| 00:02 | <othermaciej> | extra sad fact: just the MathML xmnlns declaration is as long as the LaTeX form of the quadratic formula |
| 00:03 | <gavin_> | heh |
| 00:40 | <Dashiva> | othermaciej: It's XML, expecting it to be anything else is madness ;) |
| 00:50 | <Hixie> | i really need to move the focus from use cases to problems |
| 00:51 | <Hixie> | people seem to give much less useful feedback when they think in terms of the term "use case" than the term "problem" |
| 07:02 | <hsivonen> | http://www.bluishcoder.co.nz/2008/03/refactoring-of-firefox-html-video.html |
| 07:27 | <hsivonen> | I'm a bit nervous about how the mapping from the codecs MIME parameter to and from GStreamer, QuickTime and DirectShow will work in practice |
| 12:17 | <hsivonen> | Do we have enough people going to XTech to hold some kind of official/unofficial f2f meeting? |
| 12:34 | <Philip`> | zcorpan_: compatMode/documentMode work pretty differently in the Live DOM Viewer, so I wouldn't trust what it says |
| 12:35 | <zcorpan_> | Philip`: interesting |
| 12:36 | <zcorpan_> | Philip`: i get the same results in the zombie dom viewer. what's different? |
| 12:38 | <Philip`> | zcorpan_: Er, I can't remember :-( |
| 12:38 | <Philip`> | It might have only affected documentMode and not compatMode |
| 12:39 | <zcorpan_> | i can't find any difference with documentMode either |
| 12:39 | <Philip`> | Hmm |
| 12:40 | <zcorpan_> | i haven't tried the actual http header though |
| 12:40 | Philip` | didn't either |
| 12:40 | <Philip`> | I'll have to try it out when I'm at home again |
| 12:40 | <zcorpan_> | k |
| 13:15 | <hsivonen> | why did absent system/public ids in become "" instead of null in the tokenizer tests? |
| 13:17 | <annevk> | that's a bug |
| 13:17 | <hsivonen> | annevk: ok. thanks |
| 13:18 | <hsivonen> | I take it that it's ok to check in fixed test without fixing the python impl |
| 13:18 | <hsivonen> | correct? |
| 13:18 | <hsivonen> | I can't remember what the outcome of the test and html5lib sync discussion was |
| 13:19 | <annevk> | python fails? |
| 13:19 | <annevk> | hmm |
| 13:19 | <annevk> | who did that? |
| 13:19 | <hsivonen> | I don't know if python fails |
| 13:19 | <hsivonen> | I just assume it might if I change the test cases |
| 13:35 | <hsivonen> | is it intentional that if > terminates the system id, the system id becomes the string so far instead of null? |
| 13:38 | hsivonen | finds that http://html5.org/tools/web-apps-tracker?from=1285&to=1286 wasn't annotated as conformance checker and tool-relevant |
| 13:39 | <annevk> | Philip`, did you change the tokenizer tests to use "" instead of null? |
| 13:40 | <hsivonen> | annevk: actually, "" is right in the cases I was looking at |
| 13:41 | <annevk> | oh ok |
| 13:41 | <annevk> | good |
| 13:41 | <Philip`> | annevk: No |
| 13:41 | <Philip`> | http://html5lib.googlecode.com/svn/trunk/testdata/tokenizer/test2.test still seems to use null |
| 13:42 | <annevk> | seems to be a misunderstanding |
| 13:44 | <Philip`> | http://code.google.com/p/html5lib/source/browse - hmm, why are there no files there? |
| 13:45 | <Philip`> | Oh, it's because Google is stupid and can't write HTML that works in browsers |
| 13:47 | <zcorpan_> | hsivonen: that case was brought up on the list before the spec changed wrt > in FPI so i guess it was considered, and hence, intentional |
| 13:48 | <hsivonen> | zcorpan_: ok |
| 13:54 | <zcorpan_> | i'm thinking about doctype roundtripping with innerHTML |
| 13:54 | <zcorpan_> | the current level of roundtripping of quirkyness is no doctype and doctypes with a name other than "html" |
| 13:55 | <zcorpan_> | but emitting FPI and SI if present doesn't roundtrip bogus doctypes anyway |
| 14:19 | <hsivonen> | Hixie: is there a strong perf reason to have a separate character encoding sniffing algorithm instead of running the real tokenizer with a simplified token handler and ISO-8859-1 decoding? |
| 14:26 | <annevk> | is there any browser that supports document.innerHTML to begin with? |
| 14:28 | <zcorpan_> | annevk: not that i know of |
| 14:30 | <hsivonen> | I hope ipv6.google.com is not an indication of a future naming pattern |
| 14:31 | <zcorpan_> | annevk: but i think the genshi and/or kid templating engines implemented it |
| 14:32 | <Philip`> | Sounds like they aren't very confident at IPv4-only clients' ability to ignore AAAA records |
| 14:32 | <Philip`> | which seems entirely sensible for now |
| 14:33 | <Philip`> | s/IPv4-only clients/clients which support IPv6 themselves but are on IPv4-only networks/ |
| 14:43 | <annevk> | "Any UA that prints pages so incorrectly should be considered non-conformist." |
| 18:47 | <sayrer> | annevk, what's the best selector-api test suite? |
| 18:48 | <annevk> | I guess WebKit has public tests |
| 18:48 | <annevk> | apart from that I'm afraid nothing has happened yet |
| 19:20 | <Lachy> | I will be trying to work on the test suite starting next week |
| 19:20 | <Lachy> | for selectors api |
| 19:35 | <turtlethere> | Some disinformation about HTML 5 and mp3 appeared: |
| 19:35 | <turtlethere> | http://www.ddj.com/web-development/206902334?cid=RSSfeed_DDJ_All |
| 19:36 | <turtlethere> | Kevin Carlson explains: |
| 19:36 | <turtlethere> | everyone involved can agree that the audio element should support MP3 |
| 19:36 | <turtlethere> | Is that so? |
| 19:38 | <hsivonen> | turtlethere: not as far as I've noticed |
| 19:38 | <hsivonen> | turtlethere: might be different in two years |
| 19:39 | <svl> | wow, I see _four_ silverlight ads on that page. Sliiight overkill? |
| 19:40 | <svl> | oh, no, five; there's a really small text-only one as well. |
| 19:41 | <hsivonen> | at least the Silverlight ads GIF images and not Flash animations :-) |
| 19:41 | <turtlethere> | Is anybody in position to complain about this to DDJ? |
| 19:42 | <turtlethere> | I have not found any comment feature there |
| 19:42 | <turtlethere> | and I do not think I know enough to contact them by e-mail |
| 19:44 | <turtlethere> | I mean, such false representations can influence public expectations |
| 19:44 | <annevk> | MP3 was never agreed upon fwiw |
| 19:44 | <turtlethere> | and effectively constrain the choice when it is time to make one. |
| 19:44 | <annevk> | At this point it has much the same problems as H264 so... |
| 20:31 | <annevk> | http://www.w3.org/QA/2008/03/semantic_web-telephone_game.html first I read about it |
| 20:54 | <Philip`> | The error message on http://www.xiven.com/weblog/search/results?q=%00 seems quite peculiar |
| 20:58 | <hsivonen> | Philip`: are you trying to bust everyone on the XPhiles list? |
| 21:02 | <Philip`> | hsivonen: No, I only happened to look at that one (since it was near the top of that list) |
| 21:03 | <Philip`> | It looks like many of the others are cheating because they're just static pages, and for the others it'd be too trivial to break them :-p |
| 21:05 | <hsivonen> | bertilo wennergren cheats by outsourcing searches to Google |
| 21:06 | <hsivonen> | adactio has gone text/html |
| 21:06 | hsivonen | gives up |
| 21:07 | <Philip`> | That doesn't stop http://www.bertilow.com/%ef%bf%bf.html |
| 21:08 | <hsivonen> | %ef%bf%bf always triumphs |
| 21:08 | <Philip`> | %01 is just as good |
| 21:09 | <Philip`> | (%00 only works in query strings because Apache always returns 404 if it's in the path name) |
| 21:09 | <Philip`> | Actually, %3C works perfectly well in this case too |
| 21:09 | <Philip`> | http://www.bertilow.com/<script>alert(document.cookie)</script>.html |
| 21:10 | <Philip`> | and that's not XML's fault |
| 21:10 | <annevk> | nice |
| 21:11 | <Philip`> | I think Validator.nu should have a feature that tries putting %00 and %01 and %ef%bf%bf and %3C into every form it can find on a site, and into every part of the address |
| 21:19 | <Philip`> | Hmm, could you use unescaped input bugs to insert <meta name="verify-v1" ...> into a page and take control of Google's settings for the site? |
| 21:20 | <Philip`> | (I'm not sure what pages Google accepts that on, or whether they require it to be in the <head> or something) |
| 22:32 | <gsnedders> | http://realtech.burningbird.net/?s=%ef%bf%bf — yeah U+FFFF always works |
| 23:33 | <Philip`> | Hixie: The Live DOM Viewer is broken and doesn't show attributes except in IE |
| 23:34 | <Philip`> | And don't try blaming me for that :-p |