| 09:38 | <gsnedders> | Dashiva: String.fromCharCode returns UTF-16 code-units. JS strings have no knowledge of characters, merely UTF-16 code-units. |
| 09:41 | <annevk> | so how can http://tools.ietf.org/html/rfc3490#section-4.2 never fail but still fail in e.g. step 5? |
| 09:42 | <annevk> | is "If any step fails, then the original input sequence is returned immediately in that step." some kind of catch all not explicitly mentioned in the steps? |
| 09:42 | <annevk> | so that when step 5 fails you would just do that and not verify anything as step 7 describes? |
| 09:53 | <jgraham> | AryehGregor: Yeah, I think I once got it working on IE8, but I'm not sure what happened to the patch |
| 09:53 | <jgraham> | Now, of course, I don't have a computer with IE8 |
| 09:54 | jgraham | mutters something about vendors that only let you install one version of a browser at a time |
| 09:54 | <jgraham> | gsnedders: ping |
| 09:56 | <annevk> | nobody cares about DNS? |
| 09:56 | <annevk> | I am disappoint |
| 09:56 | <annevk> | also, even though I read quite a few emails while on "vacation" I'm now at 1700... up from 600 or so |
| 09:57 | <gsnedders> | jgraham: pong |
| 09:57 | <gsnedders> | jgraham: (I guess this is about my ping earlier) |
| 09:59 | <jgraham> | gsnedders: You told me to ping you. I have compiled. Now I will go back to ignoring you |
| 09:59 | <jgraham> | *complied |
| 10:00 | <gsnedders> | jgraham: You've compiled? Compiled what? |
| 10:00 | <gsnedders> | :P |
| 10:00 | <jgraham> | Myself. I am now very fast but have a tendency to crash |
| 10:00 | <gsnedders> | jgraham: html5lib seriailizer tests are dependent upon the treewalker format. I've changed that as we discussed before. |
| 10:01 | <gsnedders> | jgraham: Should I change the tests and break other impls, or? |
| 10:01 | <jgraham> | What other implementations? |
| 10:01 | <jgraham> | I didn't think there were any of the treewalkers |
| 10:01 | <gsnedders> | Well, if there are none, then I guess there's no issue |
| 10:02 | <gsnedders> | jgraham: I've implemented what we thought was the preferred solution, a list of dicts with "namespace", "name", and "value" for attributes on elements, instead of the current mess with no foreign content support. |
| 10:02 | <jgraham> | gsnedders: Sounds good |
| 10:03 | <jgraham> | So, did anyone keep up with hybi over Christmas? |
| 10:03 | <jgraham> | Has anything changed? |
| 10:04 | <gsnedders> | jgraham: The patch is a bit ugly at the moment, may well push and then clean up |
| 10:04 | hsivonen | is emerging from holiday and sickness |
| 10:04 | <gsnedders> | Like, just a bit more duplicated code than I'd like |
| 10:04 | <hsivonen> | lots of email as usual |
| 10:07 | <annevk> | we should form a club |
| 10:07 | <annevk> | on second thought, I guess we already have one |
| 10:09 | <jgraham> | The "I went on vacation and now I have lots of email" club? Seems like it might be distressingly non-exclusivr |
| 10:09 | <jgraham> | s/r/e/ |
| 10:11 | <gsnedders> | Bah, vacation ;P |
| 10:40 | <annevk> | hsivonen, asked our encoding guy |
| 10:40 | <annevk> | (re public-html) |
| 10:40 | <hsivonen> | annevk: thanks! |
| 10:41 | <jgraham> | annevk: Oh, I just asked someone else :) |
| 12:36 | <payman> | zcorpan: typeof(window.MediaError) should be 'function' or 'object'? |
| 12:37 | <bga_> | is it exception? |
| 12:37 | <bga_> | => function |
| 12:37 | <payman> | bga_: FF seems to disagree. |
| 12:38 | <payman> | (so does IE) |
| 12:39 | <zcorpan> | payman: dunno what webidl says |
| 12:40 | <zcorpan> | http://dev.w3.org/2006/webapi/WebIDL/#interface-object |
| 12:41 | <zcorpan> | "Unless otherwise specified, the [[Prototype]] internal property of objects defined in this section is the Object prototype object." |
| 12:41 | <payman> | 'function object', so object. |
| 12:41 | <zcorpan> | MediaError is not a constructor, so 'object' |
| 12:42 | <zcorpan> | while WebSocket would be 'function' |
| 12:43 | <payman> | zcorpan: I assume the same goes for window.TimeRanges ? |
| 12:43 | jgraham | wonders if http://dev.w3.org/2006/webapi/WebIDL/#es-constants changed recently |
| 12:44 | <zcorpan> | payman: yeah |
| 12:44 | <erlehmann> | that feel when "ps auxf" and in the START line there is only a year |
| 12:44 | <jgraham> | Oh, I am confusing that with something else |
| 13:09 | <foolip> | zcorpan, what [[Prototype]] is isn't relevant for what e.g. typeof window.MediaError will return, is it? |
| 13:09 | <foolip> | it seems that the relevant part of the spec is "There exists an interface object for every interface not declared with the [NoInterfaceObject] extended attribute. If the interface is declared with a [Constructor] extended attribute, then the interface object is a function object." |
| 13:10 | <foolip> | since MediaError isn't defined as [Constructor], it's not a function object, presumably |
| 13:10 | <zcorpan> | foolip: oh, i thought [[Prototype]] was relevant for typeof, but i'm not sure of that |
| 13:10 | <zcorpan> | foolip: indeed |
| 13:10 | <foolip> | payman, ^ |
| 13:19 | <zcorpan> | webkit doesn't support apng? |
| 13:20 | <hsivonen> | I can't remember ever seeing an apng announcement for any WebKit-based browser |
| 13:20 | <gsnedders> | zcorpan: Depends on the platform, no? Safari just throws it off to CoreImage, AFAIK |
| 13:25 | <annevk> | yeah, WebKit does not have a shared image layer |
| 13:25 | <annevk> | except for SVG, which isn't really images in the same sense |
| 13:27 | <zcorpan> | i see |
| 13:27 | <zcorpan> | would be nice with wider adoption of apng |
| 13:27 | <annevk> | better be in Acid 4 |
| 13:31 | <Rik`> | zcorpan: iirc apng has been rejected by the png spec |
| 13:33 | <Philip`> | The only effect that has is on the case of the chunk identifiers, I think |
| 13:33 | <Philip`> | It doesn't seem to have stopped people implementing it |
| 13:35 | <annevk> | it's like the debate on who owns text/html |
| 13:35 | <annevk> | *boring* |
| 15:22 | <hsivonen> | was a spec bug ever filed for making script.src setting start the fetch and having readyState and progress events to go with the fetch? see http://www.w3.org/mid/AANLkTikatiEzUaM6xMe465f5edCrFopYg2-5zhcN29DL⊙mgc ? |
| 15:23 | <hsivonen> | if a spec bug was filed, I failed to notice it when I read through bugmail |
| 15:25 | <miketaylr> | i can't find one since December 20, 2010 |
| 16:53 | <jgraham> | Hixie, hsivonen: It isn't clear to me whether one is supposed to preprocess bytes from document.write before they are passed to the tokenizer. So, for example, should unpaired surrogates in document.write end up as replacement characters, or as literal characters in the source? |
| 16:56 | <jgraham> | The spec describes it as "preprocessing the input stream" which might imply that it doesn't happen for document.write (since the characters are inserted directly into the input stream). But the test explicitly suggests that it happens before the characters are passed to the tokenizer, which would suggets it also applies to document.written characters |
| 16:59 | <jgraham> | The same question mostly applies to innerHTML |
| 17:01 | <jgraham> | i.e. in step 5 of http://www.whatwg.org/specs/web-apps/current-work/multipage/the-end.html#html-fragment-parsing-algorithm , should the input stream be preprocessed? |
| 18:32 | <hsivonen> | jgraham: document.write doesn't write bytes. It writes UTF-16 code units. Unpaired surrogates pass trough. The spec isn't clear on this. |
| 18:33 | <hsivonen> | jgraham: http://www.w3.org/Bugs/Public/show_bug.cgi?id=11298 |
| 18:35 | <hsivonen> | jgraham: I disagree with what Hixie says on the bug, but I'm not at a keyboard suitable for counterarguing |
| 18:46 | <gsnedders> | So much in html5lib depends upon the treewalker format, it's quite disturbing |
| 19:26 | <karlcow> | http://www.rield.com/cheat-sheets/rel-attribute-values.html |
| 19:48 | <gsnedders> | jgraham: See my huge commit I've just pushed to html5lib |
| 20:27 | <karlcow> | http://www.youtube.com/user/kkkwwwaaakkk#p/u/4/54XYqsf4JEY "most dangerous file format -> PDF" |
| 20:40 | <Workshiva> | "Now we must play the most dangerous game" |
| 20:54 | bga_ | hate websites from which /me can not download w/o js enabled |
| 21:09 | <Evet> | services > APIs > Applications > Websites |
| 21:59 | <Hixie> | jgraham: re surrogates and document.write(), i don't know. I wrote the spec assuming document.write() worked in Unicode characters, which may have been a mistaken assumption. |
| 22:09 | <bga_> | hm |
| 22:09 | <bga_> | is events subsystem has joistick/gamepad/... events? |
| 22:12 | <bga_> | i see |
| 22:12 | <bga_> | DOM_KEY_LOCATION_JOYSTICK |
| 22:16 | <bga_> | but |
| 22:16 | <bga_> | http://www.multi-device.ru/products_pictures/RaceFXWheelForXbox_150.jpg |
| 22:20 | <bga_> | wheel is analog device |
| 22:20 | <bga_> | and its not mouse wheel :) |
| 22:22 | <bga_> | + gamepad vibration api |
| 22:51 | <Evet> | which js framework you suggest for client-side of RIA? |
| 22:52 | <bga_> | xul/amplesdk |
| 23:07 | <webben> | When the spec lists hgroup among the elements with "special parsing rules", http://dev.w3.org/html5/spec/parsing.html#parsing , what rules is it referring to? |
| 23:07 | <webben> | oh never mind found them |
| 23:28 | <Evet> | bga_: wohoo its good |
