] <@davidb> hub: you might want to email surkov directly.
- # [22:02] <@davidb> or catch him on IRC backchannel
- # [22:03] <@davidb> explain you'd like to land the fix with the test coverage, and that you'd like extra coverage not to slow this landing down.
- # [22:08] <@tbsaunde> hub: ADD?
- # [22:09] * Quits: @jprmc (jprmc@F2D29657.F60B0462.67AC9B1.IP) (Ping timeout)
- # [22:10] <@tbsaunde> so my experience of improve tests bugs is that it takes forever for them to get fixed or worked on at all
- # [22:10] <@davidb> I'm of two minds about them.
- # [22:10] <@eeejay> that's why tests should come hand in hand with fixes
- # [22:11] <@davidb> I don't like delaying things that are clear improvements.
- # [22:11] <@davidb> eeejay: or even TDD
- # [22:11] <@davidb> but we are playing catch up
- # [22:11] <@eeejay> yep
- # [22:11] <@tbsaunde> catch up with what?
- # [22:11] <@davidb> tbsaunde: TDD == test driven development. You do your tests first.
- # [22:12] <@davidb> gecko was not built that way
- # [22:12] <@davidb> nor is it built that way now really
- # [22:12] <@davidb> I've seen some pretty obscene amounts of tests land with new content work.
- # [22:12] <@hub> tbsaunde: ADD = Attention Deficit Disorder, or when you divert yourself of the thing you haven't finished to do other things
- # [22:12] <@davidb> Like Mounir's new form controls.
- # [22:12] <@eeejay> tdd and add are related
- # [22:12] <@davidb> hahah
- # [22:13] <@hub> tbsaunde: the bug was about a regression. I fixed the regression and wrote a test to ensure it happened again. but now I get asked at the same time to refactor the tests
- # [22:13] <@tbsaunde> davidb: you mean he landed very few or a lot?
- # [22:13] <@davidb> a LOT
- # [22:13] <@davidb> and jst required them
- # [22:13] <@davidb> it seems to be a characteristic of module owners
- # [22:14] <@davidb> you feel responsible for super coverage
- # [22:14] <@eeejay> hub, yeah, that is the problem with an old and imperfect code base. instead of you break it you buy it, it is you fix it you buy it
- # [22:14] <@tbsaunde> davidb: yeah, so I general want lots of tests but don't really think they should come *before* a first implementation
- # [22:14] <@davidb> I'd like to get to a point where I could just write tests for 2 months.
- # [22:15] <@davidb> take the load off
- # [22:15] <@tbsaunde> hub: yeah, so I'd tend to let you land that, but I suspect if I let you doo that the tests would never get fixed which I'd like to avoid
- # [22:15] <@hub> eeejay: I fixed it, and wrote the test that make sure I don'tr break it again
- # [22:15] <@eeejay> hub, and now you own the entire test suite :)
- # [22:16] <@tbsaunde> davidb: yeah, I'd love to see that happen but definitely have negative interest in doing it myself
- # [22:16] <@hub> *sigh*
- # [22:16] <@tbsaunde> unless you offer me barrels of vodca ;)
- # [22:17] <@davidb> this is clearly extortion :)
- # [22:17] <@davidb> hub: i'd choose vodka
- # [22:17] <@davidb> or
- # [22:17] <@davidb> wait until tbsaunde asks you for review ;)
- # [22:17] <@tbsaunde> hub: yeah, but someone needs to do it, sorry :/
- # [22:17] <@davidb> r+ if you write tests for my bugs
- # [22:18] <@davidb> correction
- # [22:18] <@davidb> for anything related to my bugs
- # [22:18] <@tbsaunde> heh
- # [22:25] <@davidb> ok guys, have a nice weekend
- # [22:25] <@davidb> i'm off monday
- # [22:26] * Quits: @davidb (davidb@F2D29657.F60B0462.67AC9B1.IP) (Quit: davidb)
- # [22:26] * khuey|away is now known as khuey
- # [22:29] * Quits: peteb-away (ptbrunet@moz-E9B02845.austin.res.rr.com) (Ping timeout)
- # [22:34] * Quits: victorporof (victorporo@717ECBD0.6465CCD9.79933D60.IP) (Connection reset by peer)
- # [22:35] * Joins: victorporo (victorporo@717ECBD0.6465CCD9.79933D60.IP)
- # [22:40] * victorporo is now known as victorporof
- # [22:43] <@hub> tbsaunde: sure someone need to do it. but if I do that now, I'm not fixing the Mac build
- # [22:51] * Parts: clown (clown@67828CC7.C1A51174.9D42CF23.IP)
- # [22:57] <@tbsaunde> hub: true
- # [22:57] * Quits: geo (geo@moz-BBE3ABD.mv.mozilla.com) (Ping timeout)
- # [22:59] * Joins: geo (geo@moz-BBE3ABD.mv.mozilla.com)
- # [23:05] * Joins: jprmc (jprmc@20347F7B.5BCEC6DB.DA78B690.IP)
- # [23:05] * ChanServ sets mode: +o jprmc
- # [23:29] * Quits: nhirata (nhirata.bu@moz-BBE3ABD.mv.mozilla.com) (Quit: nhirata)
- # [23:30] * Joins: nhirata (nhirata.bu@moz-BBE3ABD.mv.mozilla.com)
- # [23:36] * Quits: @jprmc (jprmc@20347F7B.5BCEC6DB.DA78B690.IP) (Ping timeout)
- # [23:36] * Joins: jpr_ (jprmc@D52C684D.5BCEC6DB.DA78B690.IP)
- # [23:45] * Quits: jpr_ (jprmc@D52C684D.5BCEC6DB.DA78B690.IP) (Ping timeout)
- # Session Close: Sat Feb 18 00:00:00 2012
The end :)