| 00:00 | <jgraham> | Hixie: The HTML5 spec seems to kill lxml, which I;m using to construct the tree |
| 00:00 | <Hixie> | ah |
| 00:01 | <jgraham> | (I haven't got as far as making the updating... thing go away if there is an error) |
| 00:01 | <jgraham> | specifically lxml 1.3.6 doesn't like the doctype <!DOCTYPE html PUBLIC "-//W3C//DTD HTML 4.01//EN"> |
| 00:08 | <Hixie> | heh |
| 00:10 | jgraham | fixes the bug in html5lib; will deploy tomorrow |
| 00:10 | eseidel | wonders if jgraham is the author of html5lib |
| 00:15 | <jgraham> | eseidel: I am one of the authors along with annevk and various others |
| 00:18 | <jgraham> | Hixie: It doesn't work quite right on the spec atm |
| 00:19 | <jgraham> | but that's another problem for tomorrow |
| 00:19 | jgraham | -> bed |
| 00:26 | <Hixie> | nn |
| 00:32 | Philip` | looks at some data and discovers surprisingly that the web is a mess |
| 00:33 | <Philip`> | It looks like it's actually about as broken as the rest of the internet |
| 00:55 | <Philip`> | On the positive side, 100% of pages that declare themselves as iso-8859-1 can be decoded with no errors |
| 00:56 | <Philip`> | On the less positive side, only 94% of pages declared as utf-8 can be decoded without errors |
| 00:56 | <Philip`> | and 84% of gb2312 |
| 01:02 | <Philip`> | http://validator.nu/?doc=http://tools.ietf.org/html/rfc303 - some people can't even get us-ascii right |
| 01:39 | <Philip`> | How does MetaSniffer think that pages like http://fqq.com/ have charset "x-windows-874"? |
| 01:40 | <Philip`> | There's 58 of them, and none seem to actually say "x-windows-874" anywhere inside them... |
| 01:43 | <Philip`> | Oh, that's because I'm asking it for the canonical name of the sniffed charset |
| 04:47 | <Hixie_> | these fixes are proving remarkably easy to do |
| 07:05 | <othermaciej> | Hixie: would you consider the pause of a WebKit release build on Acid3 (around test 26) to constitute a failure? |
| 07:06 | <othermaciej> | (I think it pauses less than any other browser but it's still noticeable) |
| 07:09 | <Hixie> | yes, though obviously nowhere as serious as anything else |
| 07:10 | <othermaciej> | I'm not sure it is possible to pass then |
| 07:11 | <othermaciej> | but I guess we'll look at it once we fix everything else |
| 07:11 | <othermaciej> | (up to 90/100 now) |
| 07:53 | <annevk> | Hixie, not, the current spec treats a <select> start tag specially in "in select" and different from <input> |
| 07:56 | <Hixie> | hm? |
| 07:57 | <Hixie> | how do you mean? |
| 07:58 | <annevk> | <select><option>...<select><option>... |
| 07:58 | <hsivonen> | annevk, jgraham: are you fixing the html5lib tests to match Hixie's spec changes? |
| 07:58 | <annevk> | In Firefox and Opera that generates 2 selects |
| 07:58 | <annevk> | In IE and HTML5 it generates 1 |
| 07:59 | <Hixie> | ok... |
| 07:59 | <Hixie> | that's good, no? |
| 07:59 | <annevk> | yeah, maybe it is |
| 08:00 | <annevk> | Hixie, btw, you're still going to fix EOF handling, etc right? |
| 08:00 | <annevk> | and hopefully form parsing |
| 08:01 | <Hixie> | actually nobody mailed about that, but yes, there's a big red note about that in the spec, no? |
| 08:01 | <Hixie> | oh crap did i remove that note? |
| 08:01 | <Hixie> | hm |
| 08:01 | <Hixie> | send mail! |
| 08:01 | <Hixie> | :-) |
| 08:02 | <Hixie> | bbiab |
| 08:03 | <annevk> | i e-mailed about form parsing, i'll complain about eof handling once i've read the rest of your e-mail and the spec changes |
| 08:06 | <Hixie> | k thanks |
| 08:06 | <Hixie> | please do mail me, i'll try to address it relatively soon if you send it tonigth |
| 08:07 | <om_sleep> | Hixie: fwiw, Safari would have to be 3-4x faster to avoid visibly pausing (to my eye) on test 26 |
| 08:07 | <om_sleep> | I hope SJ has some sweet hardware announcements this year |
| 08:11 | <annevk> | in revision 1345 it was removed |
| 08:18 | <annevk> | so the problem with html5lib is that all these other projects rely on it and when I want to fix parser bugs I have to fix their testcases too |
| 08:56 | <Philip`> | annevk: Also the problem is that when I want to fix the test cases so they work in my parser, I have to fix html5lib because it's not nice to leave failing tests |
| 09:04 | <annevk> | i agree that that's not ideal |
| 09:05 | <annevk> | i don't really have good ideas other than to fork html5lib to make it about parsing again |
| 09:06 | <annevk> | Because I did really like having the Python reference implementation |
| 09:06 | <annevk> | "reference" |
| 09:09 | <Philip`> | Maybe it'd help to move the testdata into a separate project, so it can be updated independently of any one implementation, and then occasionally import certain revisions into html5lib once html5lib is made to pass those tests? |
| 09:10 | Philip` | might try looking at the updated parser spec in the next week or so, to see what's changed and to fix tests and things |
| 09:11 | <annevk> | we could put the test stuff in http://code.google.com/p/html5/ |
| 09:56 | <zcorpan> | i remember someone had a master's thesis about mobiles and xhtml |
| 09:57 | <zcorpan> | ah found it |
| 09:58 | annevk | makes some new HTML parsing tests to cover the changes |
| 10:00 | <hsivonen> | zcorpan: out of curiosity, whose thesis was that? hendry's? |
| 10:00 | <zcorpan> | hsivonen: yeah |
| 10:03 | <jgraham> | annevk: I don't see why you would need to fork html5lib to update the tests |
| 10:03 | <hendry> | zcorpan: heh :) |
| 10:03 | <hendry> | not sure if any bits are relevant. I want to rewrite it. ;) |
| 10:03 | <annevk> | jgraham, the problem is all the related projects |
| 10:03 | <jgraham> | My opinion is that if we are broken per spec then there's no problem with fixing the tests and sending mail |
| 10:04 | <jgraham> | annevk: Related to what? |
| 10:04 | <jgraham> | Also note that c.g.c/p/html5lib -> c.g.c/p/html5 involves a small license change |
| 10:05 | <annevk> | true, i don't really want to fork it |
| 10:06 | <hsivonen> | fwiw, I don't mind if the html5lib tests move ahead of the Validator.nu parser |
| 10:06 | <hsivonen> | it won't break my build or anything like that |
| 10:07 | <hsivonen> | actually, I'm kind of hoping that someone other than me fixed the tests :-) |
| 10:08 | <jgraham> | Although it's not quite the tdd philosophy, I'd much rather have some already failing tests to help get the changes right |
| 10:09 | <annevk> | i'll fix the html5lib python parser and the tokenizer and parser tests |
| 10:09 | <annevk> | everything else is screwed |
| 10:09 | jgraham | still isn't sure what constitutes "everything else" here |
| 10:09 | <annevk> | i made a runparsertests.py that only does "test_tokenizer.py", "test_parser.py", "test_parser2.py" |
| 10:10 | <annevk> | and not serializer, validator, etc. |
| 10:16 | <Philip`> | I think the html5lib tokeniser tests are up to date now |
| 10:16 | <Philip`> | though I could have easily missed some bits |
| 10:17 | <hsivonen> | virtuelv: about.validator.nu now has a table of contents |
| 10:18 | <virtuelv> | hsivonen: thanks |
| 11:07 | <annevk> | jgraham, yt? |
| 11:08 | <annevk> | I need some help with the table crap. HTML5 now has this concept of tainted tables where we only have some table insertion magic. Do you (or anyone else) see an easy way to change html5lib to make it aware of the new concept? |
| 11:38 | <hsivonen> | I find it interesting how out=gnu seems to be getting more attention than the fancy XML and JSON outputs |
| 11:40 | <hsivonen> | Hixie: http://validator.nu/?doc=http://www.whatwg.org/specs/web-apps/current-work/source&out=gnu&asciiquotes=yes |
| 12:01 | <annevk> | I'm getting somewhere... |
| 12:01 | annevk | makes more tests |
| 12:46 | <Philip`> | annevk: "unpexted-hidden-input-in-table" - s/pex/expec/ |
| 12:47 | <Philip`> | Hmm, I added some new error strings in the html5lib tokeniser a short while ago, but didn't add a descriptive English version - should I have? |
| 12:47 | <Philip`> | (in r1104) |
| 12:48 | <Philip`> | (http://canvex.lazyilluminati.com/html5lib/changeset/1104 - ooh, Trac still works) |
| 12:52 | <annevk> | dunno, don't really care about it too much |
| 12:58 | <annevk> | it doesn't have my changes |
| 12:58 | <annevk> | local checkout or something? |
| 12:59 | <Philip`> | It only updates every six hours or so |
| 12:59 | <annevk> | k |
| 12:59 | <Philip`> | (since Trac requires a local repository, so I have to sync from the remote one) |
| 13:08 | <Philip`> | http://philip.html5.org/data/charsets.html |
| 13:08 | Philip` | wonders if he's missing something useful on there |
| 13:14 | <hsivonen> | Philip`: I'm surprised at how many pages claim to use MacRoman one way or another |
| 13:14 | <annevk> | the <meta content=" ... charset .." > case that doesn't have http-equiv |
| 13:17 | <hsivonen> | Philip`: it would be interesting to check if invalid gb2312 would be valid and sane if decoded as gb18030 |
| 13:20 | <Philip`> | annevk: The thing I called "http-equiv" is misleading because actually it entirely ignores the http-equiv value |
| 13:21 | <Philip`> | (It just looks at 'content', like the spec says) |
| 13:24 | <Philip`> | But I'll add a list of the actual http-equiv values that were seen |
| 13:30 | <Philip`> | Quite a lot of people put <meta> charset declarations inside <style> contents |
| 13:32 | <annevk> | <select>xxx why does "xxx" end up in the DOM per HTML5? |
| 13:33 | <annevk> | oh, IE does that |
| 13:34 | <annevk> | ok |
| 13:41 | <Philip`> | <meta http-equiv="Содержимое-Тип" content="text/html; charset=windows-1251"> |
| 13:43 | <annevk> | in select in table is now supported |
| 13:51 | <takkaria> | heh, http://www.glassdoctors.co.uk/chipsandmot.html begins with </html> |
| 13:56 | Philip` | gives up trying to use Unicode ids and #id links, and replaces everything with ASCII |
| 14:03 | <annevk> | ok, after after phases are just a more verbose trailing phase |
| 14:04 | <annevk> | hsivonen, not many tests were affected... i added tests7.dat though which has a bunch of new tests |
| 14:10 | <Philip`> | Of 760 declared gb2312 pages, 120 have errors decoding as gb2312, 8 have errors as gbk, 8 have errors as gb18030 |
| 14:18 | <annevk> | and gb2313 < gbk < gb18030 is the order? |
| 14:21 | <Philip`> | Yes |
| 14:25 | <Philip`> | (I don't know how many of these errors are caused by e.g. actually using utf-8 instead of gb*, where it happens to be a subset of gbk and not of gb2312) |
| 14:25 | <Philip`> | (*a subset of the byte sequences allowed by) |
| 14:39 | <annevk> | http://www.microsoft.com/windows/products/winfamily/ie/ie8/readiness/DevelopersNew.htm |
| 14:40 | <annevk> | "Six connections per host instead of two for broadband scenarios and a scriptable property allow for more improved performance by allowing parallelization of downloads in Internet Explorer 8." |
| 14:41 | <annevk> | "XMLHTTPRequest Enhancements include a timeout property that can be set to cancel the request if necessary, allowing developers to manage the request better." |
| 14:41 | <annevk> | it's good to know MS is still taking the W3C serious... |
| 14:58 | <hsivonen> | annevk: ok. thanks |
| 14:59 | <annevk> | hsivonen, please make more tests as you introduce changes, that will give us more coverage |
| 15:01 | <hsivonen> | annevk: ok |
| 15:01 | <annevk> | IE also has their own cross-site XHR |
| 15:01 | <annevk> | "XDomainRequest" |
| 15:02 | <annevk> | No idea if cross-document messaging matches HTML5 or not. I guess it matches given MS their editorial comments |
| 15:19 | <takkaria> | nice, they even use hatom without actually using hatom |
| 15:35 | <Philip`> | annevk: I can't click the "Install Internet Explorer 8" link on that page in Opera :-( |
| 15:35 | <annevk> | hit the lower end of the p |
| 15:35 | <annevk> | but yeah... |
| 15:36 | <annevk> | works fine in post-9.2 |
| 15:36 | Philip` | likes shift+arrow link navigation |
| 15:37 | Philip` | guesses IE8 won't work in Wine |
| 15:40 | <Philip`> | "Web sites that unknowingly access objects after they are freed may find such access failing in Internet Explorer 8" - why would an object be freed if it was still possible to access it? |
| 15:43 | <gsnedders> | namespaces are interesting |
| 15:44 | <gsnedders> | looks like they're changing what they already have |
| 15:59 | <othermaciej> | I'm certainly curious what XHR stuff they have done |
| 16:00 | <virtuelv> | hm, no word on canvas or svg? |
| 16:00 | <virtuelv> | oh, "Internet Explorer 8 offers Web developers the opportunity to write standards-compliant HTML-based Web pages that support features (such as SVG, XUL, and MathML) in namespaces, provided that the client has installed appropriate handlers for those namespaces via binary behaviors. (A binary behavior is a type of ActiveX control.)" |
| 16:01 | Philip` | wonders how they can do standards compliant namespace handling in HTML |
| 16:02 | <Philip`> | (particularly without breaking the people who depend on IE6's namespace handling) |
| 16:02 | <Philip`> | (unless the plan is to break a lot of people during the beta, so customers will demand that IE8 goes back to rendering in IE7 bug mode by default, so Microsoft will have a good reason for choosing to do so) |
| 16:04 | <annevk> | othermaciej, look like an ontimeout attribute that cancels the request after some amount of time and XDomainRequest, whatever that is... |
| 16:04 | <othermaciej> | Philip`: it looks like the only real namespace change is making it simpler to load a "binary behavior" |
| 16:04 | <othermaciej> | (from their description) |
| 16:05 | <othermaciej> | annevk: it would be nice to know how it differs from XXX |
| 16:05 | <annevk> | it seems like something completely new |
| 16:05 | <annevk> | :( |
| 16:05 | <annevk> | like a new type of object, a la JSONRequest |
| 16:06 | <annevk> | but there's not much information |
| 16:06 | <annevk> | i asked the interwebs, but so far there's nothing :) |
| 16:07 | <virtuelv> | I'm wondering why they are specifically mentioning XUL among those ActiveX controls |
| 16:08 | <annevk> | to show that they're trying not to be evil or something? dunno, XUL on the Web would not be good |
| 16:15 | <gsnedders> | Philip`: namespace change in IE8 standards mode only, I expect. Quirks doesn't change, nor does those who opt-in to IE7 Standards |
| 16:16 | <gsnedders> | Philip`: I expect most of those that rely on the namespace handling are quirks anyway |
| 16:17 | <Philip`> | gsnedders: I'd expect those that rely on it to be standards, because they're fancy scripts used by people who want to write proper XHTML 1.0 |
| 16:18 | <Philip`> | (The namespace syntax itself might be more quirky, but that's not very interesting - it's the scripted interaction with HTML-namespaces that matters) |
| 16:18 | <gsnedders> | why would the scriped based ineraction change? |
| 16:20 | <Philip`> | It's a new browser version which claims to have changes to its namespace support, so I'd assume its namespace support (quite possibly including how it interacts with DOM scripting) has changed |
| 16:22 | <gsnedders> | I was assuming the difference was an HTML serialisation |
| 16:22 | <annevk> | I also thought it was yet another quirk, but simpler, for text/html |
| 16:23 | <Philip`> | Uh, I'm not quite sure what people are saying |
| 16:32 | <zcorpan> | hsivonen: looking at http://simon.html5.org/test/aria/colon-vs-dash/ it seems that ie doesn't differ from the other browsers with the colon when it comes to .getAttribute() and .attributes |
| 16:33 | <zcorpan> | but namespace magic with \: in selectors is probably taking place, even for attribute selectors, which is why it doesn't match |
| 16:33 | <annevk> | I thought the issue was styling and it not being equivalent in HTML and XHTML |
| 16:34 | <zcorpan> | annevk: yeah |
| 16:34 | <annevk> | "just say no" |
| 16:35 | <zcorpan> | i'm just making sure we don't use non-issues as arguments :) |
| 16:38 | <annevk> | IE8 adds support for ARIA... |
| 16:41 | <annevk> | also, IE fixed their page zoom: "Primary features in this release include the elimination of horizontal scroll bars for the majority of mainstream scenarios and the introduction of persistent zoom states." |
| 16:41 | annevk | remember someone praising that.. glazou? |
| 16:42 | <Philip`> | http://www.glazman.org/weblog/dotclear/index.php?post/2008/02/11/Is-fullZoom-zooming |
| 16:43 | <Philip`> | I guess he'd say Google Maps zooming isn't zooming either, because the map changes (e.g. place names popping up) when you zoom |
| 16:44 | <Philip`> | so I'm not sure it's the most user-friendly definition of zooming |
| 17:18 | <zcorpan> | i wonder if ie8 fixed the bugs i pointed out in ie7b2 |
| 17:42 | <gsnedders> | zcorpan: I expect it has, and introduced totally new bugs |
| 17:44 | <Philip`> | gsnedders: That's what beta is for - we can test it, report all the bugs, and they will all be fixed before release |
| 17:45 | <gsnedders> | Philip`: because all bugs get found :P |
| 17:45 | <Philip`> | and fixed |
| 17:46 | <Philip`> | The web is too important to let people get away with releasing software with bugs |
| 17:47 | <gsnedders> | We need Acid4. |
| 17:47 | <Philip`> | What will it test? |
| 17:48 | <SadEagle> | Philip`: so not software should ever be released? |
| 17:48 | <gsnedders> | Everything. It will make sure nothing ever is released with bugs. |
| 17:48 | <Philip`> | http://code.msdn.microsoft.com/iemix08labs works now |
| 17:50 | <Philip`> | xdr = new XDomainRequest(); |
| 17:50 | <Philip`> | xdr.onerror = err; |
| 17:50 | <Philip`> | xdr.ontimeout = timeo; |
| 17:50 | <Philip`> | xdr.onprogress = progres; |
| 17:50 | <Philip`> | xdr.onload = loadd; |
| 17:50 | <Philip`> | xdr.timeout = tbTO.value; |
| 17:50 | <Philip`> | xdr.open("get", tbURL.value); |
| 17:50 | <Philip`> | xdr.send(); |
| 17:51 | <Philip`> | then the callbacks can read xdr.responseText, xdr.contentType |
| 17:52 | <Philip`> | XMLHttpRequest has xmlHttp.ontimeout=timeoutFired; xmlHttp.timeout = 5000; |
| 17:52 | <gsnedders> | "In this vain, Internet Explorer 8 will ship with an additional layout engine that is fully CSS 2.1 compliant upon final release." |
| 17:53 | <Philip`> | vain? |
| 17:54 | <gsnedders> | following on from other browsers approaching full support |
| 18:02 | <Philip`> | <activity category="define"> |
| 18:02 | <Philip`> | <activityAction context="selection"> |
| 18:02 | <Philip`> | <execute action="http://dictionary.reference.com/browse/{selection}"; method="get"/> |
| 18:02 | <Philip`> | </activityAction> |
| 18:03 | <Philip`> | </activity> |
| 18:03 | <Philip`> | Those braces seem like a familiar idea |
| 18:11 | <Philip`> | Oh, http://code.msdn.microsoft.com/ie8whitepapers exists too |
| 18:13 | <Philip`> | "Internet Explorer 8 hopes to implement some of the most requested CSS3 features by web developers and designers." |
| 18:13 | <Philip`> | (Apparently currently box-sizing and vertical text) |
| 18:14 | <annevk> | euhm, IE dispatches "hashChanged" while the event is named "hashchanged" |
| 18:15 | <annevk> | also, window.globalStorage["http://www.contoso.com";] |
| 18:15 | <annevk> | also, their API is asynchronous |
| 18:15 | <annevk> | ... |
| 18:16 | <annevk> | and yet they claim standards compliance |
| 18:18 | <Philip`> | Apparently they support |
| 18:18 | <Philip`> | <html> |
| 18:18 | <Philip`> | ... |
| 18:18 | <Philip`> | <svg width="600" height="300" xmlns="http://www.w3.org/2000/svg">; |
| 18:18 | <Philip`> | <rect x="0" y="0" width="100" height="100" style="fill:url(#gradient)" /> |
| 18:18 | <Philip`> | ... |
| 18:18 | <Philip`> | </svg> |
| 18:20 | <Philip`> | Also <li role="treeitem" aria-expanded="false"> and treeitem.ariaExpanded = "false"; |
| 18:20 | <annevk> | :( |
| 18:22 | <annevk> | maybe making a checklist for them will help... |
| 18:25 | <Philip`> | Maybe making test cases would help too |
| 18:29 | <annevk> | "When the title is absent, for accessibility Internet Explorer falls back to the alt attribute for a tooltip." |
| 18:30 | <Philip`> | "Attempting to navigate to a DATA URI via the address bar is not allowed. Data URIs may not be used as a source for FRAME or IFRAME elements. An OBJECT tag is limited to supporting non-navigable file types [i.e. not text/html]." |
| 18:32 | <met_> | annevk, "I was wondering if someone on the interwebs has more information on the various things listed on that page." Di you watched the IE8 streaming several minutes ago? |
| 18:32 | <annevk> | the stuff they released now is what i was looking for |
| 18:33 | <met_> | ok |
| 18:33 | <met_> | download still doesn't work |
| 18:34 | <met_> | some video are archived at http://visitmix.com/Default.aspx |
| 18:34 | <met_> | but not the official introduction yet |
| 18:35 | <met_> | only explanation of Web Slices |
| 18:41 | <annevk> | document.documentMode |
| 18:45 | met_ | is download IE8 now |
| 18:45 | Hixie_ | realises the html5 parser has draconian error handling for one class of errors |
| 18:46 | <gsnedders> | that shouldn't happen. |
| 18:46 | <Hixie_> | actually, i intend to keep it |
| 18:46 | <Hixie_> | the class of errors is "unexpected end of file" :-P |
| 18:47 | <gsnedders> | EOF when? |
| 18:47 | <gsnedders> | met_: download? where? |
| 18:47 | <jandem> | gsnedders: http://www.microsoft.com/windows/products/winfamily/ie/ie8/readiness/Install.htm |
| 18:47 | <Hixie_> | gsnedders: anywhere it's not allowed :-) |
| 18:47 | <met_> | http://www.microsoft.com/windows/products/winfamily/ie/ie8/readiness/Install.htm |
| 18:47 | <gsnedders> | so they work now |
| 18:47 | <met_> | looks like some mirros are syncing |
| 18:48 | <met_> | one minute it works, the second not |
| 18:48 | met_ | is installing... |
| 18:50 | <annevk> | Hixie_, how is it draconian? |
| 18:55 | <Hixie_> | annevk: it stops parsing! |
| 18:57 | <met_> | It has big 'Emulate IE7' button. But you have to restart IE after switching. |
| 18:58 | <annevk> | Hixie_, :p |
| 18:59 | <itpastorn> | Anne, I've written on Wikipedia that you was indeed the first to blog about Acid3. Can you confirm this? |
| 18:59 | <annevk> | the interwebs is near-infinite, how can I tell? |
| 19:00 | <met_> | Score 17/100 in Acid3 and it doesn't pass Acid2 completelly. |
| 19:00 | <itpastorn> | As far as you know? |
| 19:01 | <annevk> | maybe... |
| 19:01 | SadEagle | gives kudos to annevk & co for opera passing this incredibly evil testcase |
| 19:01 | <itpastorn> | (There was nothing on Technorati before Jan 10) |
| 19:02 | <annevk> | itpastorn, ok |
| 19:03 | <itpastorn> | Thanks! |
| 19:03 | <itpastorn> | Right now I am editing the (X)HTML 5 article on the Swedish WP. Anyone on this channel that speaks Swedish is invited to help. |
| 19:15 | <Philip`> | IE8 takes a long time to install compared to other web browsers |
| 19:20 | <annevk> | SadEagle, you make me curious :) |
| 19:22 | <SadEagle> | annevk: it's not that big of a deal, on second thought --- only mozilla gets it wrong (sequencing in reference evaluation in JS) |
| 19:22 | <annevk> | k |
| 19:23 | <annevk> | Philip`, it's 3 browsers! |
| 19:24 | <Philip`> | Also it takes an extra two years to finish installing when you reboot |
| 19:27 | <annevk> | XBL2 in 8k of JS, nice |
| 19:28 | <othermaciej> | annevk: Microsoft has a whitepaper up now, but it doesn't seem to document XDocmainRequest |
| 19:28 | Philip` | waits while Windows continues rebooting |
| 19:29 | <Philip`> | othermaciej: http://code.msdn.microsoft.com/ie8whitepapers/Release/ProjectReleases.aspx?ReleaseId=581 seems to document it (to some extent) |
| 19:29 | <annevk> | othermaciej, it seems to be XHR but slightly different |
| 19:30 | <annevk> | it doesn't send cookies/auth data and requires the server to reply with a header that's basically a boolean |
| 19:33 | <Philip`> | Hmm, they do the address bar significant-part-of-domain highlighting thing that FF3 used to do |
| 19:34 | <Philip`> | Live DOM Viewer looks ugly and says "Line: 127 Error: Not implemented" |
| 19:35 | <Philip`> | Node.attributes.length is not implemented |
| 19:44 | <Philip`> | and .item(n) is not implemented either |
| 19:51 | <hsivonen> | Hixie_: according to a new MS whitepaper, IE8 will do aria-* |
| 19:51 | <roc> | I'm sure Ian's reading those whitepapers right now |
| 19:51 | <roc> | just like we are :-) |
| 19:51 | <Philip`> | http://philip.html5.org/misc/live-dom-viewer-ie8.html |
| 19:51 | <Philip`> | It still does non-tree DOMs |
| 19:57 | <Hixie_> | actually i'm busy fixing the spec |
| 19:57 | <Hixie_> | hsivonen: sure hope they document how it interacts with html |
| 19:57 | <hsivonen> | Hixie_: yeah. |
| 19:58 | <hsivonen> | Hixie_: what I'm thinking is that with Firefox 3, Opera 9.5 and IE8 on board, having the TAG second-guess the syntax at this point wouldn't make sense |
| 19:59 | <Hixie_> | would there ever be a point where having the TAG second-guess the syntax makes sense? |
| 19:59 | <hsivonen> | well, no, but now it would make even less sense |
| 20:00 | <Philip`> | annevk: It does seem to be 'hashchange' and not 'hashChange', though attachEvent('hash[cC]hange', ...) doesn't work so I can't see how to test it except with onhashchange |
| 20:00 | <annevk> | apparently mousewheel is also dispatched for vertical scrolling, I didn't know |
| 20:00 | <hsivonen> | I want to see documentation explaining how their new namespace stuff integrates with parsing |
| 20:00 | <Hixie_> | hsivonen: less than no sense is still... no sense |
| 20:00 | <annevk> | Philip`, you can check event.type maybe? |
| 20:01 | <Philip`> | annevk: That says "hashchange" |
| 20:01 | <hsivonen> | looks like we will have to follow IE8 on the namespace thing now that HTML5 didn't spec SVG integration before they got on the case |
| 20:01 | <annevk> | Philip`, cool |
| 20:02 | <hsivonen> | does message grouping on validator.nu work in IE8? |
| 20:05 | <hsivonen> | Philip`: re: Charset usage data: my wild guess is that TIS-620 needs to be treated like ISO-8859-11. that is, treat as alias for the windows-* superset |
| 20:05 | <Philip`> | <svg xmlns="..."><circle/>foo in the Live DOM Viewer is parsed into an SVG element with attribute name 'xmlns="..."><circle' and value '...">circle', followed by a CIRCLE element |
| 20:05 | <Philip`> | (It's happier if you use xmlns:something instead) |
| 20:06 | <Philip`> | (but it does the old non-XML IE unknown-element parsing, as far as I can tell) |
| 20:08 | <Philip`> | (It's the same in real pages as well as the Live DOM Viewer) |
| 20:10 | <Philip`> | (As far as I can tell, the namespace support here is exactly the same as in IE6) |
| 20:11 | <Philip`> | (except for that xmlns parsing bug) |
| 20:11 | <annevk> | Philip`, Acid3 screenshot? |
| 20:12 | <Philip`> | Do MacBook Pro keyboards have a print-screen key? |
| 20:12 | <gsnedders> | fn+somethig |
| 20:13 | <roc> | their performance improvements are all JS and DOM |
| 20:13 | <Philip`> | That still leaves a hundred possibilities, assuming there aren't any other modifier keys :-p |
| 20:13 | <roc> | oops wrong channel |
| 20:16 | Philip` | uses the on-screen keyboard, then finds that Paint.NET has expired |
| 20:18 | <annevk> | someone else told me they got 17 |
| 20:19 | <Philip`> | I see 17 too |
| 20:19 | <Philip`> | (eventually) |
| 20:19 | <Philip`> | and some XML errors before that |
| 20:21 | <Philip`> | http://canvex.lazyilluminati.com/misc/ie8-acid3-partial.png - http://canvex.lazyilluminati.com/misc/ie8-acid3.png |
| 20:22 | <Philip`> | I think it didn't quite pass |
| 20:30 | <Philip`> | On the WHATWG HTML5 spec: "Line: 40 Error: Object doesn't support this property or method" |
| 20:30 | <Philip`> | Oh, that's just addEventListener |
| 20:34 | <met_> | i am not able to pass the http://webkit.org/perf/slickspeed/ every 15 seconds IE8 popup alert wifth Want-to-stop-the-script? |
| 20:34 | <Philip`> | It can sort of load the HTML5 spec, but it freezes for literally 20 seconds to highlight a link when you mouse over it |
| 20:34 | <Philip`> | and 10 seconds when I move the mouse out again |
| 20:43 | <Philip`> | The numbers on http://www.w3.org/TR/html5/#serialising go "1.", "1.", "2." |
| 20:43 | <Philip`> | and the next section goes 1, 2, 3, 4, 5, 6, 7, 8, 8, 9, 10, 11, 13 |
| 20:43 | <Hixie_> | is that my fault? |
| 20:43 | <Philip`> | I think it's IE8's fault |
| 20:44 | <Hixie_> | how they heck do you get _counting_ wrong |
| 20:45 | <Hixie_> | ok browsers need to stop asking me if i want to change my default browser |
| 20:51 | <othermaciej> | roc: wrong channel or not, I'm curious what exactly their perf improvements are - have you tried any benchmarks? |
| 20:52 | <roc> | totally unscientific, but one of our guys ran sunspider and they're 2x slower than FF3 trunk |
| 20:52 | <roc> | currently |
| 20:52 | <annevk> | from the release notes: "The Internet Explorer 8 AJAX features Cross Document Messaging and DOM Storage are based on the HTML 5.0 specification. The specification is in draft stage and is continually being updated. Therefore, our implementation of the specification may be based on an older version. The Internet Explorer team will look into updating the implementation when possible after Beta 1." |
| 20:52 | <roc> | I presume you know exactly where you are relative to FF3 trunk :-) |
| 20:53 | <Philip`> | The HTML5 spec does not give positive indications of IE8 performance |
| 20:53 | <hsivonen> | annevk: is there any indication where the XDomainRequest stuff came from? |
| 20:54 | <Philip`> | a:after{content:attr(href}} shows the absolute resolved URI instead of the href attribute, in case anyone is trying to collect a list of IE8 bugs |
| 20:55 | <annevk> | hsivonen, no |
| 20:56 | <annevk> | actually, maybe the open ajax alliance |
| 20:56 | <annevk> | but that's just a wild guess |
| 20:58 | <Hixie_> | oh THAT's why my site slowed down |
| 20:58 | <Hixie_> | acid3 is on slashdot |
| 21:00 | <gsnedders> | /. slow a site down? never! |
| 21:02 | <Hixie_> | the site would be fine if i increased my memory allocation |
| 21:02 | <Hixie_> | and the site is responding fine |
| 21:03 | gsnedders | pastes the entire HTML 5 spec into jgraham's outline tester |
| 21:03 | <itpastorn> | And once again the talk on /. is full of insighful remarks... NOT! |
| 21:03 | <gsnedders> | itpastorn: +1 informative |
| 21:04 | <othermaciej> | roc: I was wondering how they do compared to themselves, but 2x slower than FF3 trunk would have to be considerably faster than they used to be |
| 21:04 | <roc> | yeah |
| 21:04 | <gsnedders> | what is FF3 compared with FF2 anyway? |
| 21:04 | <roc> | if you read the whitepaper, they talk about string and array improvements in particular |
| 21:05 | <roc> | gsnedders: "a lot faster" |
| 21:05 | <gsnedders> | roc: I'm well aware :) |
| 21:05 | <othermaciej> | ~2x I think |
| 21:06 | <gsnedders> | so IE8b1 is approx FF2? that's even further than I thought |
| 21:06 | <met_> | Philip`, did you see acid2 correctly or not? |
| 21:06 | <roc> | I suggest you benchmark it yourself :-) |
| 21:06 | <met_> | this guy see it correcly http://www.arcanology.com/2008/03/05/ie8-beta-1-is-released/ but not me |
| 21:06 | <gsnedders> | nor me |
| 21:07 | <met_> | I see texarea in the right eye |
| 21:07 | <hsivonen> | itpastorn: re: (X)HTML 5 Swedish wikipedia page: the history part focuses on contrasting HTML5 with XHTML2. It would be good to mention the Opera/Mozilla position paper from The Workshop. |
| 21:07 | <gsnedders> | met_: ditto |
| 21:08 | <gsnedders> | met_: what OS? I was trying on Vista |
| 21:08 | <met_> | Win XP |
| 21:09 | <annevk> | itpastorn, Selectors API did not come from HTML5 |
| 21:09 | <annevk> | itpastorn, and if "Utvidgningar" means extensions, those did not come from HTML5 / Web Forms 2 either |
| 21:10 | <annevk> | re: talk page |
| 21:11 | <jgraham> | gsnedders: There's little point in doing that; I haven't fixed the bugs that stop it working yet :) |
| 21:11 | <gsnedders> | jgraham: and it crashes Saf :) |
| 21:11 | <annevk> | ouch, you need one account per language, that sucks |
| 21:11 | <jgraham> | gsnedders: Really? Cool |
| 21:11 | <hsivonen> | annevk: indeed. |
| 21:12 | <gsnedders> | jgraham: well, not crash. just pasting is slow in Safari. |
| 21:12 | hsivonen | so far has accounts for en, fi, fr, sv and de |
| 21:12 | <Philip`> | met_: I didn't see Acid2, and I can't now because all the sites are dead |
| 21:12 | jgraham | finishes reading scrollback and goes bug hunting |
| 21:12 | <annevk> | i have two on en by accident |
| 21:12 | <gsnedders> | jgraham: well, not noticeably slow normally, but when pasting something of the size of HTML 5… |
| 21:12 | <met_> | yes both |
| 21:12 | <met_> | looks like whole world is acid-testing now 8-) |
| 21:13 | <Philip`> | and Google Cache is not the best place for running such tests |
| 21:13 | <Philip`> | Does anybody have a mirror? :-) |
| 21:13 | <annevk> | heh, the whole Hixie_ server is down |
| 21:14 | <takkaria> | gsnedders: re: benchmarks, sunspider took ff2 ~30s, it takes ff3 ~8s |
| 21:14 | <takkaria> | gsnedders: http://ajaxian.com/archives/firefox-3-performance-numbers |
| 21:16 | <gsnedders> | takkaria: thx |
| 21:16 | <itpastorn> | Anne: You are a brave soull, trying to read Swedish (multipart) |
| 21:17 | <itpastorn> | Utvidgningar means additions. But "besl�ktad" means realted. So it is not *from* HTML 5, but related to it. |
| 21:17 | <annevk> | http://code.google.com/p/xbl/wiki/Features |
| 21:17 | <Philip`> | met_: I see the textarea, and the left eye is orange and has a red line underneath it |
| 21:17 | <met_> | yeh |
| 21:18 | <Philip`> | (when I strip the header junk out of the Google cache copy, which is hopefully about correct) |
| 21:18 | <met_> | strange this guy has it correct http://www.arcanology.com/2008/03/05/ie8-beta-1-is-released/ |
| 21:18 | <Philip`> | (running on Vista) |
| 21:18 | <met_> | (WinXP) |
| 21:18 | <Philip`> | Um, I can't scroll up or down the page |
| 21:18 | <Philip`> | Is that a(nother) bug? |
| 21:21 | <Philip`> | Ah, yes, overflow:hidden disables scrolling |
| 21:21 | <Philip`> | (or at least scrolling via user interaction) |
| 21:21 | <itpastorn> | annevk: "related" in spirit, more than formally (It's a to do list for more articles rather than the article at hand. I needed to dump it quickly somewhere...) |
| 21:22 | <annevk> | k |
| 21:23 | <annevk> | most of the CSS features came in existence because of XForms intitially |
| 21:24 | <takkaria> | I take it people have seen http://disruptive-innovations.com/zoo/css3tests/selectorTest.html? |
| 21:25 | <Philip`> | IE8 seems to render its standards mode like other browsers' standards mode, not like almost standards |
| 21:25 | <itpastorn> | updated the talk page... |
| 21:25 | <itpastorn> | Added this conversation as reference ;-) |
| 21:26 | <Philip`> | (and it doesn't have an almost standards mode of its own) |
| 21:29 | <virtuelv> | acid2 fails from acidtests.org, according to a cow-orker |
| 21:29 | <SadEagle> | probably due to the load |
| 21:31 | <virtuelv> | and passes on webstandards.org |
| 21:31 | <virtuelv> | I meant to write "acid2 in IE8" |
| 21:31 | <hasather> | virtuelv: the eyes were slow loading |
| 21:33 | <annevk> | http://realtech.burningbird.net/standards/ie8-first-looks/ is funny |
| 21:33 | <virtuelv> | quoting, "seems the object fallback fails if the 404 page is longer than a certain limit" |
| 21:36 | <roc> | annevk: I don't get that XBL2-in-JS page. How can it make the nodes anonymous? |
| 21:36 | <met_> | in commnents people also complain about noncompleted acid2 http://blogs.msdn.com/ie/archive/2008/03/05/internet-explorer-8-beta-1-for-developers-now-available.aspx |
| 21:36 | <annevk> | roc, i suppose it doesn't, unless it provides a wrapper for the actual DOM or so |
| 21:37 | <roc> | ok, so it's just a kind of preprocessor with XBL2 syntax |
| 21:37 | <roc> | that's useful |
| 21:37 | <roc> | but maybe the advertising is misleading :-) |
| 21:38 | <gsnedders> | "The test doesn't work from the other domains you listed because the call for the eyes is cross domain; IE8 currently doesn't trust that cross domain call. We're working through whether we can change that for beta 2 safely." |
| 21:40 | <met_> | some complaint how thay changed microformat hAtom http://blogmatrix.blogmatrix.com/:entry:blogmatrix-2008-03-05-0000/ |
| 21:42 | <gsnedders> | met_: Tony has just said they'll post about that later |
| 21:43 | <met_> | looks there is some problem with opacity e.g. http://www.debian-linux.cz/ in IE7 mode has opaque articles, but not in IE8 mode |
| 21:50 | <roc> | apparently opacity is just broken in the new engine. Sounds like they rushed things out for MIX |
| 21:53 | <Philip`> | http://canvex.lazyilluminati.com/misc/ie8/ is what I've stumbled upon so far |
| 21:54 | <Philip`> | Did they mix up "beta" with "untested build we copied off some developer's machine last night"? |
| 21:55 | <met_> | They cannot show Acid2 demo in MIX08 http://twitter.com/mollydotcom/statuses/767242657 |
| 21:55 | <annevk> | thanks to /. lol |
| 21:59 | jgraham | discovers that his computer gets unhappy when FF is using all the available memory |
| 22:00 | <roc> | I hope you're using FF3beta at least |
| 22:02 | <jgraham> | roc: Yeah it's the latest nightly. If I was guessing I'd say it was a problem with firebug anyway... |
| 22:17 | <annevk> | "Text nodes that have siblings appear as separate nodes in the HTML tree. If the text node is empty, an empty node will appear in the HTML tree." |
| 22:17 | <annevk> | -- http://support.microsoft.com/kb/949787 |
| 22:18 | <annevk> | (some weird stuff listed there that i don't really get) |
| 22:21 | <roc> | "Internet Explorer 8 Beta 1 crashes when you access the following Adobe AIR Download Center Web site:" |
| 22:21 | <roc> | I get that! |
| 22:21 | <roc> | boy, they still have a lot of work to do for IE Standards Mode |
| 22:22 | <roc> | accessibility |
| 22:22 | <roc> | printing |
| 22:23 | <gsnedders> | how much needs to be done for printing? I didn't see much wrong |
| 22:24 | <roc> | apparently it uses IE7 mode for printing |
| 22:24 | <othermaciej> | annevk: empty text node? |
| 22:25 | <othermaciej> | are they using that as a funny way to say "all whitespace"? |
| 22:26 | <annevk> | i was thinking the same, but i don't know |
| 22:31 | <Philip`> | http://canvex.lazyilluminati.com/misc/ie8/5.html |
| 22:31 | <Philip`> | IE8 says: 1, 2, 4 |
| 22:34 | <Philip`> | If I add more content inside the <object>, it says 4, 5, 6 |
| 22:35 | <Philip`> | Bugs are fun, particularly when they're not your own |
| 22:36 | <met_> | http://www.webstandards.org/files/acid2/test.html works again |
| 22:36 | <met_> | and the Acid2 there too |
| 22:36 | <met_> | looks like there is difference between http://www.webstandards.org/files/acid2/test.html and acid2.acidtest.org |
| 22:37 | <Philip`> | The eye lines fail on both those versions |
| 22:39 | <annevk> | as gsnedders quoted above there's something about a cross-site request |
| 22:40 | <met_> | oh looks like it, there is difference in the line |
| 22:40 | <met_> | line 132 |
| 22:40 | <met_> | <div class="eyes"><div id="eyes-a"><object data="data:application/x-unknown,ERROR"><object data="http://www.webstandards.org/404/"; type="text/html"><object data="data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAGAAAAAYCAYAAAFy7sgCAAAGsUlEQVRo3u2ZbWwcZxHHf3s%2B7LNbO3ZjXBtowprGODRX0qpNQCjmJKuVKhMl1P2AkCwhFOIKkCBSm9IXavGFKAixIAECwkmWo5MrhRI3Ub40IEwQgp6aIDg3Cd6eEqyIHEteah%2B1E69vhw%2BZtTaX8704ZzkKjHS6271nZ56ZZ%2BY%2F%2F%2BdZKF%2FCwYshx3EkkggLsD1v4FQkEZZYL |
| 22:40 | <met_> | CbAKyG9%2Ba9EIsG6hnUAf8x74K3aUC3j4%2BM54HcsR2oAIomwZOezkv%2FnSHpYNh%2BNCmAE7xv94zvFdd1bHsjMZmQkPSxAJP%2B%2FfuBLwK54PC7JZFKAVJmzXLBt2w%2FMvcDLwIb8QS8CeJ4nkURYIomw7J%2FYJ8BvSiiXptGGxWds2%2Fa9%2Bnaxh%2BYAD%2Bgt04NDgABTpQY2cvvSFLzw86gWeBVwC8SzlOSv2YeBPfmDBoBHgKmR9LBEEmHZfDTqGykqfkUE0nA78BzQGfSgUeP3wNeTXwXg7MwZDhw4UHL6ra2ti79%2FOvljgG8AZ4H64Lhm4MvAocxsRppGG%2FxcXihlwLIs6R%2FfKV2HO%2F26uA94pdDYUKUZUU7W1RQYXA98Gnhaf5%2FXWX0HeAHYoQonqa4sZSOsSWMCWeC9Yko%2BCQwBe |
| 22:40 | <met_> | 4E6oNc0Tc91XTl1%2BaTsn9gnI%2Blhyc5nZWxsrBIkKSbl2tiic3tW53YDEwOKaoFBrcOfqKee53lG9xsPMjV784r%2F4lO%2FpPvyJ9iyZcuvFSaXK5XYeAZ4CDgGvB3MS4B54LQuWYPeuy4iRFsevsXqpuYoqVQKIH2bK1CuDQNo11o4XUzh%2FcDWYIe1LEtyuZx4niee54njOGKapgfsqlL%2Bl2OjEXg8nxrc1dJ0h3hbtL%2BGCtz7KPBF4CuBe9uB15VafE8hr9qylI3HgG8C2%2FK7VyHZoJj7MrBRm30qFotJMpkU27YlHo%2F7Ha5a%2BV%2FKRkSJ4KuKRLVLKapTjB1SzAVIjY2NSXY%2BKyPpYdk%2FsU9OXT4pruv6BdZbBQfKsVGnvWlIe1VB6VQO8JxC1vZYLCbZ%2BaxsPhpdZDyRRFhG0sPiOE6ld |
| 22:40 | <met_> | KBg2lRg4xF1YCDIIIKN7DGgD3gH%2BBXwejKZfPrs2tPs%2FvPN2bKuYR1nd7xLKBSSJeqoXKnERjPwNWAG%2BLn2rZuM%2B4Tpml6vaWlp4eLcxVusZq5lCgVgOVKJjRqdX86ffL4D5wIoZACnTpw4wRMdT96i%2FImOJxERAs4uVyqxUacF%2FPdiCj%2BjdRBRGFtwXVdG0sPSdbhTmkYbpH98p2RmM2JZlig1vl0GWo4NQ%2Fn%2Bs5pKRXfwjweaxy7TND3HcRZbfC6X8xVPVQlGy7WxVWlO5XRXFXm6EZmrQuSXYyPE3SiVoEhE6Wyr0u2rumO6zv%2B21AFdQAswC1wCMuUCXCmyWQus103Qg8qlDO0lxwOb%2Fl4FiK3AB3VS%2FuKKLtK%2FgbeAnwG%2FvUODuRw%2FFrR0H1UC75fwu8oJ%2FhFsW5VIG%2FB |
| 22:40 | <met_> | UgEIN6Y65O4AHu4Ap0zQ9y7LEcZyb9lRBUHQcRyzL8unZVBW5bFWAvAp%2BhDQ2g4F47dUYtlU6obXA54DnVdFLekjUGGifh4AFy7LEdV3xj3X9I66m0QZpGm2QrsOd0j%2B%2BU0bSw5KZzYjrun6HWlAd961i4FfCj0aN1Usau%2Bc1lmuXPFwvAEumUut7tQQvAb%2FXb%2FT0bCAej9cODg7yt%2 |
| 22:40 | <met_> | ooops |
| 22:41 | <met_> | and the second <div class="eyes"><div id="eyes-a"><object data="data:application/x-unknown,ERROR"><object data="http://www.damowmow.com/404/"; type="text/html">..... |
| 22:45 | <Philip`> | "The Internet Explorer 8 AJAX features Cross Document Messaging and DOM Storage are based on the HTML 5.0 specification. The specification is in draft stage and is continually being updated. Therefore, our implementation of the specification may be based on an older version. The Internet Explorer team will look into updating the implementation when possible after Beta 1." |
| 22:46 | <annevk> | i quoted that earlire |
| 22:46 | <annevk> | earlier* |
| 22:46 | <met_> | good time for DOM storage, changed already |
| 22:46 | <Philip`> | annevk: Oh, I missed that |
| 22:46 | <Philip`> | and was too lazy to check upwards :-p |
| 22:47 | <annevk> | met_, same for postMessage |
| 22:47 | <met_> | 8-) |
| 22:47 | annevk | wonders if Firefox 3 is going to fix DOM storage in time |
| 22:47 | <met_> | someone is really lucky |
| 22:48 | <Hixie_> | my site should be back up again |
| 22:49 | <Hixie_> | i upgraded the server from like a 500MHz machine to a 2000MHz machine and quadrupled the RAM |
| 22:49 | <Hixie_> | i'll downgrade it again once /. is gone |
| 22:49 | <Hixie_> | interesting that /. is way harder on the site than reddit |
| 22:50 | <roc> | annevk: do you have the bug#? |
| 22:51 | <annevk> | roc, no |
| 22:51 | <annevk> | i wonder if there's even a definite bug on it |
| 22:51 | <gsnedders> | Hixie_: maybe reddit sux :P |
| 22:51 | <gsnedders> | Hixie_: /. is older, which probably helps the number of readers |
| 22:52 | <annevk> | roc, actually, seems to be https://bugzilla.mozilla.org/show_bug.cgi?id=407839 |
| 22:52 | annevk | likes gmail |
| 22:52 | <Hixie_> | why on earth does IE8 fail acid2 on acidtests.org |
| 22:52 | <roc> | yeah just found it thanks |
| 22:53 | <gsnedders> | Hixie_: "The test doesn't work from the other domains you listed because the call for the eyes is cross domain; IE8 currently doesn't trust that cross domain call. We're working through whether we can change that for beta 2 safely." |
| 22:53 | <roc> | it is blocking |
| 22:53 | <annevk> | Hixie_, cross-site request to damowmow.com |
| 22:53 | <Hixie_> | ? |
| 22:53 | <Hixie_> | oh, it's a bug |
| 22:53 | <Hixie_> | i see |
| 22:53 | <Hixie_> | fails on the wasp too, for me |
| 22:54 | <Hixie_> | oh, passes if i use www.webstandards.org instead of webstandards.org |
| 22:54 | <annevk> | makes sense |
| 22:54 | <annevk> | it's a full link |
| 22:55 | <Hixie_> | why would they show the iframe if it fails due to cross-domain |
| 22:55 | <Hixie_> | that's dumb |
| 22:57 | <Hixie_> | holy shit it renders my portal's display:table stuff correctly |
| 22:57 | <Hixie_> | it's ridiculously slow at painting |
| 23:00 | <gsnedders> | it's totally unoptimised, apparently |
| 23:01 | <Hixie_> | wow they implement the selector api |
| 23:01 | <Philip`> | It has some problems with the CSS on http://philip.html5.org/tests/canvas/suite/tests/ |
| 23:02 | <Philip`> | and the canvas emulation doesn't render - it just leaves empty spaces |
| 23:03 | <eseidel> | it does kinda feel like they just posted a nightly build |
| 23:03 | <eseidel> | but that's what beta's often feel like |
| 23:03 | <Hixie_> | yeah |
| 23:03 | <Hixie_> | that's what it often _is_ |
| 23:04 | <Philip`> | and the popup messages on http://philip.html5.org/tests/canvas/suite/tests/results.html don't work |
| 23:04 | <othermaciej> | it sounds like impressive progress |
| 23:04 | <othermaciej> | (haven't tried it to see though) |
| 23:04 | <othermaciej> | I wonder if they will try to reconcile XDR with XXX |
| 23:04 | <othermaciej> | it is cool that they implemented some HTML5 stuff |
| 23:05 | <othermaciej> | including the storage stuff that chris at one point said was out of scope, but that's water under the bridge |
| 23:05 | <Philip`> | and the button text alignment is wrong on http://philip.html5.org/tests/canvas/suite/tests/reportgen.html?100,0 |
| 23:05 | <annevk> | they did implement it with deviations |
| 23:05 | Philip` | wonders if he has any page that works correctly |
| 23:05 | <annevk> | DOM Storage in IE8 is asyn |
| 23:06 | <Hixie_> | wow apparently they fixed getAttribute() |
| 23:06 | <annevk> | (and it's also based on the old model) |
| 23:06 | <Hixie_> | that probably helped them get several points on acid3 |
| 23:09 | <Hixie_> | hm, sounds like they did the absolute minimum to pass acid2 as far as <object> goes |
| 23:10 | <jruderman> | http://www.flickr.com/photos/albill/2313174452/ makes it look like they're still getting 17/100 on acid3 |
| 23:10 | <jruderman> | which seems odd |
| 23:10 | <Hixie_> | i just tried it and got 10 |
| 23:10 | <Hixie_> | so |
| 23:10 | <Hixie_> | who knows |
| 23:10 | <Philip`> | http://canvex.lazyilluminati.com/misc/ie8-acid3.png |
| 23:10 | <Hixie_> | ie7 used to get 12, iirc |
| 23:10 | <jruderman> | oh |
| 23:10 | <Philip`> | after waiting quite a few seconds after it got stuck at around 10 |
| 23:11 | <Hixie_> | ah |
| 23:11 | <Hixie_> | interesting |
| 23:11 | <jruderman> | the screenshots on https://secure.wikimedia.org/wikipedia/commons/wiki/Acid3 are painful to read :P |
| 23:12 | <Hixie_> | oh god, why did they add more namespace crap to text/html |
| 23:12 | <jruderman> | "This website wants to run the following add-on: 'MSXML 3.0 SP9' from 'Microsoft Corporation'. If you trust the website..." |
| 23:12 | <jruderman> | wtf |
| 23:12 | <jruderman> | hixie: they couldn't make *all* their changes be towards standards compliance, could they? |
| 23:12 | <Hixie_> | haha |
| 23:13 | <Hixie_> | they had to explicitly NOT support the xhtml namespace |
| 23:13 | <Hixie_> | that's funny |
| 23:13 | <Hixie_> | i wonder how many pages they break due to this "namespace" support |
| 23:13 | <jruderman> | what. |
| 23:13 | <Hixie_> | aww, they got around it by not supporting xmlns="" on existing html elements |
| 23:15 | <jruderman> | they make web sites specify which SVG handler they hope you have installed? weird |
| 23:15 | <jruderman> | oh, that's ie5 |
| 23:18 | <jruderman> | do they support XHTML parsing too, or only this special text/html behavior? |
| 23:18 | <Hixie_> | good god |
| 23:19 | <Hixie_> | they couldn't even implement html5 without doing something other than what the spec said |
| 23:19 | <Philip`> | http://intertwingly.net/blog/2008/03/05/Improved-Namespace-Support#c1204755301 |
| 23:19 | <Philip`> | (about HTML namespaces) |
| 23:20 | <Hixie_> | indeed there's a blatent lie in this doc, it says that their stuff is compliant with html5 immediately after describing two methods and an event that have never been in the spec |
| 23:23 | <annevk> | it surprises me that they haven't said a single thing about it |
| 23:23 | <annevk> | on the list |
| 23:24 | <othermaciej> | Hixie_: which things does that apply to? |
| 23:25 | <Hixie_> | DOM storage |
| 23:26 | <annevk> | on globalStorage they have a method .x (forgot the name), a method .commit(), and a method .clear() |
| 23:26 | <annevk> | to do the exact opposite of what the API was intended for |
| 23:26 | <ianloic> | hey, I'm trying to describe a bunch of the patterns in DOM APIs so can model other APIs after it |
| 23:27 | <ianloic> | is the difference between a List type and a Collection type that collections are mutable and lists are immutable |
| 23:27 | <ianloic> | ? |
| 23:27 | <Hixie_> | on the plus side they're still using .globalStorage |
| 23:27 | <Hixie_> | so |
| 23:27 | <Hixie_> | no biggie |
| 23:28 | <Hixie_> | ianloic: no |
| 23:28 | <Hixie_> | ianloic: indeed i'm not really sure what the difference is |
| 23:28 | <Hixie_> | ianloic: in general i would say the DOM is highly inconsistent and any patterns are more luck than design |
| 23:28 | <Hixie_> | ianloic: and so i wouldn't use it as a good example of anything |
| 23:28 | <Hixie_> | Philip`: "<meta http-equiv="content-style-type" content="text/css; charset=iso-8859-1" />" is not legit |
| 23:29 | <Hixie_> | Philip`: so i'm happy to not worry about that |
| 23:29 | <Philip`> | Hixie_: Oh |
| 23:29 | <Philip`> | What's wrong with it? |
| 23:29 | <Hixie_> | there's no byte stream, the attributes are already in unicode by the time the content-style-type matters |
| 23:30 | <ianloic> | Hixie_, well, I'm trying to make an API that we're exposing from our browser (songbird) into web pages more web-ish - trying to do things similarly to the DOM (when it gets it right) is my best guess of a good strateg |
| 23:30 | <ianloic> | y |
| 23:30 | <ianloic> | Hixie_, but thanks for the List/Collection info - I'll go with "choose one and be consistent" :) |
| 23:30 | <Hixie_> | i recommend making a good attempt at making the API consistent, and then posting it to a list like public-webapi or the whatwg implementor's list and asking for feedback |
| 23:31 | <othermaciej> | ianloic: I'm not sure doing things "similar to the DOM" makes for good API |
| 23:31 | <ianloic> | othermaciej, well, I'd like to make things recognizable to web developers and accessible from js libraries |
| 23:32 | <ianloic> | othermaciej, right now we're exposing a mozilla-like API to web pages which is way worse |
| 23:32 | <ianloic> | Hixie_, I'll definitely be seeking feedback from the wider community - we need to get this stuff right (ish) |
| 23:32 | <gavin_> | mozilla-like? |
| 23:33 | <othermaciej> | ianloic: I guess what I'm saying is this |
| 23:34 | <othermaciej> | if you expose something that's not actually a document using a DOM tree, DOM events, and DOM-style live connections, it's likely to be an awkward fit |
| 23:34 | <ianloic> | gavin_, we have interfaces that look like nsIObserver, not dom even listeners |
| 23:34 | <othermaciej> | I would say design to your domain and align with DOM where it makes sense |
| 23:34 | <othermaciej> | I do think events can be used in a pretty general way |
| 23:34 | <ianloic> | othermaciej, yeah - I don't want to force too much of it directly into the DOM model |
| 23:34 | <othermaciej> | so being an EventTarget is a good way to align |
| 23:34 | <Hixie_> | Philip`: it would be interesting to see the distribution of how many bytes it takes for the encoding detector to make a decision, instead of stopping it at 512 or 1024. |
| 23:35 | <othermaciej> | I think the live collections are both confusing and hard to implement well |
| 23:35 | <Hixie_> | Philip`: then we could see if there's a point at which there is a sharp reduction in benefit |
| 23:35 | <ianloic> | othermaciej, but I look at the way that DOM Storage integrates as a decent model |
| 23:35 | <Hixie_> | Philip`: (great study btw) |
| 23:36 | <ianloic> | othermaciej, I don't know that they're confusing for developers, but they're a pain to implement |
| 23:37 | <othermaciej> | I don't even know what kinda collections you're working with necessarily (I assume something musical if it's songbird) |
| 23:38 | <ianloic> | othermaciej, libraries & playlists |
| 23:39 | <othermaciej> | os yeah that seems like a case where you may want to query as well as iterate |
| 23:40 | ianloic | nods. |
| 23:40 | <othermaciej> | s/os/so/ |
| 23:43 | <Philip`> | Hixie_: Has the encoding detector changed since June 2007? |
| 23:43 | <Philip`> | (since apparently that's when the implementation I used is from) |
| 23:44 | <Hixie_> | yes, i changed parts of it yesterday |
| 23:44 | <Philip`> | Oh, okay |
| 23:44 | <Hixie_> | er |
| 23:44 | <Hixie_> | i mean |
| 23:44 | <Hixie_> | over the past couple of weeks |
| 23:44 | <Hixie_> | not yesterday |
| 23:44 | <Hixie_> | dunno where that came from |
| 23:44 | <annevk> | it's significantly different now |
| 23:44 | Philip` | will try implementing it some time so he can detect encodings more up-to-datedly |
| 23:45 | <annevk> | and non-deterministic :( |
| 23:45 | <Hixie_> | how so? |
| 23:46 | <annevk> | last I checked scripts could be executed twice depending on how much you did before parsing |
| 23:46 | <takkaria> | Hixie_: are there any more changes in the pipeline for the tokenizer, or is it pretty much done now? |
| 23:47 | <Hixie_> | annevk: oh, that's not the encoding detector though, that's just the whole mechanism |
| 23:47 | <Hixie_> | but yes |
| 23:47 | <Hixie_> | takkaria: depends what people find |
| 23:47 | <annevk> | Hixie_, I don't like that part |
| 23:47 | <Hixie_> | annevk: neither do i |
| 23:47 | <Hixie_> | annevk: if you have a better proposal... :-) |
| 23:47 | annevk | wants clear yes/no answers |
| 23:47 | <annevk> | right... |
| 23:52 | <jgraham> | Hixie_: Re: outlines; I think the bit that starts "Otherwise, if the element being entered has a rank equal to or greater than the heading of the current section" has sub-optimal behavior for cases like <h1><h2><h3><h2> (the final h2 will be a sibling of the h1 not the first h2) |
| 23:53 | <Hixie_> | uh yes, that's a bug |
| 23:53 | <Hixie_> | send mail |
| 23:54 | <jgraham> | OK, just checking I'm not too tired to think straight :) |
| 23:55 | <annevk> | is "greater rank" a header with a number closer to zero or further away from zero? |
| 23:56 | <jgraham> | I believe greater rank is closer to 0 |
| 23:57 | <jgraham> | "The h1 element is said to have the highest rank, the h6 element has the lowest rank" |
