| 00:18 | <MikeSmith> | paul_irish: you got a couple minutes to chat? |
| 00:23 | <openstandards> | Hi all i'm after some advice concerning screen readers and aria |
| 00:23 | <paul_irish> | MikeSmith: yup. pm away. |
| 00:25 | <openstandards> | I've got a div thats hidden till a user clicks a link and it scrolls into the viewing area however i want it to be hidden to screen readers and picked up via google bot |
| 00:25 | <nimbupani> | openstandards: if you use display: none it should be hidden to screenreaders as well |
| 00:26 | <openstandards> | nimbupani, however from an seo view will it still be seen to the search engines |
| 00:27 | <openstandards> | like whats aria-hidden all about i haven't found many examples of it being used in the wild |
| 00:27 | <nimbupani> | http://searchengineland.com/is-hiding-content-with-display-none-legitimate-seo-13643 |
| 00:29 | <openstandards> | ok cheers |
| 00:30 | <openstandards> | just gonna get some pizza and i'll return to read that link thanks nimbupani |
| 00:30 | <nimbupani> | cheers openstandards |
| 01:21 | <openstandards> | http://realtech.burningbird.net/web/html5/remove-hidden-attribute interesting link i've found |
| 01:37 | <MikeSmith> | <sigh> just now becoming aware of other cases where existence of long-outdated CSS WG drafts in TR space has ended up wasted a lot of some peoples' time, created unnecessary confusion |
| 01:37 | <TabAtkins> | Which ones? |
| 01:37 | <MikeSmith> | all of them, I guess |
| 01:37 | <TabAtkins> | Makes sense. |
| 01:37 | <MikeSmith> | people looking at docs published in 2003 |
| 01:37 | <TabAtkins> | We *really* need to just link to the dev drafts. |
| 01:38 | <MikeSmith> | yeah, please lobby the CSS WG to do more about this |
| 01:38 | <Philip`> | Should do it like the IETF, and delete the TR drafts after six months |
| 01:38 | <MikeSmith> | I have tried myself to make it clear |
| 01:38 | <MikeSmith> | Philip`: wfm |
| 01:38 | <MikeSmith> | or sounds good to me |
| 01:39 | <MikeSmith> | and WD that is more than 6 months old should definitely be removed/replaced |
| 01:40 | <jaket> | great name |
| 01:40 | <jaket> | mike smith |
| 01:40 | <jaket> | mike smith! |
| 01:40 | <MikeSmith> | even CR drafts, we need to some way to indicate whether they are still relevant or not |
| 01:40 | <MikeSmith> | I think some of these CSS drafts are actually CRs |
| 01:41 | TabAtkins | doesnt' even know what TR stands for. |
| 01:41 | <MikeSmith> | but everybody knows they are no longer up to date with implementations and implementation plans |
| 01:41 | <MikeSmith> | well, I should say not everybody knows, clearly -- *we* know, the rest of the world does not |
| 01:41 | <MikeSmith> | TabAtkins: Technical Report |
| 01:41 | <MikeSmith> | antique term of art |
| 01:42 | <MikeSmith> | opaque now, basically |
| 01:44 | <TabAtkins> | Ah, kk. |
| 01:48 | <openstandards> | can't they just add a css class to the page to highlight it after a given time easy enough to do |
| 01:48 | <variable> | Philip`, I prefer having an archive of past ideas - don't like when things like that are deleted |
| 01:49 | <openstandards> | variable, so do I which is why they should just make it known its a previous idea and no longer in use |
| 01:51 | <othermaciej> | technically, the W3C produces Technical Reports, not "standards" |
| 01:51 | TabAtkins | doesn't know how easy it is to alter a TR document. |
| 01:52 | <othermaciej> | in general, what a TR url points to is easy to replace (that's the standard process) but are rarely modified in place |
| 01:53 | <othermaciej> | each time a WG publishes a draft it goes to TR |
| 01:53 | <othermaciej> | in theory the heartbeat requirement should keep non-REC items in TR space from getting too stale |
| 01:53 | <othermaciej> | but it doesn't seem to be very consistently followed |
| 01:55 | <MikeSmith> | othermaciej: yeah, about "standards", that is technically true on a "how many angels can you fit on the head of pin" definition of "technically" |
| 01:55 | <TabAtkins> | ...I don't think the CSSWG has *ever* done a heartbeat thing while I've followed it. |
| 01:55 | <MikeSmith> | ISO and ITU want to own the term "standards" |
| 01:56 | <MikeSmith> | and national standards bodies actually do refer to specs from W3C and others as "de facto" standards rather than "de jure" standards |
| 02:00 | <MikeSmith> | but for the actual users of our specs, they are very much standards |
| 02:00 | <MikeSmith> | or should be at least |
| 02:00 | <MikeSmith> | in the sense that they should represent what is actually implemented |
| 02:01 | <TabAtkins> | Yes. |
| 02:01 | <othermaciej> | yeah, I don't care for the fussy technicalities around the word "standard" |
| 02:02 | <othermaciej> | but I imagine that is why other standards bodies invent different terms |
| 02:02 | <othermaciej> | I wonder why IEEE gets to have "standards" |
| 02:02 | <othermaciej> | are they an official national standards body? |
| 02:04 | <MikeSmith> | othermaciej: I guess maybe they are a so-called ISO "PAS submitter" |
| 02:04 | <MikeSmith> | othermaciej: http://isotc.iso.org/livelink/livelink/fetch/-8913189/8913214/8913248/Approved_PAS_Submitters.html?nodeid=2315468&vernum=-2 |
| 02:05 | <MikeSmith> | or maybe not |
| 02:05 | <MikeSmith> | they're not listed there |
| 02:05 | <MikeSmith> | PAS Submitters can have their specs sort of fast-tracked to ISO standard status, as I understand it |
| 02:30 | <MikeSmith> | <big sigh> http://code.google.com/p/epub-revision/wiki/CSS3Relations#D3:_EPUB_specific_HTML_classes |
| 02:30 | <MikeSmith> | kennyluck: please educate them |
| 02:30 | <MikeSmith> | "Define class names" |
| 02:33 | <kennyluck> | I'd say WHATWG is very successful but IDPF is... |
| 04:15 | <jacobolus> | does anyone in here know about the state of optimization of SVGPointList in browsers? |
| 04:16 | <jacobolus> | it strikes me that it might be possible to use it, completely outside the context of SVG, to do fast array math in browser javascript |
| 04:16 | <jacobolus> | but maybe the overhead of setting the object up would make it not worth the trouble most of the time |
| 04:18 | <jacobolus> | oh, actually, nevermind. I thought maybe matrixTransform could be used on a whole SVGPointList |
| 04:18 | <jacobolus> | but apparently it only works on one point at a time |
| 04:21 | <jacobolus> | it's really too bad javascript doesn't have some better data types, and some array math/crypto/compression/etc. APIs that can run at native speed |
| 07:16 | <xfinx> | ogguh |
| 07:43 | <MikeSmith> | heh, "Burauzabenda" |
| 07:44 | <MikeSmith> | = Japanese transliteration, in roman alphabet, of "browser vendor" |
| 08:23 | <_mthz> | anyone know anything about the html5lib tests, specifically for the tokenizer? i have some questions about some tests that actually look like they're not in accordance with the spec |
| 08:31 | <MikeSmith> | _mthz: if you wait around a bit until others in central Europe time come on, there are several people who could probably respond |
| 08:32 | <MikeSmith> | jgraham for one |
| 08:32 | <MikeSmith> | hsivonen too, if you can get his attention |
| 08:32 | <_mthz> | thanks! |
| 08:32 | <MikeSmith> | _mthz: are you writing a tokenizer? |
| 08:32 | <_mthz> | yes |
| 08:32 | <MikeSmith> | cool |
| 08:33 | <hsivonen> | _mthz: what's not in accordance? |
| 08:33 | <_mthz> | ah =) |
| 08:33 | <_mthz> | one sec |
| 08:33 | <_mthz> | xmlviolationtes -- Non-XML Character |
| 08:33 | <_mthz> | it excpects \uFFFF to be replaced with \uFFFD |
| 08:33 | <hsivonen> | _mthz: those test XML infoset coercion |
| 08:34 | <_mthz> | ahhhh |
| 08:34 | <_mthz> | that makes sense |
| 08:34 | <hsivonen> | _mthz: the tests in that file don't apply to the vanilla spec |
| 08:34 | <_mthz> | ok another bucket is parsing named entities in attributes |
| 08:34 | <_mthz> | an example being Entity in attribute without semicolon ending in 1 |
| 08:35 | <hsivonen> | that's tricky, but I'm pretty sure those tests are right |
| 08:35 | <_mthz> | it expects an error for <h a='¬1'> but the spec says this is an error only if the ampersand is followed by at least one alphanumeric AND a semicolon |
| 08:35 | <_mthz> | If no match can be made, then no characters are consumed, and nothing is returned. In this case, if the characters after the U+0026 AMPERSAND character (&) consist of a sequence of one or more characters in the range U+0030 DIGIT ZERO (0) to U+0039 DIGIT NINE (9), U+0061 LATIN SMALL LETTER A to U+007A LATIN SMALL LETTER Z, and U+0041 LATIN CAPITAL LETTER A to U+005A LATIN CAPITAL LETTER Z,... |
| 08:35 | <_mthz> | ...followed by a U+003B SEMICOLON character (;), then this is a parse error. |
| 08:37 | <hsivonen> | _mthz: the para begins with "If no match can be made" |
| 08:37 | <hsivonen> | _mthz: here a match *is* made |
| 08:37 | <hsivonen> | for "not" |
| 08:37 | <_mthz> | ahh! |
| 08:37 | <_mthz> | i thought it meant no "terminal" match |
| 08:37 | <hsivonen> | nope |
| 08:38 | <hsivonen> | named character references do longest prefix match |
| 08:38 | <hsivonen> | that's what makes them "fun" |
| 08:38 | <_mthz> | gotcha =) |
| 08:39 | <_mthz> | thank you very much -- that test suite is very useful! |
| 08:39 | <_mthz> | one one last question -- the other set of tests failing for me deal with comment parsing |
| 08:40 | <_mthz> | but it looks like i see that you updated those for a pre-emptive spec change |
| 08:40 | <_mthz> | any idea if/when that change should go through? |
| 08:40 | <hsivonen> | fwiw, I think getting the ¬ and ¬in stuff right *and* linearlishly performant is the hardest part in the tokenizer |
| 08:40 | <hsivonen> | _mthz: after Hixie's vacation |
| 08:40 | <hsivonen> | _mthz: 2 to 4 weeks is my guess |
| 08:41 | <_mthz> | i use a trie structure to store the named entities so its pretty efficient i think |
| 08:43 | <_mthz> | i think the most unfortunate thing is the tokenizer's ties to the parsing for switching the content model |
| 08:43 | <_mthz> | parsing=parser |
| 08:48 | <_mthz> | wait a second |
| 08:49 | <_mthz> | to go back to the <h a='¬1'>example... |
| 08:49 | <_mthz> | that test expects an error but it shouldn't have one |
| 08:49 | <_mthz> | If the character reference is being consumed as part of an attribute, and the last character matched is not a U+003B SEMICOLON character (;), and the next character is either a U+003D EQUALS SIGN character (=) or in the range U+0030 DIGIT ZERO (0) to U+0039 DIGIT NINE (9), U+0041 LATIN CAPITAL LETTER A to U+005A LATIN CAPITAL LETTER Z, or U+0061 LATIN SMALL LETTER A to U+007A LATIN SMALL... |
| 08:49 | <_mthz> | ...LETTER Z, then, for historical reasons, all the characters that were matched after the U+0026 AMPERSAND character (&) must be unconsumed, and nothing is returned |
| 08:50 | <_mthz> | it doesn't say that case is an error? |
| 08:51 | <annevk> | doesn't it say earlier it's an error? |
| 08:51 | <_mthz> | no -- it says it's an error if no match was found |
| 08:51 | <_mthz> | but we matched ¬ |
| 08:52 | <annevk> | oh right; I think we did that to not make & an error for hyperlinks |
| 08:54 | <_mthz> | i wonder if the spec shuldn't be adjusted to have special handling for href/src or something that likely has a URL in it? -- I noticed this deficiency myself awhile ago before the '=' was included in the list of special characters |
| 09:01 | <_mthz> | ahh i see my original question now. you have a bunch of "bad named entities" tests for things ike "&alpha" that expect errors, but again, these aren't errors as per the spec, since for it to be an error you need to hve a semicolon? |
| 09:02 | <annevk> | zcorpan_ knows more about this than I do |
| 09:03 | <_mthz> | hrm.. what would be the appropriate mailing list to send this sort of question to? |
| 09:03 | <zcorpan_> | what's the question? |
| 09:05 | <_mthz> | zcorpan: there are a bunch of tests in html5lib that test bad named entities (like "&alpha") -- these tests expect an error, but my reading of the spec says that a bad named entity is only an error in the case in which the ampersand is followed by at least 1 alphanumeric character AND a semi-colon |
| 09:05 | <_mthz> | but the tests don't have a semi-colon, so they shouldn't be errors? |
| 09:06 | <zcorpan_> | the spec changed wrt this a while ago (at the same time as = was added to the list of special chars), maybe the tests haven't been updated |
| 09:06 | <_mthz> | gotcha |
| 09:08 | <_mthz> | the second question would be tests like "<h a='¬1'>" -- these also expect an error, but there shouldn't be one since the last matched character is followed by an alphanumeric |
| 09:09 | <_mthz> | same issue -- tests not updated? |
| 09:09 | <zcorpan_> | probably |
| 09:10 | <_mthz> | what's the best way for me to log a bug or get them updated? |
| 09:11 | <zcorpan_> | i guess http://code.google.com/p/html5lib/issues/list |
| 09:11 | <zcorpan_> | or you could fix them yourself |
| 09:35 | <zcorpan_> | ericsson are implementing <device>? |
| 09:36 | <Peter`> | Yes |
| 09:36 | <Peter`> | experimenting with webkit |
| 09:36 | <zcorpan_> | cool |
| 09:36 | <hsivonen> | on which OS? |
| 09:37 | <Peter`> | I've only seen examples for Mac OS X |
| 09:37 | <Peter`> | https://labs.ericsson.com/blog/beyond-html5-conversational-voice-and-video-implemented-webkit-gtk |
| 09:38 | <hsivonen> | Peter`: looks more like Ubuntu from the screenshots |
| 09:38 | hsivonen | has lost track of which operating systems Ericsson ships |
| 09:38 | <zcorpan_> | i guess <video autoplay muted> would solve their problem |
| 09:38 | <Peter`> | Uh, you're right |
| 09:38 | <hsivonen> | something Android and something with some version of Symbian on some level of the stack at least |
| 09:39 | hsivonen | wonders what Symbian is in the latest Ericsson Symbian devices and what the userland is (UIQ/Qt/S60/other?) |
| 09:40 | hsivonen | also wonders if Spotify on phones uses Vorbis or AAC |
| 09:44 | annevk | is wondering who'll be first with P2P |
| 09:45 | <annevk> | also http://code.google.com/p/chromium/issues/detail?id=55377 |
| 09:47 | <jgraham> | What's the state of postMessage support in browsers? |
| 09:47 | <annevk> | magical "/" is not in everywhere but other than that it should be good, I think |
| 09:48 | <jgraham> | Which IE version? |
| 09:48 | <annevk> | 8? |
| 09:48 | <annevk> | I forgot |
| 09:48 | <annevk> | seems like 8 per http://msdn.microsoft.com/en-us/library/cc197015(VS.85).aspx |
| 09:48 | <annevk> | but limited to strings |
| 09:49 | <zcorpan_> | firefox doesn't seem to support MessageChannel or message ports |
| 09:49 | <jgraham> | It seems like it could be used in testharness.js to communicate the test results. Dunno if the benefits are worth the costs though |
| 09:49 | <annevk> | no message ports? |
| 09:49 | <annevk> | how did they do workers then? |
| 09:51 | <zcorpan_> | they also don't support shared workers |
| 09:51 | <othermaciej> | the hybi list makes me sad |
| 09:52 | <jgraham> | othermaciej: Why? I stopped paying attention when people started talking about compression |
| 09:53 | <jgraham> | Unless the lack of focus is the problem that makes you sad of course :) |
| 09:53 | <othermaciej> | versioning, featureitis, poor signal-to-noise, some people who are clearly just straight up trolling |
| 09:53 | <annevk> | i have tuned out of hybi |
| 09:54 | <othermaciej> | ietf-standard "this is more important than mere web browsers" attitude |
| 09:54 | <annevk> | i believe we are shipping -76; guess we'll see what happens |
| 09:54 | <zcorpan_> | webkit seems to have the old order of the arguments for postMessage |
| 09:54 | <jgraham> | othermaciej: It would help if Ian Fette was more active, I think |
| 09:54 | <annevk> | the benefit of going to IETF was getting Hixie to draft saner framing (that nobody has implemented) the drawback is everything else |
| 09:55 | <annevk> | imo |
| 09:55 | <annevk> | taking way too long |
| 09:55 | <jgraham> | I don't know if hew is on holiday or what, but without someone actually editing the draft there is no direction |
| 09:55 | <zcorpan_> | he was on holiday last week apparently |
| 09:56 | <zcorpan_> | but should be back this week |
| 09:56 | <othermaciej> | he posted an update at some point |
| 09:57 | <jgraham> | 9/9 |
| 09:59 | <othermaciej> | new draft < 1 week ago is not so bad, though I guess it went for a while without |
| 10:00 | <othermaciej> | I'm more concerned about the contents of that draft than its timing |
| 10:00 | <jgraham> | I guess |
| 10:00 | <jgraham> | But no one is really discussing that |
| 10:02 | <othermaciej> | I guess 9/1 was the last actual draft posted |
| 10:19 | <hsivonen> | othermaciej: what HTML5 parsing difference from old WebKit is breaking mail apps with system WebKit? |
| 10:19 | <othermaciej> | hsivonen: <foo<foo> |
| 10:20 | <hsivonen> | othermaciej: how does Outlook deal? |
| 10:20 | <othermaciej> | I don't know |
| 10:20 | <othermaciej> | I'm not 100% sure of the origin of the emails containing this error |
| 10:20 | <hsivonen> | does Outlook use the Word engine these days? does Word parse differently from Trident? |
| 10:20 | <othermaciej> | but apparently a bunch of people already have them |
| 10:20 | <hsivonen> | othermaciej: It would be very good to have the data |
| 10:20 | <othermaciej> | I believe Outlook uses the Word engine |
| 10:21 | <othermaciej> | we could try to test the relevant emails in outlook or entourage or other mail clients |
| 10:21 | <hsivonen> | adding app-specific quirks seems very sad compared to taking this as data to uphold the old Gecko/WebKit behavior in HTML5 |
| 10:21 | <hsivonen> | esp. since the old Gecko/WebKit behavior is nicer for HTML editors that want to fix user-typed markup |
| 10:22 | <hsivonen> | othermaciej: it would good to know if the emails are generated by an email app in the wild or if they are hand-crafted advertisements |
| 10:22 | <othermaciej> | I think we made an effort to argue for the old Gecko/WebKit behavior on this, based on some actual public site breakage, but yes, this issue also hits native apps and some intranet sites |
| 10:22 | <othermaciej> | I believe at least some of them were produced by an automated reporting system of some kind |
| 10:23 | <hsivonen> | othermaciej: was Hixie's counter-argument based on fairness towards IE, based on data or based on something else_ |
| 10:23 | <hsivonen> | ? |
| 10:24 | <othermaciej> | based on data that showed either way causes problems |
| 10:24 | <othermaciej> | but I don't think the details of that data were cited |
| 10:24 | <hsivonen> | othermaciej: OK |
| 10:25 | <othermaciej> | I have a vague feeling that IE might face the same issues in reverse from dependent apps or intranet sites, but intuitively it seems like the IE/Opera behavior is harder to depend on, since it will tend to produce a result that looks wrong |
| 10:25 | <annevk> | when I looked into this a long time ago we had issues either way |
| 10:25 | <hsivonen> | annevk: wouldn't it then make sense to pick the solution that sucks less for editors? |
| 10:26 | <hsivonen> | now we've picked the solution that makes the tokenizer code simpler |
| 10:27 | <annevk> | I'm not sure what makes sense; I guess I should try to find the relevant bug reports if we are serious about changing this |
| 10:27 | <annevk> | seems kind of a fundamental change but I guess it is worth looking into considering the breakage it causes for WebKit |
| 10:31 | <hsivonen> | annevk: it would be great if you could share the data that made you go the IE way |
| 10:35 | <annevk> | hsivonen, hard to find anything at all really |
| 10:36 | <hsivonen> | annevk: interesting! |
| 10:38 | <annevk> | well, my searching fu for bug systems is limited |
| 10:40 | <zcorpan_> | http://html5.org/tools/web-apps-tracker?from=900&to=902 |
| 10:41 | <jgraham> | annevk: That is a generic problem :) |
| 10:42 | <zcorpan_> | http://html5.org/tools/web-apps-tracker?from=899&to=902 even |
| 10:48 | <zcorpan_> | http://lists.whatwg.org/pipermail/whatwg-whatwg.org/2007-June/011804.html |
| 10:49 | <zcorpan_> | iirc, it was a web compat requirement to not close script for </script<div> |
| 10:50 | <othermaciej> | annevk: it's by far the top source of breakage for us (other than just plain implementation bugs, which are mostly now fixed) |
| 10:51 | <annevk> | oh lol, autofocus became a tracker issue |
| 10:53 | <erlehmann> | „Word engine“ WTF AM I READIN |
| 10:58 | <zcorpan_> | http://www.gearthblog.com/blog/archives/2006/06/more_detail_on.html has </ul <div> (and looks broken with html5 parser) |
| 10:58 | <annevk> | othermaciej, hsivonen, zcorpan_, I am unable to find the relevant bug; I found a bug where we implemented the HTML5 behavior but the reasons for doing so were not very clear, maybe cleanup; and I found a bug that supported the old WebKit/Gecko behavior |
| 10:58 | <annevk> | (for which we got the site fixed) |
| 10:58 | <othermaciej> | annevk: I am done spamming the tracker now |
| 10:58 | <othermaciej> | I had an issue party in my pants |
| 10:59 | <jgraham> | wtf? |
| 10:59 | <Philip`> | http://philip.html5.org/data/gt-in-tag.txt has <foo<foo>s in case anyone is looking for those |
| 10:59 | <Philip`> | (Ignore the filename, it lies) |
| 10:59 | <othermaciej> | annevk: interesting - maybe preponderance of the evidence would actually favor the WebKit/Gecko behavior |
| 11:00 | <annevk> | I seem to recall there was more to it, but this was back in 2008 |
| 11:00 | <annevk> | othermaciej, I am starting to think that; it would be nice if I was able to find the other bugs too though |
| 11:00 | <othermaciej> | I'm willing to take another pass at requesting a change on this but I'd need help putting together all the data |
| 11:01 | <othermaciej> | I can try to dig up the issues we specifically ran into for WebKit but I am not familiar with whatever breakage Gecko hit and I can't commit to do data analysis on any large data sets |
| 11:02 | <erlehmann> | there should be google code search on all web sites indexed |
| 11:02 | <erlehmann> | so you could easily spot how often some construct occurs |
| 11:03 | <othermaciej> | that would be way cool |
| 11:03 | <erlehmann> | hixie probably has it already |
| 11:03 | <erlehmann> | %) |
| 11:09 | <annevk> | I'm not opposed to change the parser. So far I only found data for the old Gecko/WebKit way... Of course actually going through the data Philip` provided would give a more clear answer. |
| 11:11 | <othermaciej> | yeah, I'm just not sure how to determine which parsing behavior is "better" short of visual inspection with both kinds of parsing |
| 11:11 | <othermaciej> | which seems slow |
| 11:12 | <annevk> | e.g. for http://pageranking.cbgw-lensahn-slh.de/ HTML5 is better |
| 11:12 | <annevk> | they use markup like this: |
| 11:13 | <annevk> | <a class="footer" <a href="..." target="_blank"> |
| 11:14 | <annevk> | but yeah, inspecting these would take a lot of time |
| 11:15 | <othermaciej> | the I think the issue I am talking about might be specifically <foo</foo>, despite my statement above |
| 11:16 | <annevk> | so < in the tag name state? |
| 11:16 | <othermaciej> | or at least the cases observed to be a problem |
| 11:16 | <annevk> | Philip`, any chance you can filter your data on that? |
| 11:16 | <othermaciej> | here's a particular example: |
| 11:16 | <othermaciej> | some bad content had this in it: <style type='text/css'td{width='60%' cellpadding='20%'}</style> |
| 11:17 | <othermaciej> | which ate the rest of the page instead of making an empty style element with some bogus attributes |
| 11:17 | <annevk> | Philip`, i.e. starts with < or </ then a-Z then < |
| 11:17 | <annevk> | one or more a-Z |
| 11:17 | <annevk> | oh wait |
| 11:18 | <othermaciej> | I don't think that matches what I'm describing |
| 11:18 | <annevk> | no not at all |
| 11:18 | <othermaciej> | what I'm saying is that in the bad cases we've observed, the broken open tag is followed by a close tag |
| 11:18 | <annevk> | you mean </ anywhere in the tag name |
| 11:18 | <othermaciej> | rather than another open tag |
| 11:18 | <annevk> | euh, tag |
| 11:18 | <othermaciej> | I'm not 100% sure that's all of them, but yeah |
| 11:19 | <othermaciej> | actually I am wrong, that is not all the bad cases |
| 11:19 | <othermaciej> | another was <td class="popbox" <input type="checkbox" /> |
| 11:21 | <annevk> | that directly clashes with the one above |
| 11:21 | <annevk> | so breakage either way :/ |
| 11:22 | <Philip`> | annevk: You should be able to just grep the .txt file, if you want to try various filters |
| 11:23 | <othermaciej> | in what way does the <a> one actually break? |
| 11:23 | <annevk> | styling is not applied |
| 11:23 | <annevk> | because it lacks class=footer |
| 11:23 | <othermaciej> | ok |
| 11:23 | <othermaciej> | I saw that some links looked blue instead of orange when I compared |
| 11:23 | <annevk> | Philip`, I guess, but it's no longer needed |
| 11:27 | <annevk> | going a bit more through it quite often it does not seem to matter either way |
| 11:30 | <jgraham> | You need to weight by badness of the problem of course |
| 11:31 | <jgraham> | Like eating the whole page on a few pages is worse than slight issues on more pages |
| 11:33 | <annevk> | need to do a lot of things |
| 11:34 | <annevk> | WHATWG, great nick |
| 11:35 | <hsivonen> | did someone file a spec bug about <foo<bar> already? |
| 11:35 | <annevk> | anyway, I guess I'm good either way; in any event do not let Opera be the reason for not changing this |
| 11:37 | <othermaciej> | hsivonen: I think abarth did and it was declined (not 100% sure though) |
| 11:38 | <zcorpan_> | if we change this, we need to investigate carefully what to change to. old webkit and gecko don't agree in all cases (iirc) and they don't make </script<div> close the script, iirc |
| 11:38 | <hsivonen> | what did Opera do in 2006? |
| 11:39 | <hsivonen> | it would suck to use circular reasoning to make HTML5 do something, because of Opera if Opera changed to match HTML5 |
| 11:39 | <zcorpan_> | i think we did the same as we do now |
| 11:39 | <hsivonen> | zcorpan_: I'd be willing to special-case "script" in the tag name, even |
| 11:39 | <hsivonen> | zcorpan_: OK |
| 11:39 | <annevk> | zcorpan_, no we didn't |
| 11:39 | <zcorpan_> | annevk: you sure? |
| 11:39 | <annevk> | pretty sure yes |
| 11:40 | <annevk> | see pm |
| 11:42 | <zcorpan_> | oh, previously we parsed <p<div> as <p <div=""> i.e. with an attribute "<div" |
| 11:44 | <zcorpan_> | so we were still closer to ie than gecko and webkit for both <p<div> and <p <div> |
| 11:46 | <zcorpan_> | we fixed that in 2008 to match ie and html5 |
| 11:51 | <hsivonen> | https://bugzilla.mozilla.org/show_bug.cgi?id=507498 |
| 11:52 | <hsivonen> | https://bugzilla.mozilla.org/show_bug.cgi?id=510252 |
| 11:52 | <hsivonen> | https://bugzilla.mozilla.org/show_bug.cgi?id=523516 |
| 11:52 | <hsivonen> | https://bugzilla.mozilla.org/show_bug.cgi?id=543652 |
| 11:53 | <hsivonen> | https://bugzilla.mozilla.org/show_bug.cgi?id=590416 |
| 11:57 | <phrearch> | hi |
| 11:58 | <phrearch> | does anyone know how to determin the boundary in a Content-Type header for a multipart form post? |
| 12:03 | <zcorpan_> | also see http://lists.whatwg.org/pipermail/whatwg-whatwg.org/2007-June/011891.html |
| 12:07 | <annevk> | http://w3.markmail.org/search/?q= public-html is getting closer to www-style |
| 12:09 | <annevk> | hmm, public-webapps is not even in the full list |
| 12:09 | <annevk> | not indexed at all? :/ |
| 12:10 | <annevk> | oh, they call it public-webapi |
| 12:10 | <annevk> | and public-appformats |
| 12:10 | <annevk> | so all that is duplicated? |
| 12:30 | <hsivonen> | fwiw, the bug is http://www.w3.org/Bugs/Public/show_bug.cgi?id=9985 in case anyone wants to add more data or CC self |
| 12:36 | <zcorpan_> | hsivonen: maybe you should reopen the bug for Hixie to see it |
| 12:40 | <hsivonen> | zcorpan_: OK |
| 14:16 | <MikeSmith> | I think I've asked this question before, but I'll ask again -- is there any way to CSS selectors to express, e.g, "a <foo class=note> that is NOT followed by another <foo class=note>"? |
| 14:17 | <MikeSmith> | use case is: If I have one or more class=note instances in a row, as siblings, then I want to number the first one NOTE 1, the second NOTE 2, etc. |
| 14:17 | <MikeSmith> | but if I have only one class=note without any following class=note siblings, then I want to nub |
| 14:18 | <Rik`> | MikeSmith: afaict you can't |
| 14:18 | <MikeSmith> | *want to omit the number and just have NOTE |
| 14:18 | <MikeSmith> | Rik`: OK |
| 14:18 | <MikeSmith> | was hoping there's some magic I'm missing |
| 14:19 | <Rik`> | MikeSmith: maybe with something like ".note:first-child:last-child" |
| 14:19 | <Rik`> | if all your .note are inside an element |
| 14:19 | <MikeSmith> | they aren't, unfortunately |
| 14:19 | <MikeSmith> | it's a very flat structure I'm working with |
| 14:20 | <MikeSmith> | output converted from MS Word source |
| 14:26 | <erlehmann> | MikeSmith, format foo.note for something. format foo.note + foo.note for something else? |
| 14:26 | <erlehmann> | hmm |
| 14:26 | <erlehmann> | not |
| 14:26 | <erlehmann> | oh well |
| 14:27 | <MikeSmith> | erlehmann: thanks, but the hope is for me to not have to add additional markup for N different notes |
| 14:28 | <MikeSmith> | the document has dozens of these |
| 14:28 | <erlehmann> | MikeSmith, i would suggest you fix your ms word then. |
| 14:28 | <erlehmann> | like, dig out the source code of that export tool or something. |
| 14:29 | <erlehmann> | or use XSLT on the source. it is, like, black magic, only with XML. |
| 14:29 | <erlehmann> | disclaimer: i once wrote a CMS in XSLT. i know a guy who uses a CMS consisting mainly of XSLT. and makes money with it. sadly, it is not me. |
| 14:30 | <jgraham> | erlehmann: How much money would you have to make for a life of maintaining XSLT to be worthwhile? |
| 14:31 | <jgraham> | MikeSmith: Wold it be too optimistic to hope for the ECMAScript spec? |
| 14:32 | <MikeSmith> | erlehmann: thanks for the suggestion, I'll file it away with the many other suggestions for things that I can't do anything about |
| 14:32 | <MikeSmith> | jgraham: hope? how so? |
| 14:33 | <MikeSmith> | hope for it to be published, or …? |
| 14:33 | <jgraham> | MikeSmith: Hope that you are converting it to HTML :) |
| 14:33 | <MikeSmith> | oh |
| 14:33 | <MikeSmith> | I have done most of the conversion already |
| 14:33 | <erlehmann> | jgraham, that would depend on N squared, where N is the number of dom nodes. |
| 14:34 | <erlehmann> | MikeSmith, you cannot into XSLT? |
| 14:34 | <Rik`> | MikeSmith: maybe the safest way to go is to use javascript to add a class on single notes |
| 14:34 | <MikeSmith> | jgraham; what remains is, I now need to add consistent IDs to replace the generated ones that are in the source |
| 14:34 | <MikeSmith> | erlehmann: using XSLT for this would be one of those "now you have two problems" cases |
| 14:35 | <MikeSmith> | for one thing, the document is HTML, not XHTML |
| 14:35 | <erlehmann> | harr harr |
| 14:35 | <Philip`> | That's easy, just use XSLT4HTML5 |
| 14:35 | <jgraham> | MikeSmith: You have crossreferenced all the headings and instances of terms? |
| 14:35 | <erlehmann> | first i was like :D |
| 14:35 | <erlehmann> | but then i D: |
| 14:35 | <jgraham> | Because that is what I really crave |
| 14:37 | <jgraham> | e.g. I really really want to be able to click on [[HasOwnProperty]] and be taken to the right section |
| 14:37 | <MikeSmith> | Rik`: yeah, that is what I am thinking too |
| 14:37 | <MikeSmith> | and what I will probably do (use JS) |
| 14:38 | <MikeSmith> | jgraham: I've got it at github and welcome collaborators once I've got this initial conversion done |
| 14:38 | <MikeSmith> | cross-referencing headings I can probably do trivially |
| 14:39 | <MikeSmith> | because the spec uses, e.g., "see 15.5.6.3" pretty consisently |
| 14:39 | <MikeSmith> | terms are not marked up in the source |
| 14:40 | <MikeSmith> | so that would be pretty tough |
| 14:41 | <jgraham> | Yeah :( |
| 14:41 | <MikeSmith> | Norm Walsh calls this "dragging unstructured markup uphill" |
| 14:42 | <MikeSmith> | anyway, for ES6 I will personally volunteer to help them create the source for the spec using something better |
| 14:42 | <MikeSmith> | and will make sure that they publish and HTML version |
| 14:43 | <jgraham> | Yeah, unless you can convince the committee that microsoft word is a dumb format, it will be like that thing where you drag the stone up hill, and then it rolls back down again, for eternity |
| 14:43 | <MikeSmith> | (to the degree that I am able to get the TC39 folks to do it, at least) |
| 14:43 | <MikeSmith> | yeah |
| 14:44 | <annevk> | MikeSmith, you can select all subsequent notes using ~ |
| 14:45 | <MikeSmith> | annevk: what I need is for the case where there are no subsequent notes to be styled differently from the case where there are |
| 14:46 | <annevk> | aah |
| 15:54 | <erlehmann> | MikeSmith, jgraham, how come they use word at all? |
| 15:55 | <MikeSmith> | erlehmann: dunno, I guess it may be the common format among ecma TCs |
| 15:55 | <MikeSmith> | it's not uncommon in other SDOs at all |
| 15:56 | <MikeSmith> | ITU, for example |
| 15:56 | <erlehmann> | ECMA could at least use Office Open XML, harr harr |
| 15:57 | <MikeSmith> | I suppose ECMA has a lot of committees, most of which have nothing to do with the Web |
| 15:57 | <MikeSmith> | and they are organizationally probably not super Web-savvy |
| 15:57 | <jgraham> | erlehmann: I assumed it was because the editor works for Microsoft and because they like the change marking features |
| 15:58 | <jgraham> | That is entirely a guess though |
| 15:58 | <MikeSmith> | http://www.ecma-international.org/publications/standards/Standard.htm |
| 16:00 | <erlehmann> | jgraham, the first part sounds reasonable. the second, in the year 5 after git? the year 30something after DIFF? |
| 16:02 | <jgraham> | erlehmann: It's not like diff produces nice output for flowed text documents |
| 16:02 | <jgraham> | For code it is OK |
| 16:02 | <jgraham> | But I don't recall seeing anything that will let me go from diff to a word-like presentation of bits deleted and bits inserted |
| 16:03 | jgraham | can't believe he is defending word here |
| 16:04 | <jgraham> | For the record: I hate word. But I have heard that people in publishing like the change control functionality |
| 16:09 | <TabAtkins> | MikeSmith: Rik` is right - CSS doesn't allow you to express a selector that depends on information later in the document; that is, on information in descendants or later siblings. |
| 16:09 | <MikeSmith> | TabAtkins: ok |
| 16:09 | <Rik`> | and that's sad :( |
| 16:09 | <TabAtkins> | There's very good reasons for it, but yeah, it's kinda annoying. |
| 16:10 | <MikeSmith> | there's tradeoffs to everything |
| 16:10 | <MikeSmith> | such is life |
| 16:12 | <Rik`> | I should send my first mail to www-style about input:required::after { content: "*" } |
| 16:13 | <TabAtkins> | I can already tell you the response. |
| 16:13 | <TabAtkins> | input is a replaced element, so ::before and ::after are undefined for it currently. |
| 16:14 | <Rik`> | I know but there is a usecase to solve |
| 16:14 | <TabAtkins> | Agreed; there's just no solution currently. There are only two ways to change it: |
| 16:14 | <TabAtkins> | 1) Make <input> not a replaced element (probably not going to work). |
| 16:15 | <Rik`> | one of my ideas was involving CSS to not target the latest element of a selector |
| 16:16 | <TabAtkins> | 2) Actually implement ::outside, which *is* cool for replaced elements, so you can use input:required::outside::after (that is, generated an after pseudo for the outside pseudo). |
| 16:16 | <Rik`> | (which is what MikeSmith want and why it reminded me of this topic) |
| 16:16 | <erlehmann> | TabAtkins, what would break if replaced elements suddenly get ::before and ::after? |
| 16:16 | <erlehmann> | what is ::outside? o.0 |
| 16:16 | <TabAtkins> | erlehmann: It's not a matter of what woudl break, but rather what it even *means*. The *definition* of a replaced element is "something that CSS doesn't know what's inside". |
| 16:16 | <erlehmann> | oh. |
| 16:17 | <TabAtkins> | erlehmann: Check out the Generated and Replaced Content draft. Years old and unimplemented, but the idea's there. |
| 16:17 | <erlehmann> | <http://www.w3.org/TR/css3-content/>; This, I think? |
| 16:17 | <TabAtkins> | Yeah. |
| 16:18 | <jgraham> | TabAtkins: Well there are *lots* of things with form elements that Should Not Work per CSS but magically do anyway |
| 16:19 | <TabAtkins> | jgraham: Indeed, which is why tantek is working on that now. |
| 16:24 | <Rik`> | TabAtkins: wow, I didn't know about ::outside |
| 16:24 | <Rik`> | is it implemented somewhere ? |
| 16:30 | <TabAtkins> | No, and there are currently no plans to do so anywhere that I know of. |
| 16:49 | <softdrink> | is there a good way to discover the implicit relationship of a <td> to any <th>, <col>, or <colgroup> elements that it 'belongs' to? |
| 16:53 | <TabAtkins> | No easy way, no. You can check pretty easily for any <th> in the same <tr>, but for <col>/<colgroup> you have to do some potentially complex cell-counting (extra-hard because the position of a cell may be affected by colspans in a previous row). |
| 16:54 | <softdrink> | That's the conclusion I'm coming to. Seems there's absolutely no way to do it in css. The "headers" attribute seems super redundant to add just for targeting. |
| 16:55 | <softdrink> | The more I try to use tables for mildly complex data, the less sense the spec makes to me. ;) |
| 16:58 | <Workshiva> | Blame CSS, not HTML :) |
| 16:58 | <jgraham> | softdrink: It *can* be done see http://james.html5.org/tables/table_inspector.html |
| 16:58 | <Philip`> | Pfft, you expect specs to make sense? Are you crazy? |
| 16:58 | <jgraham> | It just can't be done from CSS |
| 16:58 | <jgraham> | or easily from anywhere |
| 16:59 | <softdrink> | right, which is a shame hehe |
| 17:05 | <TabAtkins_> | hsivonen: http://lists.w3.org/Archives/Public/www-style/2010Sep/0440.html |
| 17:07 | <TabAtkins_> | Or anyone else, ^^^, which option in the second part is easiest to implement? |
| 17:09 | <Hixie> | why do the a11y group think the drag and drop API in HTML is inaccessible? |
| 17:10 | <Hixie> | anyone know if they just mean the implementations so far, or is there some problem with the API itself? |
| 17:21 | <annevk> | Hixie, the keyboard bit got dropped from the spec; that might have prompted it |
| 17:22 | <annevk> | Hixie, but really, vacation time :) |
| 18:10 | <zcorpan_> | ie9 beta eh? |
| 18:20 | <jgraham> | They put print abouve file in the menu? Interesting |
| 18:20 | <jgraham> | *above |
| 18:43 | <espadrine> | "the browser is the stage, or backdrop, for the web, and the sites are the star of the show" |
| 18:43 | <espadrine> | That sounds like Google Chrome |
| 18:44 | <zcorpan_> | the ui is a lot better than ie8 |
| 18:44 | <espadrine> | undoubtably |
| 18:46 | <softdrink> | except that i can't seem to have the new tab page just *blank* |
| 18:48 | <hsivonen> | so is there a notification API in IE9? if yes, does it resemble Web Notifications drafts? |
| 18:48 | jgraham | has only seen screenshots |
| 18:48 | <jgraham> | I guess it won't run under wine :) |
| 18:50 | <Philip`> | That would defeat their "we're better than other browsers because we're not cross-platform" claim, so I'd assume not |
| 18:56 | <zcorpan_> | seems only 4 out of their 19 "html5 demos" demo things that are in the html5 spec |
| 18:56 | hsivonen | notes that PP4 run on Windows 7 even with graphics drivers that don't do fancy DirectX |
| 18:56 | <hsivonen> | *runs |
| 18:56 | <Peter-> | They removed their implementation of the flexbox module |
| 18:57 | <Peter-> | as well as parsing of 2D CSS Transforms |
| 18:57 | <hsivonen> | I wonder why 2D transforms were removed |
| 18:57 | <Peter-> | apparently only "filter" is supported as a css property, not "-ms-filter" |
| 19:00 | <hober> | time for my n-times-weekly ping for feedback on http://www.w3.org/html/wg/wiki/User:Eoconnor/ISSUE-41 |
| 19:04 | <annevk> | http://www.beautyoftheweb.com/assets/images/content/logo-html5.png via mattur |
| 19:06 | <hsivonen> | annevk: I wonder if it is authorized |
| 19:10 | <jgraham> | http://www.neowin.net/news/optimizing-your-site-for-internet-explorer-9039s-new-features-part-1 lots of values to register there |
| 19:21 | <espadrine> | Easy to make fun of the design choice: http://static.arstechnica.com/ie-9-beta-1/excessive-tabs.png |
| 19:30 | <espadrine> | The white 5px-tall rectangle above the page feels odd and ugly... |
| 19:30 | <hsivonen> | hmm. I discovered operaunite.com, appspot.com and blogspot.com are already or being added to the public suffix list |
| 19:30 | <hsivonen> | interesting development |
| 19:45 | <erlehmann> | hsivonen, „public suffix list“ ? |
| 19:47 | <erlehmann> | jgraham, msapplication-task tags? can anyone explain what distinguishes that from favicons? |
| 19:47 | <erlehmann> | oh. silly me |
| 19:47 | <Workshiva> | erlehmann: For determining cookie privacy and such |
| 19:47 | <erlehmann> | forget those lines. |
| 19:48 | <Workshiva> | subdomains of example.com can share, but subdomains of co.uk shouldn't be able to |
| 19:50 | <zcorpan_> | hsivonen: interesting indeed |
| 19:50 | <erlehmann> | Workshiva, I get it. Those silly uk TLD issues. |
| 19:52 | <erlehmann> | what is the default behaviour for suffixes if they are not on the list? |
| 19:52 | <erlehmann> | forbid? |
| 19:53 | <Workshiva> | Good question |
| 19:59 | gsnedders | is fairly certain the girls in this flat will shortly set off the fire alarm |
| 20:05 | <Workshiva> | How about stopping them before the fact |
| 20:10 | <gsnedders> | Workshiva: Like they heed my warnings |
| 20:11 | <gsnedders> | (I did say they probably were about to…) |
| 20:11 | gsnedders | is waiting for it |
| 20:15 | <mpilgrim> | opera folk: is it possible to remove/alter the default UI for type="number" fields? As in, I want an <input type="number"> without the up/down arrow buttons |
| 20:16 | <annevk> | nope |
| 20:16 | <bl4ckcomb> | currently you can't style the new inputs |
| 20:17 | <mpilgrim> | searching brought up my own work as the top result, which is never a good sign when you know you don |
| 20:17 | <mpilgrim> | 't know something |
| 20:17 | <mpilgrim> | ok, thanks |
| 20:19 | <mpilgrim> | and with that answer, my inbox is down to 157 |
| 20:22 | <Philip`> | http://blogs.msdn.com/b/ieinternals/archive/2010/09/15/ie9-beta-minor-change-list.aspx - "DataURIs may now be used in SCRIPT SRC attributes. The ACID3 test contains the following example: [...] …although there's no reason that you should do this." |
| 20:22 | <Philip`> | Shouldn't they be supporting data: because it's a widely-implemented standard, not simply special-casing the bits the Acid tests use? |
| 20:23 | <annevk> | mpilgrim, I got a copy btw; still need to take some time to read it :) |
| 20:24 | <mpilgrim> | annevk: copy of what? |
| 20:24 | <Peter-> | Philip`: they don't, there was a bug of it on microsoft connect which got closed with "intended behavior" as its reason |
| 20:25 | <annevk> | mpilgrim, your latest book; you know, on paper |
| 20:26 | <mpilgrim> | oh, don't read that, it's several months out of date |
| 20:26 | <gsnedders> | Paper? People still use that thing? |
| 20:26 | <annevk> | :) |
| 20:26 | <mpilgrim> | gsnedders: even worse, people still pay money for it |
| 20:26 | <mpilgrim> | people baffle me |
| 20:27 | <mpilgrim> | just in general, but also why they give me money in particular |
| 20:30 | <hsivonen> | I haven't seen anything about a JIT for VBScript in IE9. Is VBScript now officially unloved? |
| 20:43 | <AryehGregor> | hsivonen, yes, they said JS is integrated into the browser and VBScript lives outside somehow. |
| 20:44 | <AryehGregor> | There was a blog post about it. |
| 20:44 | <AryehGregor> | (in which they made vague allusions that WebKit was inferior for having multiple JS engines) |
| 20:56 | <hsivonen> | AryehGregor: ok. I had missed that |
| 21:30 | <jgraham> | Should get a copy of mpilgrim's book for the office bookshelf |
| 21:33 | <hober> | mark mailed me a review copy; I feel guilty that I haven't posted a review yet. |
| 21:55 | <karlushi> | Are there Javascript libraries to create sounds? a bit like canvas can create graphics. |
| 22:01 | <jgraham> | karlushi: Only Mozilla have the right kind of API for that atm. I don't know if webapps are standardising it |
| 22:03 | <jgraham> | https://wiki.mozilla.org/Audio_Data_API |
| 22:03 | <jgraham> | I hope they are thaking it to some standards body |
| 22:04 | <karlushi> | jgraham, oh thanks. |
| 22:05 | <karlushi> | it crossed my mind today because of this http://www.vexflow.com/ |
| 23:14 | <Rik`> | karlushi: Firefox has a working Audio API and Chrome is working on it, don't know the exact state |
