| 00:00 | <aho> | so, every heading of each box would look the same by default, but you can also change that if absolutely needed |
| 00:01 | <aho> | well, css and html was never meant to be used for this kind of stuff and it really shows, but that's how things are and we gotta work with it |
| 00:26 | <MikeSmith> | my Chrome is telling me it's up to date even though it's only at 9.0.572.1 and 9.0.576.0 is supposed to be available |
| 00:28 | <MikeSmith> | wonder if it's just me or if there's something awry |
| 00:29 | <MikeSmith> | Peter`: you were able to auto-update to 576? |
| 00:46 | <hober> | I'm at 9.0.576.0 dev |
| 00:47 | <hober> | at least on this machine (running windows) |
| 00:47 | <hober> | could be there's a delay for the other platforms |
| 00:48 | <aho> | also at 9.0.576.0 dev |
| 00:48 | <aho> | (winxp) |
| 00:51 | <nessy> | MikeSmith: I am on OS X and have the same version to you 9.0.572.1 |
| 00:51 | <MikeSmith> | nessy: OK, I'm on OS X too |
| 00:52 | <MikeSmith> | so maybe it's just a problem with the OS X build not being up to date or available yet |
| 00:52 | <nessy> | yup, would think so |
| 00:52 | <MikeSmith> | k |
| 00:52 | <aho> | there is some lag .) |
| 00:53 | <MikeSmith> | yeah, I now notice that the release announcement says, "The Dev channel has been updated to 9.0.576.0 for Windows and Linux." |
| 00:56 | <cardona507> | damn you are on chrome 9? |
| 00:56 | <cardona507> | nightly build? |
| 00:56 | <Hixie> | dev channel |
| 00:57 | <cardona507> | i c |
| 01:02 | cardona507 | is suddenly feeling ashamed of my decrepit stable channel version of Chrome |
| 01:14 | <TabAtkins_> | jgraham: What do you think of David Flanagan's latest point re: responseType? |
| 01:59 | <Hixie> | the bidi algorithm makes my head hurt |
| 01:59 | <Hixie> | like, literally. |
| 02:03 | <Hixie> | why does |
| 02:03 | <Hixie> | <span dir="ltr"><bdo dir="rtl">ב</bdo> <bdo dir="ltr">B</bdo></span> |
| 02:03 | <Hixie> | render differently (modulo the space) than |
| 02:03 | <Hixie> | <span dir="ltr"><bdo dir="rtl">ב</bdo><bdo dir="ltr">B</bdo></span> |
| 02:04 | <Hixie> | aaaaaaaaaaaaaaaaaaarrrrrrrrrrrrrrrrrrrrggggggggggggggggghhhhhhhhhhhhhhhhhhhh |
| 02:07 | <Hixie> | surely it should render aleph B regardless of whether there's a space between them or not??? |
| 02:08 | <Hixie> | why would it ever render B aleph? |
| 02:08 | <Hixie> | (alef, sorry) |
| 02:09 | <othermaciej> | my hypothesis: with no space, the whole run becomes RTL |
| 02:09 | <othermaciej> | (i.e. you can't switch directionality in the middle of a word) |
| 02:10 | <Hixie> | the sooner i'm done with these i18n bugs the sooner i'll be happy again |
| 02:10 | <othermaciej> | though that doesn't correctly predict what happens if you put a non-space character between the aleph and the B |
| 02:10 | <othermaciej> | hmm, depends on what character it seems |
| 02:11 | <othermaciej> | putting a 2 in between results in 2Bב |
| 02:11 | <othermaciej> | an x results in בxB |
| 02:11 | <Hixie> | can you get B something ALEF ? |
| 02:11 | <Hixie> | anyway, http://www.w3.org/Bugs/Public/show_bug.cgi?id=11211 is the only pre-LC1 bug i have left open |
| 02:12 | <Hixie> | and that one will get marked NEEDSINFO if I can't find an answer soon |
| 02:12 | <Hixie> | then i'm going back to whatwg mail and changing my work pattern to deal with bugs by having their bugmail be in my IMAP folders and dealing with each bug when i hit a bugmail from that bug |
| 02:13 | <Hixie> | rather than sometimes going through bugs and sometimes going through mail |
| 02:13 | <othermaciej> | it's not clear to me why you can't just use a unicode LS character for that use case |
| 02:13 | <MikeSmith> | Hixie: should I ask r12a to respond to that bug asap? |
| 02:13 | MikeSmith | reads the latest bug comments |
| 02:13 | <othermaciej> | unlike a newline, it doesn't need whitespace: pre to be preserved |
| 02:13 | <Hixie> | doesn't it? |
| 02:13 | <othermaciej> | so I don't think it needs a dedicated element |
| 02:13 | <Hixie> | it's not clear to me what CSS says to do with LS |
| 02:14 | <Hixie> | it's not a control character, technically, but it seems more control-like than glyph-like |
| 02:14 | <othermaciej> | does CSS say what to do with zero-width space, or non-breaking space? |
| 02:14 | <MikeSmith> | Hixie: you're waiting for a response to comment 8 about CSS? |
| 02:14 | <othermaciej> | I assume CSS only mangles the whitespace that it says it does |
| 02:15 | <Hixie> | oh actually CSS explicitly includes LS in its list of control characters |
| 02:15 | <Hixie> | MikeSmith: first paragraph of comment 8 or last paragraph of comment 11 |
| 02:15 | <MikeSmith> | k |
| 02:15 | <Hixie> | (same question phrased two different ways) |
| 02:16 | <Hixie> | CSS says: |
| 02:16 | <Hixie> | Control characters other than U+0009 (tab), U+000A (line feed), U+0020 |
| 02:16 | <Hixie> | (space), and U+202x (bidi formatting characters) are treated as |
| 02:16 | <Hixie> | characters to render in the same way as any normal character. |
| 02:16 | <Hixie> | LS is 202x (2028) |
| 02:16 | <Hixie> | but nowhere else does it say how to render it |
| 02:16 | <Hixie> | it in particular doesn't say it should create a new line box anywhere |
| 02:16 | <Hixie> | like it does for LF |
| 02:16 | <Hixie> | as far as i can tell |
| 02:17 | <Hixie> | anyway if we can just use LS that's fine by me |
| 02:17 | <Hixie> | "
" is shorter than "<br xxxx>", and I don't know what "xxxx" should be anyway |
| 02:18 | <othermaciej> | I see |
| 02:18 | <othermaciej> | WebKit doesn't seem to do anything for U+2028 in any whitespace mode |
| 02:18 | <Hixie> | as far as i know nobody does |
| 02:19 | <Hixie> | btw i love all my recent bidi examples, but i would really appreciate it if someone who speaks arabic or hebrew could check them |
| 02:19 | <Hixie> | my arabic is all from wikipedia or google translate |
| 02:19 | <Hixie> | so it's likely full of insults |
| 02:19 | <othermaciej> | I would guess AryehGregor can check your Hebrew |
| 02:20 | <Hixie> | the hebrew is from Aharon so I think it's ok |
| 02:20 | <Hixie> | but if someone could double check it too that would certainly be useful |
| 02:21 | <Hixie> | afk for a bit, probably bbiab |
| 02:21 | <MikeSmith> | Hixie: I've pinged r12a about it and ask him if he could help with facilitating a response from teh CSS WG |
| 02:23 | <othermaciej> | I think unicode does not consider U+2028 to be a bidi formatting character, even though it is 202x |
| 02:25 | <othermaciej> | 202A-E seem to be the bidi formatting characters |
| 02:27 | <othermaciej> | the rest of 202x are mostly punctuation characters |
| 02:27 | <jwalden> | just curious, what's the rationale for <input type=checkbox required> meaning the checkbox *must* be checked? seems odd for required to mean any validation at all for checkboxes |
| 02:28 | <othermaciej> | well, there are some forms (EULAs and the like) where you must check a certain checkbox to submit |
| 02:28 | <othermaciej> | so there is a use case |
| 02:29 | <othermaciej> | the other possible interpretation of "required" is that it has no effect on a checkbox |
| 02:29 | <othermaciej> | which seems less useful |
| 02:29 | <othermaciej> | others have suggested "at least one checkbox out of the group is checked" as a meaning, but checkboxes don't have groups |
| 02:34 | <MikeSmith> | http://unicode.org/Public/UNIDATA/PropList.txt |
| 02:34 | <MikeSmith> | 200E..200F ; Bidi_Control # Cf [2] LEFT-TO-RIGHT MARK..RIGHT-TO-LEFT MARK |
| 02:34 | <MikeSmith> | 202A..202E ; Bidi_Control # Cf [5] LEFT-TO-RIGHT EMBEDDING..RIGHT-TO-LEFT OVERRIDE |
| 02:37 | <jwalden> | ah, the EULA case -- that makes sense |
| 02:38 | <jwalden> | "at least one" sounds like a new widget cross-breed of radiobuttons and checkboxes |
| 02:38 | <jwalden> | <input type=radiocheckbutton> |
| 02:39 | <othermaciej> | there are in fact forms in existence that want you to check at least one of several checkboxes |
| 02:39 | <othermaciej> | or sometimes at least N for other values of N |
| 02:39 | <othermaciej> | or sometimes exactly N |
| 02:39 | <jwalden> | yeah, I can think of some without much trouble |
| 02:39 | <othermaciej> | (usually in cases like "check the following topics you are interested in") |
| 02:44 | <jwalden> | I'm a little surprised no one's actually gone to the effort to propose radiocheckbox as an input type, actually |
| 02:54 | <jwalden> | huh, does <input type=email> really not allow IDN domain names? if I'm reading it right, that's a bit unfortunate, if understandable |
| 02:55 | <jwalden> | pattern is |1*( atext / "." ) "@" ldh-str 1*( "." ldh-str )|, ldh-str as in http://tools.ietf.org/html/rfc1034 |
| 03:01 | <jwalden> | ah, spec permits display tricks |
| 03:06 | <jwalden> | I wonder if that's enough for something like gmail, tho, which may want to be able to display a name rather than an email address, and I don't see a way to tell the browser to do that |
| 03:10 | <MikeSmith> | if that's relevanthttp://www.w3.org/2010/02/rdfa/wiki/Subsets |
| 03:10 | <MikeSmith> | oofs |
| 03:10 | <MikeSmith> | http://www.w3.org/2010/02/rdfa/wiki/Subsets |
| 03:10 | <MikeSmith> | "Certain consumers of RDFa have indicated that they wish to offer only partial support. While the W3C has recommended full RDFa, the RDFa Working Group recognises demand for simpler, consumer-specific subsets" |
| 03:12 | <MikeSmith> | http://www.w3.org/2010/02/rdfa/wiki/Subsets#Ignoring_XML_Namespaces |
| 03:12 | <MikeSmith> | " A consumer may prefer not to support XML namespaces as a CURIE mapping method. " |
| 03:39 | <karlcow> | you had beautiful nightmares with blink in html, you will have enjoyable scary css3 animation http://www.fabulant.com/downloadcenter/textpropeller/textpropeller.html |
| 04:58 | <jwalden> | for anyone reading logs, sicking explained that the gmail-multiple-addresses use case needs a new input type so that names (in addition to addresses) are submitted, and that the topic's been discussed before |
| 05:23 | <MikeSmith> | https://lists.webkit.org/pipermail/webkit-dev/2010-November/015042.html |
| 05:23 | <MikeSmith> | "One more datum, that I just discovered, Apple has implemented XPath 2.0, XSLT 2.0, and XQuery" |
| 05:23 | <MikeSmith> | http://developer.apple.com/library/mac/documentation/Cocoa/Conceptual/NSXML_Concepts/NSXML.html |
| 05:31 | <MikeSmith> | is NSXML part of standard OSX, or is it a developer library that you need to statically link into apps? |
| 05:32 | <abarth> | there was some recent discussion of XSLT 2.0 on the webkit-dev list |
| 05:32 | <abarth> | the generally feeling seemed to be against implementing it |
| 05:32 | <abarth> | in WebKit |
| 05:32 | <abarth> | but i don't think it was ruled out |
| 05:41 | <MikeSmith> | abarth: yeah, that's where I came across mention of OS X NSXML already having XPath 2.0 and XSLT 2.0 support |
| 05:44 | <Hixie> | jwalden: i doubt gmail would use a browser-provided widget for that |
| 05:44 | <jwalden> | Hixie: that was also something sicking said :-) |
| 05:45 | <jwalden> | which seems slightly dumb to me, but whatever |
| 06:58 | <Peter`> | MikeSmith: Mac had the flash player updated last Monday, so it's a tad behind |
| 06:58 | <MikeSmith> | rok |
| 06:58 | <MikeSmith> | er |
| 06:58 | <MikeSmith> | ok |
| 08:29 | <zcorpan> | for constants on XHR: you probably have an XHR instant as a variable and the constants appear on that, too |
| 08:29 | <zcorpan> | so you'd write xhr.responseType == xhr.RESPONSE_ARRAYBUFFER |
| 08:29 | <zcorpan> | s/==/=/ |
| 08:30 | <zcorpan> | s/instant/instance/ |
| 08:43 | <annevk> | canvas API uses strings |
| 09:18 | <annevk> | hsivonen, you're fine with the IANA-based registry after all? |
| 09:18 | <annevk> | hsivonen, or did you make a mistake in your email? |
| 09:25 | <hsivonen> | oooops. |
| 09:26 | <hsivonen> | annevk: thanks |
| 09:47 | <zcorpan> | wow seems https://bugzilla.mozilla.org/show_bug.cgi?id=602028 has had a major bikeshed this morning |
| 09:49 | <annevk> | oh |
| 09:49 | <annevk> | tl;dr? |
| 09:49 | <zcorpan> | haven't read it yet |
| 09:50 | <Ms2ger> | No real new arguments |
| 09:51 | <jgraham> | Are there are no limits to the ways that WebSockets can make me sad? |
| 09:51 | <annevk> | it is waiting for you to disappear |
| 09:51 | <jgraham> | scare quotes implied? |
| 09:53 | <gsnedders> | jgraham: emo. |
| 09:55 | <jgraham> | Dude I am not an emo. I am just worried that Anne has the Mafia out to "disappear" me in the hope that it brings unity to the WebSockets worls and people ship the damn thing rather than endlessly talking about it |
| 09:58 | annevk | has the "connections" |
| 09:59 | <jgraham> | TabAtkins_: I think that using string constants violates TSBO(APOO)OWTDI |
| 10:00 | <annevk> | fortunately nobody knows what that is |
| 10:00 | <jgraham> | Because some people will write foo.responseType = "whatever" and others will write foo.responseType = XmlHttpRequest.WHATEVER_TYPE |
| 10:01 | <jgraham> | and confusion will reign |
| 10:01 | <annevk> | people are doing that with the <canvas> API today? |
| 10:01 | <annevk> | I somewhat doubt it |
| 10:01 | <Ms2ger> | s/Xml/XML/ |
| 10:03 | <jgraham> | What named string constants does it have? |
| 10:03 | <hsivonen> | IMO, the constants are entirely unhelpful, because they aren't backwards compatible |
| 10:04 | <hsivonen> | I was pretty annoyed when I tried to use the constants and they didn't work |
| 10:04 | <annevk> | you're talking about the readyState constants |
| 10:04 | <hsivonen> | it's not much of a victory if 8 years from now, we'll be able to write == 4 instead of == DONE |
| 10:04 | <hsivonen> | annevk: yes |
| 10:05 | <annevk> | we're talking about something else |
| 10:05 | <hsivonen> | I saw constants and XHR together and that was my cue to whine |
| 10:05 | <annevk> | :) |
| 10:05 | <jgraham> | hsivonen: You could define the constants yourself in the meantime :) |
| 10:05 | <annevk> | or you could just keep using the numbers |
| 10:05 | <hsivonen> | jgraham: oh like for namespace URIs? |
| 10:06 | <hsivonen> | annevk: I use the numbers now |
| 10:06 | <hsivonen> | annevk: but my time was wasted when I ended up finding out that I should use the numbers |
| 10:06 | <jgraham> | Well there are only a few and raw numbers are bad |
| 10:06 | <hsivonen> | raw numbers work! |
| 10:07 | <jgraham> | You can easilly do XMLHttpRequest.DONE = 4 at the top of your script |
| 10:07 | <jgraham> | and then use the constants |
| 10:07 | <jgraham> | (or however they work; maybe you need to put it on the prototype object) |
| 10:08 | <annevk> | I wouldn't exactly call it wasted time; this pretty much applies to each new feature we introduce |
| 10:08 | <annevk> | I agree some new features are more useful than others |
| 10:10 | <hsivonen> | in this case, the new feature is syntactic sugaring, so it's by definition not useful in the sense of having new functionality |
| 10:11 | <jgraham> | hsivonen: Did you agree with the resolution of 9659 in the end? |
| 10:11 | <jgraham> | http://www.w3.org/Bugs/Public/show_bug.cgi?id=9659 |
| 10:12 | <zcorpan> | jgraham: constants are both on the prototype and on the interface object, iirc |
| 10:13 | <zcorpan> | annevk: tl;dr is someone WONTFIXed the bug, someone else REOPENed it and the same arguments were repeated in further 36 messages and still no conclusion |
| 10:13 | <gsnedders> | That's not my memory… |
| 10:14 | <jgraham> | gsnedders: Finding things in your head that you're sure didn't happen to you? |
| 10:14 | <jgraham> | A rock and roll lifestyle will have that effect you know |
| 10:14 | <hsivonen> | jgraham: I haven't reviewed the resolution but I'm not planning on making substantial code changes |
| 10:15 | <jgraham> | hsivonen: Hmm. Ok |
| 10:15 | <jgraham> | I want html5lib to be close to the spec and I want to update the tests to match the new spec which is not quite trivial |
| 10:16 | <hsivonen> | jgraham: if the update isn't trivial, I think the bug should be reopened |
| 10:16 | <jgraham> | Pushing the null handling into each tokenizer state is lots of extra complexity |
| 10:17 | <hsivonen> | there's really no point in tweaking away from Gecko and WebKit interoperating |
| 10:17 | <hsivonen> | oh, you meant tokenizer tests |
| 10:17 | <hsivonen> | yeah, it's already in the tokenizer in Gecko, so that's good :-) |
| 10:17 | <hsivonen> | I thought you meant tree builder tests |
| 10:17 | <jgraham> | Well there is no difference between tokenizer tests and parser tests really :) |
| 10:18 | <jgraham> | s/parser/treebuilder/ |
| 10:18 | <jgraham> | Oh, so you always handle null explicitly? |
| 10:18 | <hsivonen> | I guess I really need to review the spec change as soon as I finish dealing with orange and have had lunch |
| 10:18 | <hsivonen> | jgraham: yes |
| 10:19 | <hsivonen> | jgraham: and I have a special token type for it |
| 10:20 | <jgraham> | Why? |
| 10:20 | <hsivonen> | jgraham: to avoid another pass over the data just to remap U+0000 and to be able to do context-sensitive handling |
| 10:21 | <hsivonen> | basically, I don't preprocess the input stream |
| 10:21 | <hsivonen> | that's baked into the tokenizer |
| 10:21 | <hsivonen> | always has been |
| 10:22 | <jgraham> | So you have special handling for all the characters that are converted to U+FFFD? |
| 10:22 | <jgraham> | in each tokenizer state? |
| 10:22 | <hsivonen> | jgraham: there's only U+0000 |
| 10:23 | <hsivonen> | jgraham: no other character turns into U+FFFD unless encoded as a numeric character reference |
| 10:23 | <hsivonen> | jgraham: but yeah, U+0000 is handled in every tokenizer state unless it gets reprocessed in another state as part of "anything else" |
| 10:24 | <jgraham> | hsivonen: You can't get unpaired surrogates? |
| 10:24 | <hsivonen> | jgraham: I think that belongs into the UTF-16 to UTF-16 converter |
| 10:25 | <hsivonen> | and the UTF-8 to UTF-16 converter |
| 10:25 | <jgraham> | So the input stream is actually preprocessed |
| 10:25 | <jgraham> | By the code that does bytes->unicode conversion |
| 10:25 | <hsivonen> | the byte stream is preprocessed by the bytes to UTF-16 code units converter |
| 10:25 | <jgraham> | Right |
| 10:26 | <hsivonen> | but the UTF-16 code unit stream isn't |
| 10:26 | <hsivonen> | e.g. document.write() isn't preprocessed for unpaired surrogates |
| 10:26 | <asmodai> | hsivonen: can there be CSS bugs in the Mac version of FF, but not in the Windows version? |
| 10:27 | <hsivonen> | asmodai: depends on what kind of CSS bugs |
| 10:27 | <jgraham> | Oh, that's an interesting point; should U+D800 get converted to U+FFFD in document.write |
| 10:27 | <hsivonen> | asmodai: pure style system bugs, no |
| 10:27 | <asmodai> | Colleague got a design that shows up perfectly on all kinds of browsers on Windows and Mac, but FF on Mac shows an issue |
| 10:27 | <hsivonen> | asmodai: graphics layer bugs, yes |
| 10:27 | <hsivonen> | asmodai: font metrics? |
| 10:27 | <asmodai> | wait, getting a screenie |
| 10:27 | <hsivonen> | jgraham: do browsers really do that? if not, we should make it a spec bug |
| 10:28 | <jgraham> | I'm not sure |
| 10:28 | <jgraham> | I'm not really sure what the spec requires either |
| 10:28 | <hsivonen> | jgraham: I'd be very unhappy about added complexity here if it isn't required for compat |
| 10:32 | <zcorpan> | seems opera gecko and webkit put U+D800 in the dom for document.write('\uD800') |
| 10:32 | <hsivonen> | problem solved |
| 10:33 | <jgraham> | hsivonen: So it seems likely that you will be happy with the resolution of that bug |
| 10:34 | jgraham | is surprised we don't already have tests for null in every input mode if that is how you implement it |
| 10:34 | <hsivonen> | jgraham: cool. I'll have a look after lunch |
| 10:39 | <asmodai> | hsivonen: http://www.in-nomine.org/~asmodai/firefox_windows.png and http://www.in-nomine.org/~asmodai/firefox_mac.png |
| 10:39 | <asmodai> | design is stable across Windows/Mac with Chrome, Safari, IE, only with the Mac FF does it do this |
| 10:41 | <Ms2ger> | HTML pls |
| 10:42 | <hsivonen> | asmodai: so the problem is the left gray sidebar? does its size depend on font metrics? |
| 10:42 | <asmodai> | hsivonen: Also on the right side, the grey bar is shorter suddenly |
| 10:42 | <asmodai> | moment, getting URL |
| 10:42 | hsivonen | goes get lunch |
| 10:43 | <asmodai> | hsivonen / Ms2ger : http://www.rsm.nl/home/execed/In-Company/How_We_Work |
| 10:46 | <Ms2ger> | No idea about your problem, but you have |color: #000000d;| in print.css |
| 10:48 | <asmodai> | Ms2ger: tossing that to my colleague. |
| 10:48 | <asmodai> | Thanks |
| 10:49 | <Ms2ger> | Thank whoever made Firefox dump that in the error console :) |
| 10:49 | <asmodai> | hehe |
| 10:49 | <asmodai> | He just fixed that part at least |
| 11:14 | <zcorpan> | abarth: under the attack scenario that motivates the masking, where a php script implements the handshake, can't the php script send the relevant information to the client to make it possible for the client to replay the masking and thus control the network bytes anyway (after the handshake has completed)? |
| 11:17 | <jgraham> | zcorpan: Does that not assume that the server being attacked is cooperating with the attacker in order to make itself vunerable? |
| 11:18 | <jgraham> | Although maybe that is possible in a shared-hosting scenario |
| 11:18 | <zcorpan> | jgraham: the scenario was attacking another virtual host on the same physical server |
| 11:23 | <jgraham> | Does it matter after the handshake has completed? |
| 11:27 | <Philip`> | Is test2.w3.org intentionally broken? |
| 11:35 | <hsivonen> | asmodai: I don't have an explanation to the Mac layout problem you are seeing. sorry |
| 11:38 | <annevk> | Philip`, prolly not? |
| 11:41 | <asmodai> | hsivonen: That's ok, bug for sure? |
| 11:43 | <hsivonen> | asmodai: could be a bug. could be a font metrics thing. but if it's a font metrics thing, it's at least mildly odd |
| 11:51 | <hsivonen> | jgraham: filed http://www.w3.org/Bugs/Public/show_bug.cgi?id=11298 |
| 11:58 | <hsivonen> | jgraham: the U+0000 fix looks reasonable to me |
| 11:59 | <asmodai> | hsivonen: logging a bug for core, which component should I pick do you think? |
| 11:59 | <hsivonen> | asmodai: layout |
| 11:59 | <asmodai> | ts |
| 11:59 | <asmodai> | ta |
| 12:00 | <hsivonen> | jgraham: ooh. now U+0000 gets ignored in "in select". well, ok whatever. |
| 12:01 | <hsivonen> | hmm. yeah, that's about right. nothing to see here |
| 12:09 | <jarib> | hi, the spec at http://www.whatwg.org/specs/web-apps/current-work/ has a syntax error in the IDL for HTMLVideoElement |
| 12:09 | <jarib> | [PutForwards=value attribute DOMSettableTokenList audio; |
| 12:09 | <jarib> | what's the right place to report this? |
| 12:10 | <Lachy> | jarib, use the bug reporting form at the bottom of the page |
| 12:10 | <jarib> | ok, thanks |
| 12:11 | <jarib> | you mean the "submit review comment" thing? |
| 12:11 | <asmodai> | hsivonen: Is this clear enough: https://bugzilla.mozilla.org/show_bug.cgi?id=611285 ? |
| 12:12 | <jarib> | ah, very nice |
| 12:13 | <hsivonen> | asmodai: I added a comment for clarity. thanks for filing |
| 12:14 | <asmodai> | hsivonen: no problem, thanks for the note. Hope they wont change the URL's code in the mean time :) (yay for screenies) |
| 12:14 | <Lachy> | asmodai, why do I get an Invalid Bug ID error when I look at that bug? |
| 12:15 | <Lachy> | oh, weird. It gives me an error when I'm not logged it, but works when I am. |
| 12:15 | <asmodai> | Lachy: Peculiar |
| 12:15 | <asmodai> | maybe because of the https? |
| 12:18 | <jgraham> | hsivonen: Arguably that bug is invalid, since it might be implied in the spec that preprocessing the input stream doesn't happen for things that you insert into it via document.write |
| 12:18 | <jgraham> | It is not 100% clear though |
| 12:23 | <zcorpan> | jarib: known issue |
| 12:23 | <jarib> | zcorpan: ah |
| 12:23 | <zcorpan> | iirc at least :) |
| 12:24 | <jarib> | got it |
| 12:24 | <smaug____> | annevk: ping |
| 12:24 | <smaug____> | annevk: could you paste the link to your XHR tests? |
| 12:25 | <annevk> | http://tc.labs.opera.com/apis/XMLHttpRequest/ .com/svn for subversion |
| 12:25 | <smaug____> | thanks |
| 12:53 | asmodai | pokes bugzilla.mozilla.org -- not your day is it? |
| 13:19 | <jgraham> | Is it intentional that CDATA sections no longer trandform U+0000->U+FFFD? |
| 13:19 | <jgraham> | hsivonen: ^ |
| 13:19 | <jgraham> | *transform |
| 13:20 | <hsivonen> | jgraham: nope, not at all intentional |
| 13:20 | <hsivonen> | I suck at spec review, it seems. |
| 13:21 | <hsivonen> | jgraham: that's what happens with spec writing that tries to avoid spelling out all the states |
| 13:32 | jcranmer | wonders why people find specs unreadable |
| 13:42 | <annevk> | jcranmer, because like reading code, it's difficult |
| 13:45 | <Philip`> | jcranmer: Because they're written with the assumption that the reader already knows how to read them and has a large amount of prior knowledge of the topic |
| 13:46 | <Philip`> | (They'd probably be ten times as long if they didn't make those assumptions) |
| 14:02 | <espadrine> | Is there any way at all to do syntax highlighted editing of text in html? |
| 14:03 | <espadrine> | Section "user-editing-actions" of the spec is very imprecise about javascript ways to move the caret |
| 14:03 | <jgraham> | espadrine: Build a text editor using <canvas>? |
| 14:04 | <espadrine> | jgraham: Bespin way... but that means the text is not searchable, right? |
| 14:05 | <Philip`> | It means you lose all kinds of features that OSes/browsers already provide and users often expect |
| 14:07 | jgraham | was being ironic |
| 14:07 | <espadrine> | Philip`: I think that is unfortunate. |
| 14:07 | <zcorpan> | use <textarea> and XBL |
| 14:07 | <jgraham> | The irony being that it is an obviously crazy idea, but also the way that seems to work |
| 14:07 | <jgraham> | or at least a way |
| 14:12 | <espadrine> | zcorpan: XBL seems nice, but it's a Mozilla-only api |
| 14:12 | <espadrine> | Am I wrong? |
| 14:12 | <jgraham> | It is standard |
| 14:12 | <jgraham> | An unimplemented |
| 14:12 | <jgraham> | *and |
| 14:12 | <jgraham> | (XBL2) |
| 14:13 | <jgraham> | Philip`: Hmm, this single-character-unget stuff is annoying |
| 14:14 | <jgraham> | By which I mean "makes things look confusingly unlike the spec" |
| 14:18 | <Philip`> | jgraham: How so? |
| 14:18 | <Philip`> | I thought ungetting a character then consuming it was just meant to be equivalent to the spec's notion of reconsuming |
| 14:20 | <jgraham> | Because when you lookahead you can't unconsume all the characters that you just looked at |
| 14:20 | <jgraham> | So you have to add extra modes and extra not-in-the-spec logic |
| 14:22 | <Philip`> | I thought that only happened with character references, and in that case the characters will all be (?i)[&0-9a-z] so they won't be handled as anything other than character tokens |
| 14:22 | <Philip`> | or something like that |
| 14:23 | <jgraham> | Doctypes |
| 14:24 | <Philip`> | Oh, that too |
| 14:24 | <Ms2ger> | Doctypes add a whole new sense of fun |
| 14:25 | <gsnedders> | Nah, they're not that fun until you end up on insane amounts of switching like IE |
| 14:26 | <Philip`> | The sad thing is that the entire meaningful output of the complex doctype syntax is 1.6 bits of data |
| 14:29 | <Philip`> | jgraham: I suppose the current implementation of unget means it possibly wouldn't be hard or slow to support ungetting multiple characters |
| 14:29 | <Philip`> | jgraham: I vaguely remember it was originally an issue due to keeping track of line number, perhaps |
| 14:32 | <karlcow> | hmm the number of times in a form when I press Command+S instead of clicking on Save Draft and having the stupid dialog window. :) |
| 14:32 | <Philip`> | jgraham: The code will never really look like the spec though, since lookahead is incompatible with streaming input |
| 14:38 | <hsivonen> | it's uncool that the spec has lookahead |
| 14:40 | <jgraham> | Philip`: True, but extra states seem quite bad |
| 14:41 | hsivonen | has extra states |
| 14:42 | <jgraham> | Philip`: I can't see why unget wouldn't worj with > 1 character at the moment |
| 14:49 | <jgraham> | This new null handling doesn't seem to maintain the relationship that 1 null character === 1 parse error |
| 15:18 | <jgraham> | Is it me or does the end tag open state eat the first character of the bogus comment in the "anything Else" case? |
| 15:18 | <jgraham> | http://www.whatwg.org/specs/web-apps/current-work/complete.html#end-tag-open-state |
| 15:18 | <jgraham> | That seems wrong |
| 15:23 | <zcorpan> | jgraham: "Emit a comment token whose data is the concatenation of all the characters starting from and including the character that caused the state machine to switch into the bogus comment state, ..." |
| 15:24 | <hober> | hsivonen: in your latest blog post, in the definition of External Script: s/and/an/ |
| 15:24 | <zcorpan> | jgraham: or do you want the slash to be part of the data? |
| 15:25 | <jgraham> | zcorpan: Oh, I missed that because it is a dumb place to put that requirement :) |
| 15:26 | <zcorpan> | yeah the bogus comment stuff is a bit different to the rest of the parser |
| 15:26 | <zcorpan> | tokenizer* |
| 15:26 | <hober> | hsivonen: in the definition of Executing a script, s/text a JavaScript/text as JavaScript/ |
| 15:27 | <jgraham> | The rest isn't so bad; it's just that particular bit that basically reads "collect some characters then emit a tokejn containing them and one more character that you probably discarded by now" |
| 15:30 | <hober> | jgraham: that bit was really easy in elisp |
| 15:31 | <hober> | I don't discard any characters, since I'm tokenizing a buffer in-place :) |
| 15:31 | <hober> | "consume the next input character" is just (forward-char 1) |
| 15:31 | <jgraham> | hober: That is cheating :) |
| 15:31 | <hober> | indeed |
| 15:32 | <jgraham> | hober: How is that going? |
| 15:32 | <jgraham> | I am totally psyced about a decent HTML mode for emacs |
| 15:32 | <jgraham> | (sp?) |
| 15:33 | <hober> | I haven't worked on it in a couple of weeks |
| 15:33 | <jgraham> | (although in my dream world, I would get to use js2-mode inside <script> elements and css mode inside <style> elements) |
| 15:33 | <hober> | insofar as the tokenizer is concerned, it's essentially done, but I can't prove that yet, until I improve my html5lib-test-runner |
| 15:33 | <jgraham> | (I think if I got that I would probably cry) |
| 15:33 | <hober> | yeah, that bit isn't going to happen |
| 15:33 | <jgraham> | :) |
| 15:34 | <hober> | clone-indirect-buffer, narrow-to-region, foo-mode |
| 15:34 | <Philip`> | jgraham: (And you'd get HTML mode inside document.write inside <script>) |
| 15:34 | <jgraham> | Philip`: Hah |
| 15:34 | <jgraham> | Not even I am that unreasonable |
| 15:35 | <hober> | I suppose I could work on the editing mode in parallel to improving my test passing score |
| 15:35 | <Philip`> | I wonder what it'd be like if web technology had a strict concept of layers |
| 15:36 | <hober> | hsivonen would like the mode to be able to run relax ng validation with nxml's rng-valid stuff |
| 15:36 | <Philip`> | instead of being lots of components which can all embed each other and be embedded in each other |
| 15:36 | <zcorpan> | Philip`: and css mode in style="" in document.write inside <script>? |
| 15:36 | <hober> | hopefully that isn't too tightly coupled to a) xml and b) nxml |
| 15:36 | <hober> | zcorpan: and js2-mode in an IE behavior expression in style="" in document.write inside <script> :) |
| 15:36 | <Philip`> | (HTML containing PHP using SQL to return HTML which contains scripts which print HTML which call scripts which ...) |
| 15:37 | <zcorpan> | hober: naw, IE behaviors are non-standard |
| 15:37 | <Philip`> | (It's not surprising there's so many security vulnerabilities due to incorrect escaping) |
| 16:25 | eric_carlson | works on a build fix |
| 16:26 | <Ms2ger> | Wrong window? ;) |
| 16:27 | <jgraham> | It's like twitter, he's just broadcasting some irrelevant details from his life to people he has never met |
| 16:30 | <eric_carlson> | wrong window, but I have met some of the people here :-) |
| 16:30 | <othermaciej> | MikeSmith: to answer your earlier question, NSXML is part of Foundation and a Mac OS X system API |
| 16:31 | <othermaciej> | MikeSmith: it doesn't generally get random Web content thrown at it in uncontrolled conditions, so the XSLT 2.0 compatibility breaks are probably less of an issue |
| 16:31 | <MikeSmith> | ok |
| 17:08 | <MikeSmith> | Hixie: please note http://lists.w3.org/Archives/Public/public-html/2010Nov/0192.html if you've not already |
| 17:08 | <MikeSmith> | about issue 128, "Authors should be able to use <figure> where <img> can be used"" |
| 17:09 | <MikeSmith> | "if either of the original proposal authors would like additional time to revise their proposals, please respond to this message by Wednesday, November 17" |
| 17:19 | <AryehGregor> | Hixie, I don't see where the Hebrew stuff is. I only see Arabic examples in <bdi> and dir=auto. |
| 17:24 | <Hixie> | AryehGregor: there's a hebrew example that contains the word "F#" |
| 17:24 | <Hixie> | MikeSmith: are there edits I should make that I haven't made? |
| 17:24 | <MikeSmith> | Hixie: up to you |
| 17:25 | <MikeSmith> | I just wanted to give you a heads-up that the deadline had been set |
| 17:25 | <MikeSmith> | in case you might want to change anything |
| 17:25 | <Hixie> | don't worry, i read all the mail from the chairs sent to the list :-) |
| 17:25 | <AryehGregor> | Hixie, does that imply that you expect me to open up the complete version of the spec so I can search for it? Because right now I'm using Firefox, and even in Firefox 4, it basically freezes all the other tabs for a minute or two whenever I do that. |
| 17:25 | <MikeSmith> | hai hai |
| 17:25 | AryehGregor | finds it anyway |
| 17:25 | <AryehGregor> | "בחר שפת תכנות" |
| 17:25 | <Hixie> | AryehGregor: it's in the rendering section, i think, but honestly i don't remember at this point |
| 17:25 | <Hixie> | does ff4 really still not handle the complete spec? |
| 17:26 | <Hixie> | i thought they fixed that |
| 17:26 | <Hixie> | it was working last i tried |
| 17:26 | <AryehGregor> | It does, if you wait a minute or two. |
| 17:26 | AryehGregor | wonders if IE9 renders it faster |
| 17:26 | <AryehGregor> | (not that it helps me on Linux) |
| 17:27 | <AryehGregor> | Awesome, RTL text jumps around in XChat if you even try to select it. |
| 17:27 | AryehGregor | sighs |
| 17:27 | <AryehGregor> | "בחר" certainly means "choose". |
| 17:27 | <AryehGregor> | Yay first-string! |
| 17:27 | <AryehGregor> | The word "בחר" certainly means "choose". |
| 17:27 | <AryehGregor> | Er, "Yay first-strong!" |
| 17:28 | <Ms2ger> | All you need is bdi on IRC |
| 17:28 | <AryehGregor> | In the masculine singular. Wikipedia seems to use the plural for things like this. No idea if "שפת תכנות" means "programming language", but it seems plausible. |
| 17:29 | <AryehGregor> | This is rather modern stuff for me to be sure. |
| 17:29 | <AryehGregor> | Give me a prayer or something, that's more up my alley. |
| 17:29 | <AryehGregor> | But I'm guessing Aharon is Israeli and knows what he's talking about. |
| 17:29 | <Hixie> | hehe |
| 17:40 | AryehGregor | notices that Linus Torvalds seems to use Chrome: https://bugzilla.redhat.com/show_bug.cgi?id=638477#c38 |
| 18:04 | <AryehGregor> | hsivonen, "The old Firefox behavior was to execute non-async, non-defer scripts in the order in which they were run." http://hsivonen.iki.fi/script-execution/ |
| 18:04 | <AryehGregor> | Surely you mean "the order in which they occurred in the document" or something? |
| 18:05 | <gsnedders> | execute them in the order they were run? Wow. |
| 18:05 | <Ms2ger> | run != execute |
| 18:06 | <AryehGregor> | Oh, wait. |
| 18:07 | <AryehGregor> | I didn't read the definitions. |
| 18:07 | <AryehGregor> | How confusing. |
| 18:07 | <AryehGregor> | "(Blame the HTML5 spec for giving distinct meanings to “run” and “execute”.)" |
| 18:07 | AryehGregor | throws tomatoes at Hixie |
| 18:07 | <AryehGregor> | That is exceptionally horrible terminology. |
| 18:07 | <Hixie> | feel free to file a bug suggesting better terminology |
| 18:08 | <Hixie> | this stuff just evolves and often ends up silly |
| 18:08 | <AryehGregor> | http://www.w3.org/Bugs/Public/show_bug.cgi?id=11301 |
| 18:08 | <Ms2ger> | Hixie, btw, are you planning to remove the stuff that's in DOM Core now? |
| 18:50 | <Hixie> | AryehGregor: awesome, thanks |
| 18:50 | <Hixie> | Ms2ger: haven't been following DOM Core, but if there's stuff that needs removing, file a bug or send mail |
| 18:50 | <Ms2ger> | Okay |
| 19:26 | <annevk> | Hixie, plenty of stuff can be removed |
| 19:32 | Ms2ger | wants RESOLVED JUNK |
| 19:34 | <annevk> | INVALID is not enough? |
| 19:35 | <mdelaney> | Philip`: ping |
| 19:35 | <annevk> | I guess sometimes you want to express "your bug report is so stupid it is not even worth classifying as INVALID"; especially for the spam |
| 19:35 | <Ms2ger> | Yeah |
| 19:38 | Ms2ger | sighs |
| 19:39 | <kennyluck> | SPAM |
| 19:39 | <Ms2ger> | We are now discussing the wording of the sentence "Entire novels have been written about the security considerations that apply to HTML documents." |
| 19:50 | <Philip`> | mdelaney: Pong |
| 19:51 | <mdelaney> | Philip`: just sent you an email |
| 19:51 | <Philip`> | mdelaney: I saw :-) |
| 19:51 | <mdelaney> | kk |
| 19:52 | <Philip`> | mdelaney: Will reply when I have a bit more time, if it's not urgent |
| 19:54 | <mdelaney> | No, it's not. I'm just on a canvas clean-up rampage and it's a quick thing. No rush. |
| 20:17 | <heycam> | i like the idea of novels having been written about security considerations of html |
| 20:17 | <heycam> | http://mcc.id.au/temp/novel.html |
| 20:25 | <Hixie> | rofl |
| 20:25 | <Hixie> | that's awesome |
| 20:26 | <TabAtkins_> | Hahahahahaha |
| 20:28 | <annevk> | heycam, want more |
| 20:28 | <heycam> | :) |
| 20:30 | <annevk> | also, that should not be in temp |
| 20:30 | <Philip`> | Hmph, I was just about to complain bitterly about your inconsistent use of "--" vs "—", but now it looks like you've fixed it already so I've got nothing to complain about :-( |
| 20:30 | annevk | makes a copy |
| 20:31 | <heycam> | more permanent - http://mcc.id.au/2010/novel.html |
| 20:32 | <heycam> | (in practice, my temp dir is reasonably permanent though) |
| 20:32 | <gsnedders> | heycam: +1 |
| 20:33 | <annevk> | I hope we see more of this now you have permanent free time ;p |
| 20:33 | <heycam> | good point |
| 20:33 | heycam | returns to work |
| 20:37 | <nimbupani> | heycam: is this something I can tweet :) |
| 20:37 | <annevk> | it is on the interwebs but you cannot link to it |
| 20:37 | <nimbupani> | its polite to ask annevk :P |
| 20:38 | <annevk> | you must be new here |
| 20:38 | <heycam> | nimbupani, if you want |
| 20:38 | <nimbupani> | ok TabAtkins_ oredi did :) |
| 20:39 | <Philip`> | nimbupani: Surely it's too long to tweet |
| 20:39 | <nimbupani> | so I am no longer responsible for its widespread popularity. |
| 20:40 | <Philip`> | though you could do it in serialised instalments |
| 20:40 | <annevk> | Philip`, are you even on twitter? |
| 20:40 | <Philip`> | No |
| 20:41 | <nimbupani> | serialised twitter novels are not that fun Philip` |
| 20:45 | <Peter-> | heycam: haha! :D |
| 20:46 | Ms2ger | expects to see the novel on public-html |
| 20:47 | <annevk> | any minute now |
| 20:47 | <annevk> | taking bets in #whatwg-bets |
| 20:53 | <hober> | Max Barry's _Machine Man_ is being published, one page at a time, on a blog: http://www.maxbarry.com/machineman/ |
| 20:54 | <hober> | you could do that with _Security Considerations_ :) |
| 21:07 | <Ms2ger> | Does the item attribute still exist? |
| 21:07 | <zcorpan> | hmm, the http-like header parsing didn't go away in the introduction of json in abarth-01 |
| 21:08 | <Ms2ger> | ...they can quite legitimately say that the aforementioned random junk is just that, junk... |
| 21:09 | <annevk> | Ms2ger, not mentioned as global attribute |
| 21:10 | <TabAtkins_> | Ms2ger: Renamed to @itemscope |
| 21:26 | <zcorpan> | although i guess it became a bit simpler for the server side, at least if there's an off-the-shelf json parser available |
| 21:28 | <Ms2ger> | TabAtkins_, thanks, filed a bug already |
| 21:32 | <Ms2ger> | Hixie, annevk, filed bug 11305 for DOM Core stuff |
| 21:44 | <zcorpan> | Hixie: the length of the attribute name is not so relevant imho, it should be easy to remember and easy to type |
| 21:44 | <zcorpan> | Hixie: <br foobar> wins on both compared to 
 for many values of foobar, i think |
| 21:45 | <aho> | you cant go wrong with asdf :> |
| 21:49 | <zcorpan> | Hixie: the behavior in browsers without the feature is also worth considering: is it more important to get a line break than to not mess up the direction, or the other way around |
| 21:49 | <zcorpan> | if the line break is more important, then <br foobar> is better; otherwise a new element is probably better |
| 21:58 | <Philip`> | zcorpan: Would a named character reference be harder to remember than <br foobar>? |
| 21:59 | <zcorpan> | Philip`: probably not, but seems worse than an element in browsers without the feature |
| 22:03 | <Hixie> | it's better to not break than to break with the direction wrong |
| 22:03 | <Hixie> | the direction being wrong can actually reverse the meaning in some edge cases |
| 22:04 | <zcorpan> | thus: invent a new element |
| 22:06 | <jcranmer> | I shall call it... ununbium |
| 22:07 | <zcorpan> | does <ubi><br></ubi> have the desired effect in browsers that support <ubi>? |
| 22:08 | <zcorpan> | or bdi or what it's called |
| 22:10 | zcorpan | wonders if it's possible to get the desired effect of bdi with :before and :after in legacy browsers |
| 22:15 | <zcorpan> | hsivonen: you should get a favicon :) |
| 22:26 | <zcorpan> | hsivonen: s/a JavaScript/as JavaScript/ |
| 22:28 | <Dashiva> | I didn't know people still capitalized the s |
| 22:29 | gsnedders | just calls it ECMAScript unless he's actually referring to the dialect supported by Mozilla, pretty much |
| 22:29 | <gsnedders> | Which, of course, confuses people. :P |
| 22:35 | <TabAtkins_> | I just call it js. That works well enough. |
| 22:35 | <Dashiva> | I write javascript, but the spellchecker keeps complaining |
| 22:38 | <Hixie> | zcorpan: a new element or a character reference, yeah |
| 22:39 | <Hixie> | &br; |
| 22:39 | <Hixie> | or &ls; |
| 22:39 | <Hixie> | which, as some people sometimes say, is probably how it should have been defined in the first place |
| 22:39 | <TabAtkins_> | What precisely is the problem caused by <br> in bidi situations? |
| 22:39 | <Hixie> | the text on the next line ends up backwards |
| 22:40 | <TabAtkins_> | Does <br> have some weird directionality or something? |
| 22:40 | <Hixie> | it is treated as a paragraph break |
| 22:40 | <Hixie> | liek <p> |
| 22:41 | <TabAtkins_> | Ah, and that restarts the algorithm. |
| 22:41 | <TabAtkins_> | We discussed having an attribute on <br> that would make it just act like a neutral character for the purposes of the bidi algorithm, but didn't end up recommending anything. |
| 22:42 | <Hixie> | so if you have "english1 english2 HEBREW3 <br> HEBREW4 english5 english6" you'll have "english1 english2 3WERBEH" "4WERBEH english5 english6" or something |
| 22:42 | <Hixie> | no wait |
| 22:42 | <Hixie> | i have it wrong |
| 22:42 | <Hixie> | but whatever |
| 22:42 | <Hixie> | it's complicated |
| 22:42 | <Hixie> | my head hurts when i try to page in the bidi algorithm |
| 22:42 | <TabAtkins_> | Yeah, I don't pretend to understand the bidi algorithm, but at least I can grasp what some of the inputs to the algo are. |
| 22:45 | <Hixie> | there was a brief period of time a few years ago where i understood it completely |
| 22:45 | <Hixie> | and could write multilevel complicated test suites for it |
| 22:45 | <Hixie> | test cases, rather |
| 23:12 | <zcorpan> | hsivonen: here's one you could use if you want: http://simon.html5.org/temp/hsivonen/favicon.png |
| 23:13 | zcorpan | is not very creative with icons |
| 23:15 | <zcorpan> | maybe that looks like it refers to Hun Stuff |
| 23:19 | <Rik`> | fun, HS in French means Hors Sujet or Hors Service which means out of order or off topic |
| 23:20 | <Hixie> | i've never seen that abbreviation in practice |
| 23:20 | <Hixie> | but my time in french-speaking land is limited these days |
| 23:21 | <gsnedders> | And I guess as a kid you don't have much reason to see it… |
| 23:23 | <david_carlisle> | hixie: if you mint a new entity name, I'd have to shoot you |
| 23:23 | <Rik`> | Hixie: all school teachers use this as comments on students copies |
| 23:24 | <Hixie> | david_carlisle: lol |
| 23:24 | <Hixie> | Rik`: interesting |
| 23:24 | <Hixie> | david_carlisle: i'd have _you_ mint it! :-) |
| 23:25 | <Rik`> | and saying "je suis complètement HS" is pretty common too |
| 23:25 | <zcorpan> | Hixie: then maybe you risk him comitting suicide |
| 23:25 | <david_carlisle> | hixie: seriously, i see that you;re just batting ideas around but if it does look like that's what you want to do, could you keep me in the loop |
| 23:26 | <Hixie> | david_carlisle: absolutely, i wouldn't do anything to the entities without the mathml wg, don't worry |
| 23:26 | <david_carlisle> | as adding another one on th exml side would probably be better than getting out of sync |
| 23:26 | <Hixie> | david_carlisle: to start with, actually adding one would be a huge pain for me if it wasn't in unicode.xml :-) |
| 23:26 | <Hixie> | david_carlisle: the html5 spec's entity stuff is literally autogenerated from unicode.xml each time, i don't have anything in the spec for it |
| 23:28 | <david_carlisle> | have you seen https://bugzilla.mozilla.org/show_bug.cgi?id=603716 |
| 23:28 | <david_carlisle> | the lawyers will probably say we can't use entities at all.... |
| 23:28 | <david_carlisle> | and I can get my life back |
| 23:30 | <zcorpan> | lol @ http://mcc.id.au/2010/novel.html |
| 23:30 | <zcorpan> | abarth pwns Hixie |
| 23:31 | <Hixie> | david_carlisle: i saw that yeah :-) |
| 23:32 | <Dashiva> | I would buy it |
| 23:34 | <jamesr_> | i dunno how much sense the story makes given both parties are employed by Google |
| 23:34 | <gsnedder1> | Oh come on, who cares. abarth kicks ass. |
| 23:37 | <Dashiva> | It should say "the Google", though |
| 23:37 | <Dashiva> | Sounds more ominous |
