| 01:35 | <Hixie> | can anyone try and see if this weasyprint thing can handle the html spec? |
| 01:48 | <guzzlefry> | Hixie: It says it passes Acid2? |
| 01:49 | <Hixie> | yes? |
| 01:53 | <guzzlefry> | Hixie: Going to assume I need more knowledge about HTML than I currently have to help with this. :P |
| 01:53 | <Hixie> | why? |
| 01:54 | <Hixie> | i just mean can someone install weasyprint and try to print the html spec and see whether it does it or not :-) |
| 01:54 | <Hixie> | as opposed to crashing or taking four hours |
| 01:54 | <guzzlefry> | ah, I might be able to do that. |
| 01:55 | <Hixie> | we used to use prince to print the PDF version but it required so much RAM to create the PDF that my server would kill the process before it could finish |
| 01:56 | <guzzlefry> | Is the spec on whatwg.org or w3c? |
| 01:56 | <Hixie> | http://whatwg.org/c |
| 03:09 | <guzzlefry> | Well, weasyprint doesn't like Gentoo. :P |
| 04:26 | <abarth> | MikeSmith: you around? |
| 04:26 | <MikeSmith> | abarth: yup |
| 04:26 | <MikeSmith> | what's up? |
| 04:26 | <abarth> | I was wondering if we could give Mike West commit access to http://dvcs.w3.org/hg/content-security-policy/ |
| 04:27 | <MikeSmith> | sure |
| 04:27 | <MikeSmith> | he has a W3C account already? |
| 04:27 | <abarth> | thanks |
| 04:27 | <abarth> | he says he had an old individual account from a while back |
| 04:27 | <abarth> | but he just requested a member account a few minutes ago |
| 04:27 | <abarth> | (he works for Google now) |
| 04:27 | <MikeSmith> | yeah found it |
| 04:29 | <MikeSmith> | OK he should have access now |
| 04:29 | <abarth> | many thanks |
| 04:29 | <abarth> | you're the best :) |
| 04:29 | <MikeSmith> | hah |
| 04:29 | <MikeSmith> | cheers man |
| 05:24 | <MikeSmith> | Hixie: I installed weasyprint but it fails when I try it with the HTML spec |
| 05:24 | <MikeSmith> | never gets past validating the stylesheet |
| 05:25 | <MikeSmith> | it appears to be first trying to validate the stylesheet before processing the document |
| 05:26 | <Hixie> | k |
| 05:26 | <Hixie> | oh well |
| 05:27 | <MikeSmith> | I did get it to work on some smaller docs |
| 05:27 | <MikeSmith> | it doesn't seem to use much memory but it's extremely slow and hogs up CPU |
| 05:29 | <MikeSmith> | Hixie: so yeah anyway I don't think it would work even if somebody took time to figure out how to workaround the thing that's causing it to fail with the HTML spec now |
| 08:51 | <zcorpan> | matjas: why doesn't http://mathias.html5.org/tests/javascript/string/ test fontsize? |
| 08:52 | <matjas> | zcorpan: because whoops. good catch |
| 08:54 | <zcorpan> | matjas: a wrapper function in tests is usually unnecessary |
| 08:54 | <matjas> | zcorpan: no test() calls? just the assert*()s? |
| 08:55 | <zcorpan> | no, i mean the outer function |
| 08:55 | <zcorpan> | (function() { .... })(); |
| 08:55 | <odinho> | You can make them globals, no prob ;] |
| 08:55 | <matjas> | ah, yeah that’s just me. don’t want to leak the typeError var |
| 08:56 | <matjas> | must not pollute global scope… not even in tests </ocd> |
| 08:56 | <odinho> | I did that at first as well. Then I learned to relax and stop worrying. |
| 08:56 | <jgraham> | That's just weird :) |
| 08:56 | <zcorpan> | best practice for tests is not exactly the same as best practice in general :-) |
| 08:56 | <matjas> | jgraham: you don’t know me very well, do you |
| 08:57 | <zcorpan> | tests should usually cut the crap |
| 08:57 | <odinho> | Figured out, less code -> better. Nice looking tests is best. |
| 09:05 | <zcorpan> | matjas: what's the moz bug number for " in link()? |
| 09:05 | <matjas> | https://bugzilla.mozilla.org/show_bug.cgi?id=352437 |
| 09:06 | <zcorpan> | thanks |
| 09:57 | <MikeSmith> | anybody know where's the place to report bugs in the mercurial sources? |
| 09:58 | <MikeSmith> | nm |
| 09:58 | <jgraham> | http://bz.selenic.com/ |
| 09:58 | <MikeSmith> | found it |
| 09:59 | <smaug____> | oh, they have switched to a saner bug tracker |
| 10:06 | <zcorpan> | annevk: i think backslash to forward slash in URLs can be restricted to file, http and https |
| 10:08 | <zcorpan> | annevk: and should only happen before the first ? or # |
| 10:09 | <zcorpan> | annevk: (except # for file seems to not be treated as a fragment delimiter in chrome) |
| 10:09 | <MikeSmith> | Content-Type: text/html; charset="UTF-8" is not valid, right? |
| 10:09 | <MikeSmith> | (note the quotes) |
| 10:09 | <MikeSmith> | a media-type should not have quotes in there, right? |
| 10:10 | <MikeSmith> | or are they optional or something? |
| 10:10 | <zcorpan> | quotes are allowed iirc |
| 10:12 | <jgraham> | annevk: re: notifications, it does seem to make somse sense to me to add language metadata even if all notification mechanisms don't support it. But it doesn't seem like 1 language / field is enough in many cases, and that can't be solved without markup (which is a really bad idea) |
| 10:13 | <zcorpan> | MikeSmith: http://www.w3.org/Protocols/rfc2616/rfc2616-sec3.html#sec3.7 |
| 10:13 | <zcorpan> | 3.6: value = token | quoted-string |
| 10:13 | <MikeSmith> | yeah, I remember I had this discussion somewhere before |
| 10:13 | <MikeSmith> | so it does allow a quoted string |
| 10:15 | <zcorpan> | quoted-string = ( <"> *(qdtext | quoted-pair ) <"> ) http://www.w3.org/Protocols/rfc2616/rfc2616-sec2.html |
| 10:15 | <zcorpan> | so only allows *double* quotes |
| 10:15 | <MikeSmith> | ah yeah I remember that now too |
| 10:15 | <MikeSmith> | nuts |
| 10:16 | <zcorpan> | xrefs in rfcs would be nice |
| 10:17 | <Ms2ger> | HTML in rfcs would be nice |
| 10:20 | <SimonSapin> | like this? http://tools.ietf.org/html/ |
| 10:21 | <Ms2ger> | No, not like that |
| 10:21 | <MikeSmith> | anyway filed http://bz.selenic.com/show_bug.cgi?id=3531 |
| 10:22 | <MikeSmith> | and I hacked the hgweb sources on dvcs.w3.org to drop the quotes around the charset value |
| 10:22 | <MikeSmith> | that was causing the validator to complain |
| 10:22 | <Ms2ger> | Fix the validator? ;) |
| 10:22 | <MikeSmith> | I remember having a discussion with Henri about that |
| 10:23 | <MikeSmith> | iirc it's because of third-party code |
| 10:23 | <MikeSmith> | maybe that chardet things |
| 10:23 | <MikeSmith> | but yeah it should be fixed, because it's wrong |
| 10:35 | <zcorpan> | Ms2ger: if you (or someone else) have a preference for what to do with https://bugzilla.mozilla.org/show_bug.cgi?id=352437 please say so in a comment |
| 10:36 | <Ms2ger> | I've got a patch to match matjas's spec, just need to clean it up a bit |
| 10:37 | <matjas> | FWIW, I’ve got a V8 patch ready to get rid of the unneeded escapes |
| 10:37 | <matjas> | let’s align some browser implementations \o/ |
| 10:38 | <zcorpan> | Ms2ger: ok, nice |
| 10:50 | <Von_Davidicus> | I have a question about website coding--particularly the use of a database. |
| 10:51 | <Von_Davidicus> | If you had a field that required a massive amount of text, would you store it in the database, or in a seperate text file, and refer to it via database? |
| 10:51 | <Philip`> | What's "massive"? |
| 10:51 | <Philip`> | (Megabytes? Gigabytes?) |
| 10:52 | <Von_Davidicus> | Anywhere from 2 KB to 25 KB. They're short stories. |
| 10:53 | <Philip`> | That sounds tiny :-) |
| 10:53 | <Von_Davidicus> | My question still stands. |
| 10:53 | <Philip`> | so I'd store it in a MySQL TEXT column (or equivalent in other databases) |
| 10:54 | <Philip`> | since storing it externally would be a lot of hassle for no real benefit |
| 10:54 | <Von_Davidicus> | Aside from editability, but I think I see your point. |
| 10:55 | <Philip`> | (Oh, I suppose not actually TEXT, since that's stupidly small; might as well use LONGTEXT) |
| 10:56 | <Von_Davidicus> | The reason I had them in seperate text files was so that I could edit them in a text editor. |
| 10:58 | <Philip`> | I imagine it'd be useful to do things like store historical revisions of every story, and store metadata like who last modified the data and when, so you'd always be required to edit them through the proper database-based interface instead of using a text editor directly |
| 10:58 | <Von_Davidicus> | Okay... |
| 10:59 | <Philip`> | unless you're absolutely certain your design will never need any data other than what the filesystem already provides |
| 10:59 | <zcorpan> | matjas: have you filed a bug with IE about this? |
| 10:59 | <matjas> | zcorpan: not yet, only about the fact that IE uppercases the tag names |
| 10:59 | <jgraham> | Philip`: Obviosuly you should put git into the mix to get revisions |
| 11:00 | <Von_Davidicus> | Meaning? |
| 11:00 | <matjas> | zcorpan: https://connect.microsoft.com/IE/feedback/details/752283 if you’re interested. I’ll file another one for the " escape thing |
| 11:00 | <Philip`> | Von_Davidicus: Knowing jgraham, that was probably sarcasm |
| 11:00 | <zcorpan> | matjas: thanks. i'm interested in the " thing :-) |
| 11:01 | Von_Davidicus | was directing the "meaning" to you, Philip` |
| 11:01 | <Von_Davidicus> | With the text files, all I do is use PHP's include() function. |
| 11:01 | <Philip`> | Ah :-) |
| 11:01 | <Von_Davidicus> | Heck, I could almost get away with using SSI. |
| 11:01 | <Philip`> | Sure, you could choose to use horrid technologies if you really wanted to :-p |
| 11:02 | <Von_Davidicus> | SSI's not bad. I've used it on a page with a several -massive- tables. |
| 11:03 | <Von_Davidicus> | As I said, I could -almost- get away with it; I get the title and story rating info from the database. |
| 11:07 | <Von_Davidicus> | Besides, you want horrid? Try using XML files as your database. |
| 11:08 | <Von_Davidicus> | BOY, that gets complicated. |
| 11:09 | <Philip`> | If you store the story text in the database too then you get simpler code, more flexibility, more nice ACID properties, easier backups, better performance, etc, and if you want to edit a story in a text editor then just write a script that downloads it from the database into a text file and then you edit it and then upload it back again |
| 11:10 | <zcorpan> | matjas: maybe you could put in a recommendation for authors saying something along the lines of "use of these methods is not recommended; using DOM APIs is recommended. however, if these methods are used nevertheless, the characters & and " should be escaped as & and ", respectively" |
| 11:10 | <Philip`> | so I don't see why you'd make life harder for yourself :-) |
| 11:10 | <Von_Davidicus> | ACID properties? |
| 11:10 | <Philip`> | Like in http://en.wikipedia.org/wiki/ACID |
| 11:10 | <matjas> | zcorpan: filed https://connect.microsoft.com/IE/feedback/details/752391 |
| 11:11 | <matjas> | zcorpan: i have a note here: http://mathias.html5.org/specs/javascript/#escapeattributevalue but i guess it would be clearer to have one near the actual string method definitions |
| 11:11 | <Philip`> | (e.g. the database system will prevent the data and metadata getting out of sync with each other, but you get no such guarantees if you involve the filesystem unless you're extremely careful) |
| 11:12 | <Von_Davidicus> | Hmmm... never had a problem with it before. |
| 11:13 | <zcorpan> | matjas: hmm, i don't have permission to view it |
| 11:13 | <matjas> | zcorpan: http://i.imgur.com/G73ww.png |
| 11:13 | <Philip`> | Von_Davidicus: Rare bugs are the worst :-) |
| 11:14 | <Von_Davidicus> | I suppose I could throw this into the database... and have PHP insert the tab characters and line breaks in the code where they should be. |
| 11:14 | <Philip`> | Von_Davidicus: Everything will look like it's running fine for several years as you get more and more users, and then the server's PSU blows up and you're very unlucky and you find that half of your data is corrupted and irrecoverable |
| 11:15 | <Von_Davidicus> | That, sir, is why I have a backup site on my own computer. And I back --that-- up, too. |
| 11:17 | <Philip`> | If your backup is a snapshot of the server while it's in a temporarily inconsistent state (e.g. because it's updated its metadata but not yet written the actual data) then that still won't solve the problem |
| 11:18 | <Von_Davidicus> | You have it backwards, sir. What's on the -server- is a snapshot of what's on my computer. I edit my site on my computer, then I upload whatever files I need to and run whatever SQL I have to. |
| 11:23 | <zcorpan> | annevk: hmm, the chrome file # thing seems is just for file://#/ or file:\\#\ - hash anywhere else is fragment. but that seems bogus and opera/firefox don't do that |
| 11:26 | <zcorpan> | http://# resolves to http:/# in chrome and http:///# in opera/firefox |
| 11:34 | <matjas> | zcorpan: “using DOM APIs is recommended” — should I add a reference there? to [DOM2HTML], of something else? |
| 11:34 | <matjas> | s/of/or/ |
| 11:36 | <zcorpan> | matjas: dom2html is obsolete, so preferably not :-) |
| 11:36 | <zcorpan> | i guess dom4 and html if you want to reference something |
| 11:37 | <matjas> | i couldn’t find property attributes in DOM4 right away so i got confused |
| 11:41 | <zcorpan> | matjas: i get an empty doc now |
| 11:42 | <matjas> | zcorpan: fixed |
| 11:46 | <zcorpan> | matjas: thanks. i guess the dom api recommendation applies to all of those, but it seems annoying to repeat it over and over. maybe put the note at the top of the section, and say that the escaping recommendation only applies to those that take an argument, or so? |
| 11:46 | <matjas> | zcorpan: but there’s also substr |
| 11:46 | <AryehGregor> | matjas, I suggest explicitly saying in the note that browsers have historically not escaped " in the argument, so anyone using the function is likely to be bitten by XSS unless they escape. |
| 11:48 | <annevk> | zcorpan: not ftp? |
| 11:48 | <zcorpan> | matjas: i'd solve that by removing substr from that section and say in a separate section that annex b must be supported :-) |
| 11:48 | <annevk> | zcorpan: I figured if we have a whitelist of hierarchical schemes we can do all that magic there |
| 11:48 | <annevk> | zcorpan: including path normalization and such |
| 11:48 | <zcorpan> | annevk: no occurrences of backslashes in ftp in web200904 |
| 11:49 | <zcorpan> | annevk: see CT-3084 |
| 11:50 | <zcorpan> | annevk: isn't it good to keep the backslash rewriting to a minimum? |
| 11:50 | <annevk> | zcorpan: I rather keep the special scheme lists to a minimum |
| 11:50 | <zcorpan> | annevk: why? |
| 11:51 | <annevk> | less code paths |
| 11:51 | <annevk> | one code path for hierarchical schemes, one for the rest |
| 11:51 | <annevk> | not sure what the point is in not admitting defeat on \ |
| 11:52 | <zcorpan> | well, i dunno, new schemes get to use \ as \ |
| 11:53 | <zcorpan> | but i don't care about it strongly |
| 11:53 | <zcorpan> | it can be escaped |
| 11:54 | <annevk> | right, and it seems better if new hierarchical schemes behave consistently with existing hierarchical schemes |
| 11:57 | <annevk> | jgraham: so apparently there's a way to embed language metadata through Unicode |
| 11:57 | <annevk> | jgraham: but it's "deprecated" or some such |
| 11:57 | <jgraham> | :-o |
| 11:57 | <jgraham> | Really? |
| 11:57 | <jgraham> | Wow |
| 11:58 | <annevk> | http://en.wikipedia.org/wiki/Unicode_control_characters#Language_tags |
| 12:01 | <zcorpan> | matjas: assert: rfc2119 words in a note :-P |
| 12:01 | <matjas> | how do you check for that? is that some anolis setting? |
| 12:02 | <zcorpan> | it isn't. but that's not a bad idea, actually |
| 12:05 | <annevk> | can you nest dictionaries? |
| 12:05 | <annevk> | dictionary UIString { DOMString value; DOMString lang; Direction dir; } |
| 12:05 | <annevk> | or some such |
| 12:06 | <zcorpan> | var a = { dir: a } |
| 12:12 | <annevk> | boo |
| 12:19 | <Lachy> | zcorpan, that doesn't work, since a is undefined at time of assignment. This works: var a = {}; a.dir = a |
| 12:20 | <zcorpan> | Lachy: good point |
| 12:38 | <matjas> | V8 will remove the unneeded escapes in its HTML String methods as soon as other browsers add the " escape: http://code.google.com/p/v8/issues/detail?id=2217#c8 |
| 12:44 | <gsnedders> | matjas: They're not pointless if the web relies on them. :) |
| 12:44 | <matjas> | gsnedders: no other browser does it so i doubt the web relies on them |
| 12:45 | <annevk> | opposite is more likely |
| 12:45 | <gsnedders> | It seems better to escape too much than too little, IMO |
| 12:47 | <zcorpan> | gsnedders: let's not muddy the waters |
| 12:48 | <gsnedders> | I'd rather have the methods not lead to trivial XSS holes. |
| 12:48 | <zcorpan> | <>' does't |
| 12:48 | <zcorpan> | " does |
| 12:48 | <zcorpan> | the spec says to escape only " |
| 12:49 | <zcorpan> | let's stick to that |
| 12:49 | <gsnedders> | Oh, matjas's new spec? |
| 12:49 | <zcorpan> | yes |
| 12:49 | <zcorpan> | we can't escape & since that would lead to double-escaping for sites that "do the right thing" |
| 12:49 | <zcorpan> | so escaping <>' is pointless |
| 12:53 | <matjas> | zcorpan: if & was escaped into &, would it be useful to escape <>'? |
| 12:53 | <annevk> | matjas: why would you escape & or < and >? |
| 12:54 | <matjas> | annevk: i wouldn’t, just wondering… i can’t think of any XSS vector using <>' even if & gets escaped |
| 12:54 | <matjas> | in quoted attribute values with double quotes, that is |
| 12:54 | <zcorpan> | matjas: no |
| 12:54 | <zcorpan> | innerHTML doesn't escape those either |
| 12:55 | <zcorpan> | for innerHTML it was a compat problem to escape <> |
| 12:57 | <zcorpan> | (opera always did it, ie never did it, webkit and firefox usually escaped it but didn't for href="javascript:<>", and of course a sitecompat bug was filed on opera where it expected it to not be escaped for href="javascript:<>") |
| 12:57 | <zcorpan> | (iirc) |
| 12:57 | <gsnedders> | zcorpan: In Carakan or Futhark? |
| 12:57 | <zcorpan> | gsnedders: innerHTML |
| 12:58 | <zcorpan> | don't recall if it was pre- or post- carakan shipping |
| 12:58 | <gsnedders> | zcorpan: Ah, not String.prototype.* |
| 13:01 | <zcorpan> | matjas: so "Note: For security reasons, Web authors should not rely on any such escaping, as older, non-JavaScript-compliant implementations may not perform this operation correctly." shouldn't use "should" or "may" |
| 13:01 | <matjas> | zcorpan: yeah, havent gotten around to it yet :) which verb to pick… |
| 13:02 | <zcorpan> | k |
| 13:02 | <zcorpan> | matjas: it kinda repeats the other note, doesn't it? |
| 13:06 | <matjas> | the one near EscapeAttributeValue? yeah |
| 13:06 | <matjas> | but this way, Web authors who only read a specific part of the spec, get to see the warning anyway |
| 13:06 | <zcorpan> | that's the one i quoted |
| 13:07 | <matjas> | only having the warning once would assume everyone read the full spec |
| 13:07 | <annevk> | matjas: instead of class=note, use class=warning |
| 13:07 | <annevk> | matjas: and keep it normative |
| 13:07 | <matjas> | annevk: thanks |
| 13:07 | <matjas> | you’re all so helpful, i love it |
| 13:08 | <zcorpan> | that works, though "may not" should still be changed |
| 13:09 | <matjas> | s/may not/might not/ |
| 13:10 | <zcorpan> | yep |
| 13:29 | <zcorpan> | i wrote an introduction, http://simon.html5.org/specs/quirks-mode#introduction |
| 13:33 | <annevk> | "standards mode but expected quirks" missing comma? |
| 13:33 | <zcorpan> | dunno, does it need a comma? |
| 13:33 | <annevk> | "apart from CSS" -> "outside of CSS" |
| 13:33 | <annevk> | I thought but would typically be preceded by a comma |
| 13:35 | <gsnedders> | annevk: Matter of style. |
| 13:36 | <zcorpan> | annevk: added a comma but a bit earlier. also changed the other thing. thanks |
| 13:37 | <matjas> | s/further/since/ but I suppose that’s style too |
| 13:38 | <zcorpan> | fixed |
| 13:39 | <matjas> | zcorpan: what’s your workflow for writing specs and hosting them on simon.html5.org? did you manage to get Python + Anolis running on the server? I just run everything locally and rsync |
| 13:39 | <zcorpan> | i run it locally |
| 13:40 | <zcorpan> | my plan is to move it to dvcs.w3.org |
| 14:32 | <smaug____> | jgraham: ping |
| 14:33 | <smaug____> | or any Opera-er |
| 14:33 | <AryehGregor> | matjas, I edit things like that on my server. Works for me because I use vim anyway. |
| 14:34 | <matjas> | AryehGregor: cool. I was asking ’cause I had no rights to install Python / Anolis on the html5.org server |
| 14:35 | <zcorpan> | smaug____: pong |
| 14:35 | <matjas> | I use https://github.com/cowboy/grunt to “watch” the source files, so whenever I save, Anolis generates a new spec version |
| 14:35 | <smaug____> | zcorpan: do you happen to know about Presto's JS<->C++ handling |
| 14:35 | <smaug____> | especially, is it difficult to handle cases when C++ keeps JS alive |
| 14:35 | <smaug____> | like event listeners |
| 14:36 | <zcorpan> | smaug____: i don't really, sorry |
| 14:36 | smaug____ | is amazed that it seems to be so hard in WebKit that webkit-guys want to change an API just because of that |
| 14:37 | <smaug____> | zcorpan: who might know? |
| 14:37 | <zcorpan> | gsnedders or jgraham maybe |
| 14:37 | <odinho> | I don't understand the question. But I've only scratched Opera's code. Done some very small DOM stuff. |
| 14:38 | <zcorpan> | else we can ask some relevant developer |
| 14:38 | <smaug____> | odinho: when you do .addEventListener("foo", somefunction), native code needs to keep that JS function object alive |
| 14:38 | <smaug____> | odinho: based on comments from abarth it is hard to not leak in that case in webkit |
| 14:39 | <smaug____> | eventlisteners are apparently very much special cased there |
| 14:39 | <jgraham> | smaug____: I will ask someone who knows |
| 14:39 | <jgraham> | But I think their proposed replacement looks horrible |
| 14:39 | <odinho> | Where is this, I want to see it :D |
| 14:39 | <smaug____> | yeah, the new DOMTranction() thing is very odd |
| 14:39 | <smaug____> | odinho: webapps wg |
| 14:40 | <odinho> | Oh, that. |
| 14:40 | <zcorpan> | i know we have special cases for event listeners to make sure we don't gc the object if it can still get events for those listeners (like an open websocket or xhr) |
| 14:40 | <smaug____> | about UndoManager |
| 14:40 | <gsnedders> | smaug____: The only real hard case for us is stuff like XHR event handlers when the XHR object itself is GC'd, I believe. |
| 14:40 | <odinho> | Yea, I saw undomanager and was like "oh, wait till the end of the day" |
| 14:40 | <gsnedders> | (Which I imagine is fairly inelegant for everyone) |
| 14:40 | <smaug____> | In Gecko XHR object can't be collected if the connection is still open |
| 14:41 | <odinho> | smaug____: Yea, -- someone said here that we have no such problem. |
| 14:41 | <jgraham> | odinho: Who? |
| 14:41 | <odinho> | jgraham: joaoe :-) |
| 14:41 | <smaug____> | odinho: was that about XHR, or JS callbacks in general? |
| 14:42 | <odinho> | smaug____: js callbacks in general. |
| 14:42 | <smaug____> | ok, good |
| 14:43 | gsnedders | is somewhat amused by how he avoids DOM code, despite dealing with both JS and CSS at times |
| 14:44 | <zcorpan> | and i try to avoid JS :-P |
| 14:49 | <AryehGregor> | We apparently have this magical thing called a cycle collector that's built out of macros and pixie dust, which mysteriously makes things not leak if you enlist one of the two people who know how it works to tell you the right incantation to add to your classes. |
| 14:53 | <MikeSmith> | hah |
| 14:54 | <MikeSmith> | I read that thread |
| 14:55 | AryehGregor | discovers that Element.style is readonly . . . why? |
| 14:55 | <AryehGregor> | Wouldn't you expect setting it to work like setAttribute(), same as practically all such attributes? |
| 14:57 | <hober> | you would |
| 14:58 | <zcorpan> | you do style.foo = 'bar' instead |
| 14:58 | <hober> | but el.style is not a string corresponding to <el style=""> |
| 14:59 | <AryehGregor> | hober, sure, but it could still be settable. .classList is settable, right? |
| 15:00 | <zcorpan> | nope |
| 15:10 | <AryehGregor> | Hmm, so it's not. |
| 15:10 | <AryehGregor> | Oh well. |
| 15:10 | <AryehGregor> | It confused me. |
| 15:15 | <zcorpan> | MikeSmith: how do i create a new repo on dvcs.w3.org? |
| 15:15 | <Ms2ger> | zcorpan, ask MikeSmith :) |
| 15:16 | <MikeSmith> | yah |
| 15:16 | <MikeSmith> | just lemme know the name |
| 15:16 | <zcorpan> | ok, how about 'quirks-mode' |
| 15:18 | <MikeSmith> | OK |
| 15:18 | <MikeSmith> | zcorpan: who should have write access? |
| 15:19 | <zcorpan> | um. me, to begin with, i guess we can add more later if needed |
| 15:21 | <zcorpan> | annevk: why http://dvcs.w3.org/hg/encoding/rev/55a9e6c1aa32 ? |
| 15:22 | <annevk> | zcorpan: lawyers? |
| 15:22 | <annevk> | zcorpan: tantek reverted that for Fullscreen |
| 15:23 | <annevk> | zcorpan: so maybe it's okay again |
| 15:23 | <annevk> | MikeSmith: spec-writers group? |
| 15:23 | <MikeSmith> | annevk: for quirks-mode draft? |
| 15:24 | <annevk> | zcorpan: in particular it's something chaals combined with W3C cared about and nobody else |
| 15:24 | <AryehGregor> | annevk, I was asked to get rid of it but politely said no, and no one sicced any lawyers on me yet. |
| 15:25 | <annevk> | AryehGregor: took them about six months to get back to me and then I lost some silly internal argument and gave up for now |
| 15:29 | <zcorpan> | annevk: ok :-/ |
| 15:30 | <annevk> | should not be a problem now though |
| 15:34 | <matjas> | where does the web ecmascript identifier thing originate? |
| 15:34 | <matjas> | trying to understand if it can be dropped or not |
| 15:36 | <annevk> | if IE/Gecko dropped it chances are the others want to follow |
| 15:36 | <zcorpan> | matjas: probably at the time we implemented carakan, jgraham noticed that everyone else did it so assumed it had to be supported |
| 15:37 | <matjas> | yay \o/ |
| 15:38 | <annevk> | matjas: filing bugs on the other engines with "remove this because IE/Gecko don't" is usually sufficient |
| 15:39 | <jgraham> | matjas: Right, what zcorpan said. I was generally pretty liberal with the definition of "everyone" since it is better to have too much compat than too little |
| 16:00 | <MikeSmith> | zcorpan: https://dvcs.w3.org/hg/quirks-mode created |
| 16:01 | <zcorpan> | MikeSmith: thanks! |
| 16:06 | <annevk> | zcorpan: I guess the best you can do is to just keep it under the current copyright and publish it through WHATCG once the W3C resolves some of the outstanding questions |
| 16:07 | <zcorpan> | http://dvcs.w3.org/hg/quirks-mode/raw-file/tip/Overview.html |
| 16:07 | <annevk> | or that :) |
| 16:07 | <annevk> | do you want me to make it official? |
| 16:07 | <zcorpan> | uh, the copyright shouldn't say fullscreen.... |
| 16:08 | <annevk> | yeah |
| 16:09 | <annevk> | zcorpan: the only other thing that's semi-required is an email like this one: http://lists.w3.org/Archives/Public/public-whatwg-contrib/2012Apr/0008.html |
| 16:11 | <zcorpan> | hmm does MakeFile need a special chmod or something again? |
| 16:11 | <matjas> | chmod +x Makefile |
| 16:12 | <zcorpan> | still, "make: Nothing to be done for `all'." |
| 16:12 | <annevk> | zcorpan: "touch Overview.src.html" |
| 16:12 | <annevk> | in doubt |
| 16:13 | <zcorpan> | oh wait i edited the wrong file |
| 16:13 | <zcorpan> | heh |
| 16:17 | <zcorpan> | annevk: i don't need to subscribe first, right? |
| 16:17 | <zcorpan> | shit, need to run |
| 16:25 | <matjas> | what’s the advantage of having a spec be hosted on w3.org? |
| 16:28 | <smaug____> | matjas: comparing to what? |
| 16:28 | <matjas> | smaug____: compared to anywhere else, e.g. on your own site |
| 16:29 | <smaug____> | well, some credibility |
| 16:29 | <Lachy> | matjas, depends what you mean. The patent policy can help if you develop the spec in a WG, task force, community group or whatever. |
| 16:30 | <Lachy> | but there are specs that have been implemented despite being written and hosted on personal sites, at least initially. |
| 16:31 | <odinho> | And it kinda helps finding it... |
| 16:37 | <matjas> | I see. Use of the bug tracker is nice too, I imagine |
| 17:45 | <annevk> | zcorpan: hmm dunno |
| 17:45 | <annevk> | zcorpan: if you join http://www.w3.org/community/whatwg/ you are subscribed |
| 17:46 | <annevk> | zcorpan: I'll just publish it and you can sort that out later |
| 17:46 | <annevk> | zcorpan: oh also, ij prefers if instead of "Standard" it says "Specification" |
| 18:25 | <hober> | Lachy: are you still editing the about: spec? |
| 18:26 | <Lachy> | yes, sort of. |
| 18:26 | <Lachy> | though I haven't looked at it in a while. |
| 18:26 | <hober> | where should people send feedback? to the whatwg lists with [about] in the subject? |
| 18:26 | <smaug____> | what is about: spec about? |
| 18:26 | <smaug____> | about:blank ? |
| 18:26 | <Lachy> | probably better to send it to the IETF list. |
| 18:27 | <Lachy> | smaug____, yes, and otehrs |
| 18:27 | <Lachy> | *others |
| 18:27 | <smaug____> | what all others |
| 18:27 | <hober> | do you mean uri⊙wo? |
| 18:28 | <smaug____> | oh, it is about the protocol, not about the about:* pages |
| 18:29 | <Lachy> | hober, ietf⊙io |
| 18:29 | <hober> | ahh, uri-review⊙io the number of url mailing lists is too damn high. |
| 18:29 | <Lachy> | oh, yes |
| 18:29 | <Lachy> | that's the one |
| 18:29 | <Lachy> | discussion has shifted so many times for that draft, even I lose track of where it should be |
| 18:30 | <Lachy> | at one stage, they even started having discussions on a list that I wasn't subscribed to without telling me |
| 18:30 | <hober> | oh awesome |
| 18:30 | <Lachy> | that was fun catching up on everything I missed |
| 18:31 | <Lachy> | smaug____, http://www.ietf.org/id/draft-ietf-appsawg-about-uri-scheme-07.txt |
| 18:31 | <Lachy> | that's the latest draft |
| 18:35 | <Lachy> | WTF? someone has completely re-written that and taken over |
| 18:35 | <Lachy> | I'm not even listed as editor any more |
| 18:36 | <smaug____> | "nice" |
| 18:37 | <Lachy> | oh well, I guess I'm not the editor of that any more. |
| 18:42 | <hober> | Lachy: weird |
| 18:42 | <hober> | there's an about registry now? ugh. |
| 18:43 | <Lachy> | yeah, I know. They pushed that through in the previous draft, despite my objections. |
| 18:43 | <Lachy> | After that, I just sort of gave up |
| 18:43 | <hober> | it doesn't even say where comment should go anymore |
| 18:44 | <Lachy> | I think the latest location for discussion is here now. apps-discuss⊙io |
| 18:45 | <Lachy> | I'm pretty sure that's the group that took over the work |
| 18:53 | <annevk> | hober: Lachy: not sure why you still bother with discussing it there |
| 18:54 | <annevk> | hober: Lachy: did anything useful come out of it? |
| 18:54 | <annevk> | hober: Lachy: last I heard it was mostly bad (e.g. setting up another IANA registry...) |
| 18:54 | <Lachy> | annevk, I haven't done anything on it for about a year |
| 18:55 | <Lachy> | If there's enough interest, I could republish the draft elsewhere and just completely ignore the IETF |
| 18:55 | <annevk> | http://tools.ietf.org/html/draft-ietf-appsawg-about-uri-scheme is a better link |
| 18:56 | <annevk> | maybe we should fold it into the URL spec |
| 18:57 | <annevk> | it should really not be more than a couple of paragraphs anyway |
| 19:23 | <Lachy> | annevk, fine with me. If you like, I can write up a draft, dropping a whole bunch of unnecessary and overly complex stuff, and make that easier to fold in to the URL spec. |
| 19:24 | <Lachy> | where is the current URL spec though? |
| 19:41 | <pablof> | http://dvcs.w3.org/hg/url/raw-file/tip/Overview.html ? |
| 19:44 | <jgraham> | timeless: Thank you |
| 19:54 | <jgraham> | pablof: Was there some context for that "question"? |
| 19:55 | <pablof> | jgraham: "Lachy: where is the current URL spec though?" |
| 19:55 | <pablof> | the question mark was meant to imply that i'm somewhat, but not totally sure :-) |
| 19:56 | <jgraham> | I see |
| 19:56 | <Lachy> | thanks |
| 19:56 | <jgraham> | I might have got that if the line had started "Lachy: " |
| 19:56 | <pablof> | yeah, i should've done that |
| 19:56 | <jgraham> | Or if I read the scrollback more closely ;) |
| 20:01 | <jgraham> | Argh gmail rage |
| 20:03 | <jgraham> | It makes me so sad that they can't get basic formatting of email replies correct |
| 20:03 | <jgraham> | How hard can it possibly be? |
| 20:07 | <hober> | jgraham: srsly |
| 20:07 | <hober> | fortunately, you can point gnus at gmail's imap. :) |
| 20:32 | <Yuhong> | Hey, I have an idea: Why not use the versioned HTML specs produced by the HTML WG as *authoring* requirements only? |
| 20:33 | <Yuhong> | That validators can validate. |
| 20:34 | <timeless> | jgraham: what did i do? |
| 20:34 | <timeless> | oh |
| 20:35 | <timeless> | um, i don't think i sent it to enough groups |
| 20:35 | <timeless> | that was on my todo list for one group that hasn't started messing up yet |
| 20:35 | <timeless> | i need to send it out to a couple of others |
| 20:35 | <timeless> | i'm not even sure i sent it to the right offender group |
| 20:35 | <timeless> | iirc the primary offenders are in DAP and i sent it to coremob |
| 20:37 | <timeless> | i sent out a better welcome for the intents list, and they were well behaved |
| 20:39 | <hober> | timeless: ? |
| 20:46 | <timeless> | hober: i'm assuming jgraham thanked me for sending an eitquette email to some list |
| 20:47 | <timeless> | http://lists.w3.org/Archives/Public/public-coremob/2012Jul/0015.htmlhttp://lists.w3.org/Archives/Public/public-coremob/2012Jul/0015.html |
| 20:47 | <timeless> | grr |
| 20:47 | <timeless> | http://lists.w3.org/Archives/Public/public-coremob/2012Jul/0015.html |
| 20:47 | timeless | kicks chrome for being awful |
| 20:48 | <hober> | ahh. yes, lovely, thanks. :) |
| 20:48 | <timeless> | please please feel free to plagiarize that |
| 20:48 | <timeless> | or reference it, whichever, i'm not picky |
| 20:50 | <Velmont> | timeless: Cool. But do you think people understand it? |
| 20:50 | <timeless> | dunno, i got a thanks from an offender |
| 20:50 | <timeless> | but he used outlook style --- original message--- w/ top quoting as he madehis reply... |
| 20:51 | <othermaciej> | timeless: I have to admit, I hate the footnote style for links in w3c mailing list mail |
| 20:51 | <othermaciej> | timeless: where did that originate? |
| 20:51 | <timeless> | it predates me |
| 20:52 | <timeless> | it probably relates to wanting to have a clen message |
| 20:52 | <othermaciej> | (I hate it because my mail client, and probably most reasonably modern ones, know how to make the URL clickable, but not the little bracketed number, so it would work better to just put the URL inline) |
| 20:52 | <timeless> | and also avoiding stupid line wrapping breaking links |
| 20:52 | <timeless> | or quoting breaking links |
| 20:52 | <timeless> | and uoting w/ line wrapping dfefinitely breaks links |
| 20:52 | <timeless> | this chrome session is causing tremendous lag to my typing, so i can't see anything |
| 20:52 | <Velmont> | I normally write urls after a paragrah with its own line, 4 spaces to indent it. |
| 20:52 | timeless | blames chrome for sucking |
| 20:53 | <Velmont> | timeless: Try Opera ;-) |
| 20:53 | <timeless> | Velmont: hrm |
| 20:53 | <othermaciej> | yeah, I like them inline on their own line, unless there is an excessive number of links and/or they are very short |
| 20:53 | <Velmont> | We did it first! |
| 20:53 | <othermaciej> | doing the [1] [2] thing makes it much harder to follow links with context in mind, especially for a long email |
| 20:53 | <timeless> | Velmont: you did sucking first? |
| 20:54 | <timeless> | othermaciej: you can cheat by copying the bottom into a notepad and having it alongside |
| 20:55 | <Velmont> | timeless ;-) |
| 20:55 | <othermaciej> | timeless: I think you just convinced me to never do the [1] [2] thing again |
| 20:55 | <othermaciej> | (I do it sometimes in my emails, but no more) |
| 20:55 | <Hixie> | i generally do the [1] [2] thing if I don't expect people to follow the links but just want to include them for completeness |
| 20:55 | <Hixie> | e.g. to back up a point i'm making |
| 20:55 | <timeless> | i think Hixie 's explanation is better |
| 20:55 | <Hixie> | whereas if i want people to look at the link, i include it inline |
| 20:55 | <Velmont> | Hixie: Seems reasonable. Doing it for "real" footnotes. |
| 20:55 | <othermaciej> | I like to actually check people's references |
| 20:56 | <othermaciej> | if emails could be HTML then of course you could have clickable links to the footnotes, getting sort of the best of both worlds |
| 20:56 | <Hixie> | Velmont: yeah, sometimes they're not even URLs, they're just textual footnotes |
| 20:56 | <othermaciej> | by W3C email has the same modern text technology as IETF RFCs |
| 20:56 | <othermaciej> | s/by/but/ |
| 20:57 | <Velmont> | othermaciej: Well helped by useless html implementation by email clients... |
| 20:57 | <othermaciej> | but yeah, I guess it is a good technique if you have references that you don't care to have checked |
| 20:57 | <othermaciej> | people send HTML email to each other all the time at Apple and no one ever even notices, but we have a mostly monolithic set of client software with known good html rendering and editing |
| 20:58 | <timeless> | Velmont: Sorry, we’re only supporting browsers with WebSockets enabled at the moment |
| 20:58 | <timeless> | othermaciej: and people don't use fancy watermarks / backgrounds? :) |
| 20:59 | <othermaciej> | mostly they use bold, italic, monospace fonts for code sections, and the occasional distinguished section header, as well as, of course, links |
| 20:59 | <othermaciej> | and including images inline when relevant |
| 20:59 | <timeless> | iow you have people who aren't stupid |
| 21:00 | <timeless> | silly company self-selecting for intelligence |
| 21:00 | <othermaciej> | well, the mail client doesn't have a ready affordance for setting an ugly background |
| 21:00 | <othermaciej> | I expect most html email with fugly backgrounds is generated by a script, not a conventional mail client |
| 21:00 | <timeless> | silly company trying to restrict users from the freedom to send mail that's unreadable! |
| 21:01 | <othermaciej> | but perhaps I am naiive |
| 21:01 | <timeless> | i should try to do a survey of that, but i'm not sure where a valid sample input would be |
| 21:01 | <timeless> | none of my mailboxes are valid |
| 21:01 | <timeless> | all but one of them have 99.99% bugzlila bugmail |
| 21:01 | Hixie | notes that he never sees the backgrounds, since his e-mail client is a black-and-white monospace text-mode client with one font size :-) |
| 21:02 | <timeless> | does it support black-on-black or white-on-white? :) |
| 21:02 | <timeless> | Velmont: anyway, opera was a total failure |
| 21:02 | <timeless> | maybe i should try ie9? |
| 21:02 | <Velmont> | timeless: It has WebSocket, but it's disabled by default :P The old protocol that was vulnerable. |
| 21:02 | <Velmont> | timeless: ie10 then. If Opera fails, I would be sad if IE9 worked. |
| 21:02 | <timeless> | is is there a version of opera that's useful? |
| 21:03 | <Hixie> | timeless: it essentially doesn't support colour at all, just text |
| 21:03 | <Velmont> | timeless: No need to be snarky, -- ofc it's useful. I use it every day. |
| 21:03 | <timeless> | Velmont: for my purposes |
| 21:03 | <timeless> | i'm ok w/ a beta or alpha as long as someone tells me how to get it |
| 21:03 | <timeless> | just as you suggested ie10 |
| 21:04 | <Velmont> | timeless: I wouldn't know. Don't know what you do etc. |
| 21:04 | <timeless> | alpha.irccloud.com |
| 21:04 | <timeless> | i don't browse the web or read email or anything like that |
| 21:04 | <Yuhong> | Hixie: Why not use the versioned HTML specs produced by the HTML WG as *authoring* requirements only (for validators)? |
| 21:04 | <Hixie> | Yuhong: ? |
| 21:05 | <Yuhong> | I am referring to the W3C HTML WG. |
| 21:05 | <Hixie> | what is the problem you are trying to solve? |
| 21:05 | <timeless> | this is #whatwg, not irc.w3.org #html ... |
| 21:05 | <Hixie> | timeless: there's no topic restriction in this channel |
| 21:05 | <jgraham> | timeless: New websockets is coming |
| 21:05 | <timeless> | jgraham: nick highlight me when they're available |
| 21:06 | <jgraham> | Hah, like I will remember that |
| 21:06 | <timeless> | i'll gladly replace chrome canary, since righ now it paints random triangular wedges across my screen |
| 21:06 | <timeless> | ie9 didn't work either.. |
| 21:07 | <timeless> | Velmont: just so you don't feel bad :) |
| 21:07 | timeless | considers replacing chromium w/ nightly for a bit |
| 21:07 | <Yuhong> | Browsers typically treat all content the same, but *validators* often lets you choose different HTML versions to validate against. |
| 21:08 | <Hixie> | Yuhong: a validator that lets you test against an old version seems kinda silly |
| 21:08 | jgraham | also uses alpine rather often fwiw |
| 21:08 | <Hixie> | Yuhong: why would you want to know if a page was valid against some know-buggy rules? |
| 21:08 | timeless | uses mail.google.com/mail/u/0/x/ often enough |
| 21:08 | <Velmont> | timeless: I couldn't test irccloud because it's not possible to make a user :P |
| 21:08 | <Yuhong> | But most validators do support it. |
| 21:08 | <timeless> | Velmont: i can probably get you an invite if you're interested |
| 21:08 | <timeless> | Hixie: your patience continues to impress me |
| 21:09 | <Hixie> | Yuhong: most validators don't support any version of HTML from the past decade |
| 21:09 | <Velmont> | timeless: I'll make an operatester account. I like irssi for IRC myself, and have dogfooded on the Opera IRC client a bit, -- but then I went back to irssi :P |
| 21:09 | <timeless> | (most developers don't use any validators) |
| 21:09 | <Hixie> | Yuhong: so i don't think argument from popularity is a good idea in this case :-) |
| 21:09 | <Yuhong> | Except that Validator.nu does support validating against HTML4. |
| 21:10 | <Hixie> | Yuhong: so? |
| 21:10 | <timeless> | i don't think that's from the last decade |
| 21:10 | timeless | wonders what the pub date on html 4 is anyway |
| 21:10 | <Hixie> | timeless: 1997-2001 depending on which minor revisionversion you mean |
| 21:10 | <Hixie> | revision/version even |
| 21:10 | <timeless> | July 8, 1997 -- The World Wide Web Consortium (W3C) today announced the first public working draft of HyperText Markup Language (HTML) 4.0 |
| 21:11 | <timeless> | Yuhong: note that the past decade doesn't go past 2002 :) |
| 21:12 | <Yuhong> | That is because W3C was screwing up with XHTML2. |
| 21:13 | timeless | shrugs |
| 21:13 | <MikeSmith> | Yuhong: I think the HTML4 support in validator.nu is likely to be dropped eventually |
| 21:14 | <timeless> | i think it's a transitional thing |
| 21:14 | <timeless> | to make it easier for validator.nu to replace the w3 validator |
| 21:14 | <MikeSmith> | yeah |
| 21:14 | <timeless> | when html5 is published, i'd expect them to formally favor html5 |
| 21:14 | <timeless> | and then probably drop it a few months later |
| 21:14 | <timeless> | or at least hide it |
| 21:14 | <Hixie> | html5 was published in like 2007 |
| 21:15 | timeless | shrugs |
| 21:15 | <Velmont> | timeless: irccloud seems a bit strange in their free plan, -- because you can only connect to irc.irccloud.com. I'm at loss as to how they'll get users hooked in the first place then. |
| 21:15 | timeless | doesn't want to get into html5 pub dates |
| 21:15 | <timeless> | Velmont: the beta doesn't enforce the limits |
| 21:15 | <timeless> | which makes addiction *much* easier :) |
| 21:16 | <timeless> | Velmont: everyone knows why drug dealers are successful |
| 21:16 | <timeless> | you give the product away for free :) |
| 21:16 | <timeless> | and i don't mean "you can only smoke it here" |
| 21:16 | <Yuhong> | Yea, my point is that validators do let you validate against different HTML versions. |
| 21:16 | <timeless> | i mean "free" |
| 21:16 | <timeless> | Yuhong: that's a transitional thing |
| 21:16 | <timeless> | but i don't see how that helps you make a request to hixie |
| 21:16 | <timeless> | nor do i understand what the request is |
| 21:17 | <timeless> | if you're actually doing maintenance on a web site |
| 21:17 | <Hixie> | Yuhong: my point is that such features are pointless |
| 21:17 | <timeless> | you want it to work w/ modern browsers |
| 21:17 | <Yuhong> | So why not use the versions produced by W3C HTML WG as targets? |
| 21:17 | <timeless> | and the current spec already works to ensure it doesn't break older browsers (in general) |
| 21:17 | <Hixie> | Yuhong: because there's no point using _any_ fixed reference as a target, as it is by definition going to be out of date |
| 21:17 | <timeless> | and future specs also do that |
| 21:17 | <timeless> | a validator won't tell you if a page/site will work w/ a given browser |
| 21:17 | <Hixie> | Yuhong: whether that's one of the w3c fixed versions or any of the thousands of whatwg fixed versions |
| 21:17 | <timeless> | you'd need to use that actual browser |
| 21:18 | <timeless> | if you have validation errors while you're actually fixing your site, you fix them. |
| 21:18 | <jgraham> | Validating against some old version of a spec doesn't ensure that the page will work either in browsers that were contemporary with the spec or in modern browsers |
| 21:19 | <Hixie> | Validating against some old version of a spec ensures only that you are using known-buggy rules for validatior |
| 21:19 | <Hixie> | validation |
| 21:20 | <Hixie> | it is an entirely futile task imho |
| 21:22 | <timeless> | Yuhong: why does validation interest you in the first place? |
| 21:22 | <timeless> | ... back to Hixie 's original question: What actual problem are you trying to solve? |
| 21:22 | <Yuhong> | OK, I realized it now. |
| 21:25 | <Velmont> | :-) |
| 21:25 | <Yuhong> | And consolidating features into versions is good for buzzwords but don't fit reality. |
| 21:25 | <Velmont> | Yuhong: You're fast man. |
| 21:26 | <Velmont> | Normally people use a long time to realize stuff like that. |
| 21:26 | <Yuhong> | I wonder it is time to stop usage of the "HTML5" buzzword. |
| 21:27 | <Yuhong> | I once saw a comic in which |
| 21:27 | <Yuhong> | I once saw a comic in which "HTML5" was defined as what is not in IE8. |
| 21:28 | <Yuhong> | I think I posted a link to it before. |
| 21:28 | <karlcow> | comic sans |
| 21:28 | <Yuhong> | http://msdn.microsoft.com/en-us/library/ff405772(v=vs.85).aspx |
| 21:28 | <Yuhong> | Not the comic, but "IInternet Explorer HTML 4.01 Standards Support Document" |
| 21:30 | <timeless> | i'd like to +1 Velmont 's you're fast note |
| 21:35 | <Yuhong> | AFAIK the use of "HTML5" as a buzzword began before WHATWG declared it a living standard. |
| 21:35 | <Hixie> | it was one of the reasons we gave up on the name :-) |
| 21:38 | <Yuhong> | BTW, on the Quirks Mode "Standard", I think this has a good explaination: http://hsivonen.iki.fi/almost-precedent/ |
| 21:43 | <Yuhong> | So what do you think the "HTML5" parser should be called? |
| 21:43 | <Hixie> | "HTML parser"? :-) |
| 21:44 | <Yuhong> | Except there was also the old one. |
| 21:44 | <Yuhong> | "spec-compliant parser"? |
| 21:44 | <Hixie> | old one? |
| 21:46 | <Yuhong> | I mean, browsers was touting the "HTML5" parser as a feature. |
| 21:46 | <Hixie> | there is in theory not supposed to be much difference between the old parser browsers had and the specced parser |
| 21:46 | <Hixie> | that's the whole point :-) |
| 21:47 | <Yuhong> | The most recent being IE10. |
| 21:52 | <Yuhong> | Most features are easy to refer to without using the "HTML5" term, but... |
| 21:53 | <Hixie> | there's not, in my mind, any reason to refer to "implementing the parser according to the spec" |
| 21:53 | <Hixie> | at least not in marketing documents |
| 21:57 | <Yuhong> | http://meta.slashdot.org/comments.pl?sid=2769161&cid=39610913 |
| 21:57 | <Yuhong> | Notice no use of "HTML5" anywhere. |
| 22:31 | <zewt> | ... can anyone explain why <title> is "required"? so far, nobody's ever been able to give any rationale, and I can't even contrive any reason, heh |
| 22:34 | <edwardbc_> | well, a document should have a title as part of its metadata, so why shouldn't it be there? |
| 22:34 | <zewt> | "should" != "must" |
| 22:35 | <zewt> | http://lists.w3.org/Archives/Public/public-whatwg-archive/2012Jun/0282.html suggests there's some more pressing reason than "titles are nice" |
| 23:43 | <annevk> | data:text/html,<title>test</title> vs data:text/html,<title></title> is interesting |
| 23:44 | <zewt> | afaik empty title usually behaves the same as not specifying any |
