| 00:32 | <Hixie> | hrm |
| 00:32 | <Hixie> | it's not clear how to proceed on the <device> thing until we have a codec |
| 01:38 | <AryehGregor> | Oh, I wanted to link roc to <https://bugzilla.mozilla.org/show_bug.cgi?id=535043>;. Oh well. |
| 08:57 | <jgraham> | fantasai: Interesting. Al the documentation seems to have vanished |
| 08:59 | <jgraham> | Still in svn though: http://html5lib.googlecode.com/svn/wiki/UserDocumentation.wiki |
| 09:00 | <jgraham> | I wonder if we are supposed to migrate it to mercurial somehow |
| 09:29 | <hsivonen> | gsnedders: do we really need contentModelFlags.test to support initializing the tokenizer to an arbitrary? |
| 09:29 | <hsivonen> | gsnedders: also, what's spec for initialStates? |
| 09:32 | <hsivonen> | whew. the failures went away when I fixed the test harness to use the new test case syntax |
| 09:32 | <hsivonen> | except a couple of real failures I had introduces in the tokenizer |
| 09:35 | <hsivonen> | hg outgoing says |
| 09:35 | <hsivonen> | (deprecated '%' in path default-push=https://hsivonen%%40iki.fi⊙hgc/hg/ from /Users/Shared/Projects/html5lib/.hg/hgrc:3) |
| 09:35 | <hsivonen> | should I take one of the % signs away? |
| 09:35 | <hsivonen> | apparently, yes |
| 10:10 | <mia^^> | hi |
| 10:10 | <jgraham> | The W3C should really do whatever is needed to update WebIDL under /TR/ |
| 10:11 | <mia^^> | I am here to learn |
| 10:12 | <annevk2> | it's not really up to the W3C |
| 10:12 | <annevk2> | more up to weinig |
| 10:13 | <Philip`> | I thought the whole purpose of /TR/ was for hosting outdated documents |
| 10:13 | <jgraham> | Doesn't there have to be some sort of group decision or so? |
| 10:13 | <hsivonen> | mia^^: welcome |
| 10:14 | <jgraham> | Anyway having a draft that is greater then 1 year out of date under TR seems unusually poor (and would certainly faila per-document rather than per-group heartbeat requirement) |
| 10:14 | hsivonen | wonders what crazy UA sniffing WebKit and Opera are going to face when they try to implement <script defer> |
| 10:14 | <mia^^> | hsivone thank you |
| 10:16 | <mia^^> | bbs phone |
| 10:16 | <jgraham> | mia^^: Anything in particular you are here to learn, or just trying to soak up the ambiance? |
| 10:16 | <annevk2> | jgraham, the reason is that weinig wanted to update some more things before pushing out a new version |
| 10:16 | <annevk2> | jgraham, based on TC39 comments or something |
| 10:16 | <hsivonen> | does ES5 standardize the IE @cc_on stuff/ |
| 10:16 | <hsivonen> | ? |
| 10:17 | hsivonen | wonders why the document.write in http://www.mnot.net/templates/hinclude.js runs in Gecko |
| 10:17 | <jgraham> | annevk2: I would prefer "push early puish often" as a policy. The current situation confuses the hell out of e.g. Google |
| 10:17 | <jgraham> | hsivonen: @cc_on? |
| 10:18 | <hsivonen> | jgraham: JScript conditional comments |
| 10:18 | <jgraham> | No |
| 10:18 | <hsivonen> | hmm. |
| 10:19 | <annevk2> | jgraham, sure, I'm just explaining the situation |
| 10:21 | <annevk2> | Hixie, what if you invoke StreamRecorder.stop() several times? seems you get ever-larger file objects with the current text |
| 10:21 | <annevk2> | Hixie, you might want file objects per fragment |
| 10:24 | <annevk2> | http://dev.w3.org/2009/dap/camera/ looks quite ugly |
| 11:24 | <gsnedders> | Philip`: yt? |
| 11:37 | <Philip`> | gsnedders: y |
| 11:42 | <hsivonen> | oops. sorry. It seems I broke one test I thought I was fixing. I'll push a fix shortly. |
| 12:11 | <hsivonen> | did we agree that we want foster-parented text nodes to coalesce? |
| 12:16 | <jgraham> | hsivonen: I think we did but I'm not sure Hixie did |
| 12:17 | <hsivonen> | hmmkay. should I change the "excite me" test case now or later? |
| 12:17 | gsnedders | didn't think anyone apart from him and hsivonen was in that thread |
| 12:17 | <gsnedders> | _We_ did. Dunno about anyone else. |
| 12:17 | <jgraham> | Well I did too |
| 12:25 | <hsivonen> | I'll go ahead and change the test |
| 12:29 | <hsivonen> | is it intentional that we don't have a tree builder test file numbered 13? |
| 12:31 | <hsivonen> | lovely. I see test failures in script escapes and in the AAA |
| 12:38 | <gsnedders> | It used to exist, IIRC |
| 12:40 | <Philip`> | hg log says it never did |
| 12:40 | <hsivonen> | gsnedders: if you wrote it, did you forget to hg add? |
| 12:41 | <gsnedders> | No |
| 12:41 | <gsnedders> | I didn't write it |
| 12:41 | <gsnedders> | I thought it had existed once |
| 12:41 | <hsivonen> | whoever wrote 14 must have assumed 13 existed |
| 12:41 | hsivonen | goes see where 14 came from |
| 12:42 | <hsivonen> | jgraham: blame points to you |
| 12:43 | <jgraham> | I thought we once did and then it got moved to tests99 or so |
| 12:43 | <jgraham> | But it could just have been a mistake :) |
| 12:43 | <hsivonen> | 99? |
| 12:44 | <jgraham> | There were some tests that you made for not-yet-specified behaviour that got moved to a magic test file |
| 12:44 | <annevk> | airplanes skip row 13 too |
| 12:44 | <annevk> | it confused me recently |
| 12:44 | <jgraham> | By "magic" I mean "numbered 99" |
| 12:45 | hsivonen | doesn't see a file numbered 99 in the repo |
| 12:46 | hsivonen | wonders if the list of bad numbers in the airplane row range has much more than 13 on it or if only skipping 13 is a culturally-biased thing and proper i18n would require skipping many more rows |
| 12:46 | <gsnedders> | Then that was spec'd and then it was moved |
| 12:57 | <hsivonen> | my script escape states are shamefully buggy |
| 12:57 | <hsivonen> | it's a small wonder they work for dogfooding in practice |
| 13:01 | <jgraham> | I think if you skipped all the numbers that some culture considers unlucky you would have to go to a totally different system like named rows or something |
| 13:34 | <hsivonen> | ok if I change tests to anticipate Hixie fixing http://www.w3.org/Bugs/Public/show_bug.cgi?id=8373 ? |
| 13:36 | <annevk> | jgraham, not sure that'd go down better :) |
| 14:07 | <hsivonen> | now I have only two classes of test case failures left: |
| 14:07 | <hsivonen> | 1) surrogates (wontfix) |
| 14:07 | <hsivonen> | 2) space characters in the AAA |
| 14:07 | hsivonen | doesn't like touching the AAA |
| 14:07 | <hsivonen> | also, we don't have enough test coverage for frameset-ok |
| 14:07 | <annevk> | surrogates? |
| 14:08 | <hsivonen> | annevk: the test cases test lone surrogates in JSON input |
| 14:08 | <hsivonen> | annevk: I've decided checking for those doesn't belong in the tokenizer |
| 14:08 | <hsivonen> | but belongs on a lower layer that doesn't participate with JSON |
| 16:34 | <Hixie> | annevk3: i intend to make stop() make the object dead so a second call won't do anything |
| 16:34 | <Hixie> | or might throw an exception |
| 16:36 | <annevk3> | k |
| 16:57 | <annevk3> | whoa, what's up with all the new accounts on wiki.whatwg.org? |
| 17:02 | drry_ | changes topic to 'WHATWG: http://www.whatwg.org/ -- logs: http://krijnhoetmer.nl/irc-logs/ -- stats: http://gavinsharp.com/irc/whatwg.html -- Please leave your sense of logic at the door, thanks!' |
| 19:30 | <Hixie> | i don't understand why people have such an adverse reaction to the websocket protocol spec |
| 19:31 | <Hixie> | when probed, people always say it should be more like other specs, e.g. HTTP or the BWTP proposal, but that was exactly the kind of mistake I set out to avoid... |
| 19:31 | <Hixie> | (mistake because those specs describe the wire protocol, but generally don't define client or server requirements, and don't define error handling) |
| 19:51 | <franksalim> | Hixie, I think they are looking for a formal grammar describing WebSocket frames |
| 19:53 | <franksalim> | Hixie, which, unless I am missing something, is not in the latest draft |
| 19:53 | <Hixie> | what for? |
| 19:55 | <franksalim> | Hixie, to understand and discuss the frame structure. FWIW, I also think you are right to precisely describe client and server requirements and error handling |
| 19:55 | <franksalim> | without a grammar, you have to do a lot of reading to understand what the protocol would typically look like on the wire |
| 19:56 | <Hixie> | well, there's a diagram in section 1.2 showing the wire format |
| 19:56 | <Hixie> | but i guess i could add a non-normative abnf blob if that's what people want |
| 19:58 | <franksalim> | That is only my best guess of what people want |
| 20:05 | <franksalim> | judging from comments on hybi |
| 20:10 | <Hixie> | huh. i don't think you can describe the wire protocol as ABNF. |
| 20:11 | Hixie | uses the <> escape hatch |
| 20:18 | <annevk3> | oh man |
| 20:18 | <annevk3> | Forms WG comments are generating all the email |
| 20:21 | <annevk3> | I guess we should just go the W3C way about this; make a WG decision to not address their comment and move on... |
| 20:21 | <annevk3> | I mean sure, ideally everything would be separate and completely unreadable and even less understandable, but that's not the case |
| 20:37 | <Hixie> | franksalim: well i tried adding some ABNF for the wire protocol, we'll see how that goes |
| 20:54 | <annevk3> | Hixie, so to prevent the race condition I remove all currently queued tasks as part of the steps for abort()? |
| 21:09 | <annevk3> | MacBook fail |
| 21:09 | <annevk3> | I was planning on buying a new one but now this one is freezing/crashing on me all the time I'm not so sure anymore... |
| 22:58 | <AryehGregor> | Have implementers said they don't want to implement any kind of fullscreen() API for video? |
| 22:59 | <AryehGregor> | You'd think they could apply whatever heuristics they use for window.open(). |
| 22:59 | <AryehGregor> | For that matter, does Flash already allow scripts to fullscreen video? If so, it doesn't seem to be abused much. |
| 22:59 | <AryehGregor> | It at least allows them to make buttons that will trigger fullscreen. |
| 23:00 | <annevk3> | the request was not for fullscreen video |
| 23:00 | <annevk3> | at least not this time around |
| 23:00 | <AryehGregor> | Oh, dear, I was totally misreading it. |
| 23:00 | <AryehGregor> | Wait, he really does want fullscreen video. |
| 23:00 | <AryehGregor> | Just he thinks you should be able to arbitrarily fullscreen pages. |
| 23:00 | <AryehGregor> | That seems unnecessary. |
| 23:01 | <AryehGregor> | Michael Dale is the one who works on video for Wikimedia, so I figured he was asking something about video. |
| 23:01 | <annevk3> | anyway, bedtimes |
| 23:01 | <AryehGregor> | Night. |
| 23:27 | <Hixie> | annevk3: yeah i guess that'd be one way (so long as you only pull the tasks for this xhr) |