| 00:16 | <othermaciej> | tantek: awesomesauce |
| 00:17 | <gavinc> | Turtle in HTML uses Turtle in HTML! ... for all 12 people that might care |
| 00:24 | <zewt> | really really close to giving up on webgl |
| 00:24 | <zewt> | they just seem to fuck up on purpose |
| 00:32 | <tantek_> | gavinc - insert obv turtles all the way down joke |
| 04:14 | <MikeSmith> | just heard that a lot of people at Opera Japan got layed off yesterday |
| 04:14 | <MikeSmith> | sad news |
| 06:46 | <MikeSmith> | seems oddly coincidental that in Chrome the only test failure of 2231 test cases from http://w3c-test.org/html/tests/submission/MathiasBynens/named-character-references/001.html is the very first case |
| 06:47 | <MikeSmith> | matjas: ↑ |
| 06:48 | <MikeSmith> | wonder if that's an artifact of the test automation or if it's a real failure |
| 06:48 | <MikeSmith> | everything passes in all browsers I ran it in |
| 06:49 | <zcorpan> | it passes in chrome for me |
| 06:50 | <MikeSmith> | really? |
| 06:50 | <MikeSmith> | weird |
| 06:50 | MikeSmith | tries canary |
| 06:50 | <MikeSmith> | I confirmed it fails also outside of the test harness with just a static file |
| 06:50 | <paul_irish> | passes here in canary. fails in stable. |
| 06:50 | <zcorpan> | i guess it was fixed then |
| 06:51 | <zcorpan> | the first test fails in my safari |
| 06:51 | <MikeSmith> | hmm, passes in my canary too |
| 06:51 | <paul_irish> | matjas was filing a few bugs with webkit over the past few weeks, so that's likely |
| 06:51 | <zcorpan> | (along with all the double-char entities) |
| 06:51 | <MikeSmith> | doesn't fail in my safari |
| 06:51 | <MikeSmith> | or WebKit nightly, rather |
| 06:52 | <paul_irish> | https://bugs.webkit.org/show_bug.cgi?id=87465 yup |
| 06:52 | <MikeSmith> | r120658 |
| 06:52 | <MikeSmith> | paul_irish: aha |
| 06:53 | <MikeSmith> | ok, I guess I haven't updated my chrome dev for a while |
| 06:53 | <MikeSmith> | was at 20.0.1115.1 |
| 06:54 | <MikeSmith> | while canary was at 21.0.1164.0 |
| 06:54 | <MikeSmith> | oh I guess 21.0.1164.0 is pretty old now too |
| 06:55 | MikeSmith | makes updates, spins fan on laptop |
| 06:55 | <MikeSmith> | 21.0.1179.0 |
| 07:10 | <matjas> | MikeSmith: yeah, I kept track of the bugs I found here: http://mathias.html5.org/tests/html/ |
| 07:11 | <matjas> | Trident still gets wrong |
| 07:11 | <zcorpan> | wat? |
| 07:11 | <zcorpan> | wrong how? |
| 07:11 | <matjas> | i’m not even kidding |
| 07:11 | <matjas> | https://connect.microsoft.com/IE/feedback/details/743819 |
| 07:12 | <matjas> |     all match U+0020 instead of U+00A0 |
| 07:12 | <zcorpan> | MS blocked my account |
| 07:13 | <matjas> | zcorpan: what? why? |
| 07:13 | <zcorpan> | but it still acts like a non-breaking space? |
| 07:13 | <zcorpan> | matjas: no idea |
| 07:13 | <matjas> | i guess… else you’d think someone would have noticed before |
| 07:14 | <zcorpan> | matjas: i spent like hours going through their "re-enable your account" procedure only to find out that it wasn't enough |
| 07:14 | <zcorpan> | so then i just went "fuck you" |
| 07:15 | <zcorpan> | not that i have filed like dozens of useful bug reports with that account or anything |
| 07:15 | <zcorpan> | i was probably a spammer |
| 07:17 | <hsivonen> | matjas: "The content that you requested cannot be found or you do not have permission to view it." |
| 07:18 | <hsivonen> | zcorpan: clearly not worthwhile trying to interact with connect.microsoft.com |
| 07:18 | <matjas> | hsivonen, zcorpan: screenshot: http://i.imgur.com/w0BcO.png |
| 07:19 | <hsivonen> | matjas: thanks. I guess "Access Restriction: Public" means something different in the MS land |
| 07:20 | <zcorpan> | hsivonen: yeah they have really tested my patience. in the ie7 era i filed some bugs, most of which got marked as "won't fix" (for ie7) or "by design" (even though it literally violated the spec), and then when ie8 development started, all my bugs had vanished |
| 07:23 | <MikeSmith> | matjas: thanks for the link |
| 07:23 | <hsivonen> | zcorpan: if you find a bug in IE, it's probably a better idea to submit the test case to the W3C test suite than to submit it to connect.microsoft.com |
| 07:23 | <zcorpan> | not to mention that while actually filing a bug, after having filled in all details in the bug, and then hitting "submit", they thought i spent too long filling the 89 required fields, so logged me out and forgot everything i filled in |
| 07:24 | <zcorpan> | FUCK. YOU. |
| 07:24 | <matjas> | wow |
| 07:24 | <zcorpan> | hsivonen: yes |
| 07:24 | <MikeSmith> | hsivonen: yeah, plus a public blog posting even |
| 07:25 | <matjas> | MikeSmith: let me know when you have time to set up the JS tests |
| 07:26 | <MikeSmith> | oh yeah |
| 07:26 | <MikeSmith> | thanks for reminding me |
| 07:26 | <MikeSmith> | we just need to set up a new repo, right? |
| 07:26 | <MikeSmith> | I can do that right now |
| 07:27 | <MikeSmith> | we just need to decide on a name |
| 07:27 | <MikeSmith> | can also set it up so it gets mirrored to github automatically |
| 07:27 | <MikeSmith> | as we did for testharness.js sources |
| 07:35 | <matjas> | MikeSmith: I don’t know, or a subdirectory in one of the existing repositories… whatever works |
| 07:35 | <matjas> | as for a name, how about “web-javascript” or “web-ecmascript”? |
| 07:36 | <MikeSmith> | matjas: OK, will make it a separate repo on dvcs.w3.org |
| 07:36 | <MikeSmith> | matjas: you have a W3C user account? |
| 07:36 | <matjas> | it’s’ll be like, test262 meets Web reality |
| 07:36 | <matjas> | MikeSmith: I do |
| 07:36 | <MikeSmith> | OK |
| 07:37 | <zcorpan> | "web ecmascript" matches the wiki page |
| 07:37 | <MikeSmith> | I will add your username to push perms for the repo |
| 07:37 | <MikeSmith> | oh |
| 07:37 | <MikeSmith> | should we do "web ecmascript" then? or change the name on wiki page? |
| 07:38 | zcorpan | doesn't actually care |
| 07:38 | <matjas> | “javascript” is more commonly used, but I don’t feel strongly either way |
| 07:38 | <matjas> | ECMAScript is the name of the spec after all |
| 07:40 | <zcorpan> | would be nice if someone revamped web ecmascript into a spec |
| 07:41 | <zcorpan> | currently only the string prototype methods are good enough for a spec |
| 07:52 | <MikeSmith> | matjas: ok, https://dvcs.w3.org/hg/web-ecmascript created and you are the only one who has push perms for it |
| 07:52 | <MikeSmith> | can add others later as needed |
| 07:54 | <MikeSmith> | just came across Apache Wicket, which looks really sensible (for an Apache project) |
| 07:54 | <MikeSmith> | http://wicket.apache.org/ |
| 07:54 | <MikeSmith> | "a POJO data model, and a refreshing lack of XML, Apache Wicket makes developing web-apps simple and enjoyable again" |
| 07:54 | <matjas> | zcorpan: I’d love to give it a whirl, but I need to read up on how spec writing is done |
| 07:54 | <MikeSmith> | (POJO is "plain old Java objects", apparently |
| 07:55 | <matjas> | MikeSmith: thanks! will tweak my test case to use testharness.js and push the result later today |
| 07:55 | <MikeSmith> | super |
| 07:55 | <MikeSmith> | if you run into any problems lemme know |
| 07:55 | <MikeSmith> | but it should just work |
| 07:57 | <Von_Davidicus> | Everyone seems to have a hate-on for XML. |
| 07:58 | <MikeSmith> | Von_Davidicus: no, just don't like what it's used for |
| 07:59 | <Von_Davidicus> | Please explain. :) |
| 07:59 | <MikeSmith> | Maven. |
| 08:00 | <MikeSmith> | or, alternative, just make that "Most every Apache software project." |
| 08:01 | <Von_Davidicus> | Sorry, that doesn't explain much to me. :/ |
| 08:01 | <zcorpan> | matjas: cool! check out http://wiki.whatwg.org/wiki/Howto_spec |
| 08:01 | <MikeSmith> | Von_Davidicus: and I'm sorry but I don't have the energy to spell it out |
| 08:02 | <zcorpan> | matjas: then just try it, and ask for advice and feedback along the way :-) |
| 08:02 | <Von_Davidicus> | Okay. |
| 08:03 | <veosotano_> | I think the biggest con of XML is the verbosity. That, and that a lot of people try to use it to express data types, which it wasn't created for |
| 08:04 | <Von_Davidicus> | What do you mean by verbosity, veo? |
| 08:05 | <veosotano_> | well if you wanted to do something like Apple's plist |
| 08:05 | <veosotano_> | JSON would fit the bill way better |
| 08:07 | <veosotano_> | I think XML's biggest strength is being able to express structured information, when it's not about ints or floats or strings, but real world concepts |
| 08:07 | <Von_Davidicus> | I liked XML because it let me write smaller files than HTML did--which is why I have a couple of my sites the way I do. |
| 08:08 | <veosotano_> | you use xslt? |
| 08:08 | <Von_Davidicus> | Yes, one site I'm changing back over to SQL+PHP, but right now it's XML+XSLT(+DTD+XSD) |
| 08:09 | <annevk> | matjas: MikeSmith: zcorpan: could call it "JavaScript" to fix that mistake at the same time :) |
| 08:10 | <zcorpan> | (+LSD) |
| 08:10 | <annevk> | matjas: JavaScript is ECMAScript plus some modifications to make it compatible with the interwebs |
| 08:10 | <zcorpan> | annevk: too late :-P |
| 08:10 | <jgraham> | MikeSmith: BTW the github mirrors are really useful |
| 08:10 | <annevk> | aaah |
| 08:10 | <matjas> | so… Web ECMAScript == JavaScript |
| 08:11 | <MikeSmith> | jgraham: yeah, glad you suggested setting those up |
| 08:11 | <jgraham> | MikeSmith: Saves me having to do the same thing at this end to clone repositorities into git-happy Opera |
| 08:11 | <MikeSmith> | ah |
| 08:11 | <jgraham> | Well that wasn't the intention at the time |
| 08:11 | <jgraham> | But it is a happy side effect :) |
| 08:12 | <Von_Davidicus> | zcorpan... are you suggesting I was on drugs when I made that site? |
| 08:13 | <zcorpan> | Von_Davidicus: no, i'm suggesting the site is on drugs :-) (though i haven't looked at the site, and surely don't mean to offend you) |
| 08:17 | <jgraham> | Oh, did we reach XML hour already? |
| 08:20 | <Von_Davidicus> | Let me show you, then. http://coachrandom.zzl.org It's a webcomic. Some of the pages are HTML--the rest are XML. |
| 08:20 | <zcorpan> | matjas: i just realized much of Howto_spec doesn't apply for javascript specs |
| 08:21 | <zcorpan> | matjas: since it's trying to address DOM specs |
| 08:25 | <annevk> | "for JavaScript specs" as there's so many of them? :) |
| 08:27 | <zcorpan> | happens to be one matjas considered writing |
| 08:27 | <zcorpan> | i'm not saying Howto_spec should be changed :-) |
| 08:27 | <annevk> | for JavaScript I guess you want to model it after the es5.github.com |
| 08:28 | <annevk> | most of the concepts like using list of steps and such still seem to apply though |
| 08:28 | <zcorpan> | yeah |
| 08:29 | <matjas> | MikeSmith: as for the directory structure, web-ecmascript/test-name/001.html etc. or do you need an extra /tests/ in there? |
| 08:29 | <matjas> | /tests/{submissions,approved}/, even |
| 08:33 | <jgraham> | Please don't use submitted/approved for new things |
| 08:33 | <jgraham> | I think it is now generally agreed to be a bad idea |
| 08:37 | <zcorpan> | matjas: or test-name.html |
| 08:38 | <matjas> | zcorpan: i’d prefer directories as I store a lot of test data in separate .js files |
| 08:38 | <MikeSmith> | matjas: what zcorpan said |
| 08:38 | <MikeSmith> | assume the tests will stay at the same URL forever |
| 08:38 | <zcorpan> | matjas: yeah that's fine |
| 08:38 | <matjas> | would you be okay with directories instead? |
| 08:38 | <matjas> | k |
| 08:39 | <zcorpan> | matjas: just that 001.html is sort of an anti-pattern (that i've followed for far too long) and useful filenames are generally better |
| 08:39 | <matjas> | ah, i got that from annevk and you :) |
| 08:40 | <zcorpan> | yeah and i guess we got that from Hixie |
| 08:40 | <matjas> | it is kinda useful if you have multiple separate tests as .html files |
| 08:41 | <zcorpan> | it can be useful sometimes, like right now i'm writing lots of CORS tests that follow a table in an .ods file so each file corresponds to a row in the table |
| 08:42 | <zcorpan> | but for e.g. the websocket testsuite i now wish i used useful filenames instead |
| 08:43 | <zcorpan> | since it's a PITA to find the test you're looking for |
| 08:43 | <jgraham> | I have been using 001 type filenames and I don't really like it |
| 08:44 | <jgraham> | But I don't really want to try and condense the title even further into the filename either |
| 08:45 | <matjas> | MikeSmith: where can I browse the https://dvcs.w3.org/hg/web-ecmascript repository over HTTP (i.e. as a dir listing)? |
| 08:46 | <jgraham> | (or have titles like link_with_onclick_form_submit_to_javascript_url_with_document_write_and_href_navigation.html) |
| 08:46 | <MikeSmith> | matjas: you can't really (not the way you can at github or bitbucket) |
| 08:46 | <matjas> | MikeSmith: i was expecting something like http://w3c-test.org/ |
| 08:47 | <MikeSmith> | ah |
| 08:47 | <MikeSmith> | yeah |
| 08:47 | <MikeSmith> | I can set that up |
| 08:47 | <MikeSmith> | but it takes some time |
| 08:47 | <MikeSmith> | will need to wait and do it later today |
| 08:47 | matjas | URI-guessed http://w3c-test.org/web-ecmascript/ but failed |
| 08:47 | <MikeSmith> | right |
| 08:47 | <matjas> | MikeSmith: ok, let me know when you do :) |
| 08:47 | <MikeSmith> | the repos don't get mirrored automatically |
| 08:48 | <MikeSmith> | OK |
| 08:48 | <zcorpan> | jgraham: at least the latter is easy to find again later, but i agree it can be ugly or annoying to write |
| 08:48 | <MikeSmith> | it requires me getting this kludgy rsync thing configured |
| 08:48 | <MikeSmith> | matjas: will ping you when it's ready |
| 08:48 | <matjas> | thanks |
| 08:51 | <zcorpan> | jgraham: does the spec define execution order for http://software.hixie.ch/utilities/js/live-dom-viewer/saved/1598 ? |
| 08:54 | <jgraham> | zcorpan: AFAIK, no but I expect if has to be document order |
| 08:54 | <jgraham> | i.e. depth first |
| 08:55 | <jgraham> | (I wrote a test like that recently) |
| 08:55 | <jgraham> | (I guess I should file a bug or something) |
| 08:55 | <jgraham> | (unless someone knows this is defined somewhere?) |
| 09:00 | <zcorpan> | "When a script element that is not marked as being "parser-inserted" experiences one of the events listed in the following list, the user agent must synchronously prepare the script element: |
| 09:00 | <zcorpan> | * The script element gets inserted into a document." |
| 09:01 | <zcorpan> | that doesn't say in what order |
| 09:01 | <zcorpan> | so, yeah, file a spec bug :-) |
| 09:02 | <jgraham> | Right, I know that bit doesn't say |
| 09:03 | <jgraham> | But there could be a bit somewhere that says that when any bulk modifications happen it has to be as if they happened in document order or something |
| 09:04 | <zcorpan> | http://software.hixie.ch/utilities/js/live-dom-viewer/saved/1599 also |
| 09:04 | <zcorpan> | opera/chrome/firefox use document order |
| 09:05 | <Von_Davidicus> | Hmmm... which to use... XHTML 1.0, HTML 4.01, or HTML5.... |
| 09:07 | <annevk> | matjas: you can do stuff like http://dvcs.w3.org/hg/url/raw-file/tip/Overview.html |
| 09:07 | <zcorpan> | Von_Davidicus: use HTML |
| 09:08 | <annevk> | matjas: e.g. http://dvcs.w3.org/hg/web-ecmascript/raw-file/tip/identifiers/index.html |
| 09:08 | <zcorpan> | Von_Davidicus: (i.e. http://whatwg.org/html ) |
| 09:09 | <matjas> | annevk: thanks |
| 09:10 | <zcorpan> | annevk: wasn't that feature going to be disabled or something since tests generally aren't trusted (which is why we have w3c-test.org)? |
| 09:10 | <annevk> | zcorpan: if they disable that, it's going to be hell for all those permanent links to editor's drafts from TR/ |
| 09:10 | <annevk> | zcorpan: and from elsewhere on the web |
| 09:10 | <annevk> | zcorpan: so they'd need to have some kind of plan in place unless they're going to violate http://www.w3.org/Provider/Style/URI.html |
| 09:11 | <jgraham> | annevk: Aren't you the one that originally pointed out that it's a security risk? |
| 09:11 | <annevk> | jgraham: I have pointed that out, yes |
| 09:12 | <annevk> | jgraham: that's why we have a separate domain for tests |
| 09:12 | <annevk> | jgraham: but for some reason they put the test repository together with the more trusted spec repository |
| 09:16 | <Von_Davidicus> | Yes, but which one should I use? |
| 09:18 | <annevk> | Von_Davidicus: HTML is not versioned, see http://whatwg.org/html |
| 09:22 | <Von_Davidicus> | XHTML 1.0 (with appropriate doctype) it is. :-) I have consented to take a step back from XML+XSLT, but I have my limits. :D |
| 09:26 | <annevk> | haha |
| 09:26 | <zcorpan> | Von_Davidicus: the XHTML 1.0 doctype is a valid HTML doctype, so that's fine :-) |
| 09:28 | <Von_Davidicus> | My D2:LoD rune pages, however, remain XML+XSLT+XSD. *Stubborn nod* |
| 09:30 | <zcorpan> | turns out i don't quite know the xhtml 1.0 strict doctype by heart anymore |
| 09:31 | <zcorpan> | i guess that's a good thing |
| 09:31 | <zcorpan> | means i'm capable of forgetting useless information |
| 09:32 | <zcorpan> | but i had it almost right :-( |
| 09:37 | <Von_Davidicus> | I never did know it by heart. |
| 09:39 | <zcorpan> | <!doctype html> is much simpler to learn :-) |
| 09:39 | <Von_Davidicus> | Which is entirely the problem! |
| 09:40 | <zcorpan> | what's the problem? |
| 09:41 | <Von_Davidicus> | Have you any idea what a simple doctype DOES to my nerdacity, which is, sadly, based almost entirely on the fact that I do webcoding? |
| 09:42 | <Von_Davidicus> | What good is it, if most people's minds don't go ~sproing~ from the first line? :D |
| 09:46 | <zcorpan> | why not put in some zalgo art on the first line instead? :-) |
| 09:50 | Von_Davidicus | looks it up. |
| 09:50 | <Von_Davidicus> | Because it doesn't seem text-based, so I don't know how to stick that in via Notepad. |
| 09:57 | <zcorpan> | <!DOCTYPE html><!--PUBLIC "͞-̴//W3C̡//̡DT͢D X͠H̡T͡ML̴ 1̶.0 S̀t҉ric̵t//̧EN͟"҉ "͚h҉͈̠̲̠̯t̪̟̠̹̳͉t̜̻͚̙̜̙͇͡p̴̩͚ͅ:͙̳̳̮̰͜/̸͍̗/͡ẁ̙w͓w̡͈̩̭̫.̨̯̙̲̜w͉̼3̜̬̮̳̤̤.̹o̙̘r͈̼ͅg̤̣͚̬̱͚ͅ/͏̥̜̗Ṭ͕R̸̘͚̘/̡͍̼̻̣x͔̱̞̘̮̖̱͞h͙͖̘̗t̳̲͚m̟̯̬l͍̳̺1̣̻̖͉͈ͅ/͔D̶T̮D͇/̤̟͔̹̗̭x̀h̪̲̕t̸̠̠̼̼̙m̷̘l̶1̨̝̩̮͍ͅ-̴̰̞̯̦ |
| 09:57 | <zcorpan> | ́ͯ̇͏̶̫̯̯͚͚̜̮̼̬̬͞t̷̘̞̮̳̲̞͈͎̍ͩ̃ͣ̊̄͐̓̑͑͆ͨ̊̚̚͟r̈́ͮ͆͗̽̈̐̿̿͘͞͏̴̤̺͍̲͔̣̘̟̜̘̲̘̬͉͕͚͢i̶͇͉͙̩͍̺̣͖͕͇̦̘̺̫̹̳̥̼͈̓̉̐ͬͩ̆̈̒̉͛ͭ̎̈́ͭ̊͋͌͆͘c̴͓̥̲͈̼̋̔̊̃ͧ̆̓ţ͕̘̖͖̗ͩ̄̍̍̿ͣͭ͂ͮ̃̍̅̐͌̈̍ͮ̎͘͠.̶́ͥ̽̍̊ͨ̓͐ͫͨͥ̃͌͏̠̪͙̝̦͍̜̘̘͕͠ď̶̢͙͙̭̖̞̺̰̝͇ͯ̌̌ͭ͗͝ţ̸̴̡͌͂̒ͤ̑ͭ̑͏̝͕� |
| 09:57 | <zcorpan> | ̱̣̠̜d̢͍͈͙̫̲̰̥͈̗͇̱̦͎ͥͫͣ̍̑͗ͭ̆ͅ"̯̗̺̼̼̖̬͎͎̟͉̠̺̥̾͒͗̽ͣ̆̽̔́ͦ͗̉̾̔ͨ́̕͘͡ͅ--> |
| 09:57 | <Von_Davidicus> | Hmmmm..... |
| 09:58 | zcorpan | isn't sure that got through correctly |
| 09:58 | <Von_Davidicus> | If you meant for a load of strange characters to show up, it got through just fine. :) |
| 09:59 | <zcorpan> | see http://eeemo.net/ |
| 10:00 | <reggna> | Obfuscated HTML.. Nah. |
| 10:02 | <Von_Davidicus> | I agree. Obfuscated HTML is for wannabes. |
| 10:03 | <zcorpan> | a comment with funny characters isn't really what i'd call obfuscated html |
| 10:03 | <zcorpan> | anyway |
| 10:04 | <zcorpan> | seems it made it through correctly to the logs but my monospace fonts aren't capable of doing a sensible rendering |
| 10:04 | <Von_Davidicus> | No, if I'm going to make things horrendously complicated and mindscrewsome, I shall do it properly: with code that actually DOES something. |
| 10:04 | <zcorpan> | Von_Davidicus: the xhtml doctype doesn't actually do anything |
| 10:06 | <Von_Davidicus> | Well, it sort of does. 1) it gives the validator something to start with. 2) omit it, and try using character entity references. If you use firefox, I hope you like yellow. |
| 10:06 | <Philip`> | Someone should do an HTML Golf contest, where you're given a target web page and you have to write a new page that has pixel-identical rendering (in some selection of modern browsers) using as few bytes as possible |
| 10:07 | <zcorpan> | Von_Davidicus: oh i thought you were gonna use text/html |
| 10:07 | <Philip`> | (as with e.g. http://en.wikipedia.org/wiki/Perl_golf) |
| 10:08 | <Von_Davidicus> | Well, I will. But, still, it does more than a comment. |
| 10:09 | <zcorpan> | not really |
| 10:10 | <Von_Davidicus> | Anyways, I gotta sign off and get some sleep. |
| 10:39 | <kennyluck> | Does anyone know how old DOM4 (or Web DOM previously) is? |
| 10:40 | <zcorpan> | kennyluck: old as in when work on it started? |
| 10:41 | <kennyluck> | zcorpan, yeah, am I am sort of interested in the earliest version. |
| 10:41 | <kennyluck> | s/am/and/ |
| 10:41 | <kennyluck> | (for historical interest) |
| 10:42 | <zcorpan> | http://simon.html5.org/specs/web-dom-core is the earliest version still online, unfortunately i haven't kept earlier versions around anywhere. but iirc i started working on it in 2007 |
| 10:43 | <zcorpan> | probably late 2007 |
| 10:43 | <kennyluck> | zcorpan, great thanks! |
| 10:45 | <zcorpan> | if pimpmyspec keeps logs, might have more information |
| 10:45 | <zcorpan> | jgraham: does it? ^ |
| 10:47 | <jgraham> | Not apart from apache logs, and I don't know how long they are kept for |
| 10:55 | <zcorpan> | kennyluck: http://web.archive.org/web/20081211192358/http://simon.html5.org/specs/web-dom-core.src |
| 10:55 | <kennyluck> | zcorpan, thanks! |
| 10:57 | <kennyluck> | I am just curious about the history of createElementNS(). It looks like problematic part I spot has been like this for a while. I'll file a bug later. |
| 12:11 | <zewt> | oh cool, a ui change in recent ff that's actually really good, the backdrop for viewing images is grey instead of blinding white |
| 12:24 | <matjas> | that zcorpan has already written out the `ToHTMLTag` algorithm as well as the algorithms for the String.prototype extensions in spec lingo ♥ |
| 12:32 | <jgraham> | So… how synchronous is it to navigate to a javascript URL? |
| 12:53 | <MikeSmith> | matjas: can you please try pushing a change to web-ecmascript? |
| 12:53 | <MikeSmith> | doesn't matter what |
| 12:53 | <matjas> | MikeSmith: sure |
| 12:53 | <MikeSmith> | I think I have the mirroring set up correctly |
| 12:54 | <Ms2ger> | MikeSmith, can you give me access too? |
| 12:54 | <MikeSmith> | sure |
| 12:55 | <matjas> | MikeSmith: done, I got some rsync-related messages |
| 12:55 | <MikeSmith> | OK |
| 12:55 | <MikeSmith> | as long as you didn't see any errors, I think it worked |
| 12:55 | <MikeSmith> | yeah |
| 12:55 | <matjas> | where’s the github repo? |
| 12:55 | <MikeSmith> | http://w3c-test.org/web-ecmascript/ |
| 12:55 | <matjas> | ooh that was for the dir listings |
| 12:56 | <matjas> | nice! |
| 12:56 | <MikeSmith> | I have not set up the github rop |
| 12:56 | <MikeSmith> | *repo yet |
| 12:56 | <MikeSmith> | I will do that later |
| 12:56 | <matjas> | no rush |
| 12:56 | <MikeSmith> | OK |
| 12:58 | <MikeSmith> | Ms2ger: oK, you should also now have push perms for that repo |
| 13:00 | <Ms2ger> | matjas, do you want me to put my tests in directories or can I just dump them all under the root? |
| 13:01 | <matjas> | Ms2ger: IMHO directories are neat and tidy, but whatever works for you |
| 13:01 | <MikeSmith> | holy crap that's a lot of test cases there at http://w3c-test.org/web-ecmascript/identifiers/ |
| 13:02 | <MikeSmith> | 98719 is nice number |
| 13:02 | <MikeSmith> | eh wtf is this "Cached metadata not present. Click for source code." thing I keep seeing now |
| 13:03 | <zewt> | cool, i clicked that and firefox is angry with me |
| 13:03 | <zewt> | 1 fail :( |
| 13:03 | <Ms2ger> | There we go |
| 13:04 | <Ms2ger> | MikeSmith, plinss |
| 13:04 | <MikeSmith> | ah |
| 13:04 | <Ms2ger> | He wants each test to have two copies of all the metadata |
| 13:04 | <MikeSmith> | hmm |
| 13:04 | <Ms2ger> | Which I think is silly, but I haven't had time to complain yet |
| 13:05 | <Ms2ger> | Otoh, at least it doesn't complain about "cached metdata" anymore |
| 13:06 | <MikeSmith> | OK |
| 13:06 | <MikeSmith> | hey is iframe sandbox landed in gecko yet? |
| 13:07 | <Ms2ger> | Not yet |
| 13:07 | <Ms2ger> | It sounded like it was getting close, though |
| 13:08 | <MikeSmith> | OK |
| 13:08 | <MikeSmith> | heh https://www.w3.org/Bugs/Public/show_bug.cgi?id=17453 resolved=needsinfo |
| 13:08 | <MikeSmith> | cute |
| 13:08 | <MikeSmith> | I have not idea what that one is about |
| 13:10 | <Ms2ger> | Me neither |
| 13:10 | <Ms2ger> | Do you know people who use IE? |
| 13:11 | <MikeSmith> | Ms2ger: yeah, a few |
| 13:39 | <matjas> | how do you add a reference to a separate spec in Anolis? <some-markup-here-i-suppose>[FOO]</…> |
| 13:41 | <jgraham> | javascript: url navigation: synchronous or asynchronous? |
| 13:42 | <Ms2ger> | <span data-anolis-spec=html>foo</span> |
| 13:43 | <jgraham> | For example if I have <a target=some_iframe href=javascript:void(x++)><script>x=1;a.click();a*=3;alert(a)</script>, what should be alerted? 3? 6? |
| 13:44 | <jgraham> | What about if I submit to a javascript: url? |
| 13:44 | gsnedders | would rather not use jgraham's harness for Web-ECMAScript tests. |
| 13:44 | <gsnedders> | (I'd rather have the tests usable in a JS shell) |
| 13:45 | <gsnedders> | http://w3c-test.org/web-ecmascript/builtins/Array.prototype.join-order.html are just ES5 behaviour, no? |
| 13:46 | <matjas> | Ms2ger: thanks |
| 13:46 | <Ms2ger> | Np |
| 13:46 | <matjas> | Ms2ger: and where are these references defined? in Anolis itself? is there a list somewhere? |
| 13:46 | <Ms2ger> | The data part of http://wiki.whatwg.org/wiki/Anolis |
| 13:47 | <Ms2ger> | MikeSmith++ |
| 13:47 | <Ms2ger> | tantek++ |
| 13:47 | <Ms2ger> | for docs |
| 13:47 | <MikeSmith> | eh? |
| 13:47 | <MikeSmith> | which docs |
| 13:47 | <MikeSmith> | oh |
| 13:47 | <Ms2ger> | Those |
| 13:48 | <matjas> | “docs” :P |
| 13:48 | <MikeSmith> | yeah, I added that cause tantek asked me nicely :) |
| 13:48 | <matjas> | thanks though, that helps. So you need to clone that repository and modify it if you need other references? |
| 13:48 | <Ms2ger> | Yep |
| 13:48 | <Ms2ger> | Do you have a bitbucket account? |
| 13:49 | <jgraham> | THe correct way to install lxml is pip install lxml (you are using a virtualenv, right?) |
| 13:49 | <matjas> | I should note I’m using http://pimpmyspec.net/ for now rather than setting up Anolis locally |
| 13:50 | <matjas> | Ms2ger: I might have |
| 13:50 | <Ms2ger> | Mm |
| 13:50 | <MikeSmith> | volkmar: Web Activities is not proposing any new markup, right? e.g., nothing parallel to the proposed <intent> element? |
| 13:50 | <Ms2ger> | pms doesn't have support for cross-spec references yet |
| 13:51 | <MikeSmith> | jgraham: please update that wiki page :) |
| 13:51 | <matjas> | I see. Ok, at least I know what markup to use |
| 13:51 | <volkmar> | MikeSmith: no |
| 13:51 | <MikeSmith> | volkmar: no means no new element? |
| 13:51 | <volkmar> | MikeSmith: I thought I mention that in the wiki page: we don't have a strong opinion on that yet |
| 13:51 | <MikeSmith> | OK |
| 13:51 | <volkmar> | I'm not for an element in <body> |
| 13:52 | <MikeSmith> | OK |
| 13:52 | <MikeSmith> | I see that section now |
| 13:53 | <MikeSmith> | we don't want to add a new element in the head for this |
| 13:53 | <MikeSmith> | IMHO |
| 13:55 | <MikeSmith> | volkmar: I'm trying to write a general one-sentence description that covers the baseline common scope of both proposals |
| 13:55 | <MikeSmith> | how about "enabling Web applications to delegate user intents/activities to be handled by other applications"? |
| 13:56 | <volkmar> | sounds fine |
| 13:57 | <MikeSmith> | thanks |
| 14:04 | <Philip`> | jgraham: (I presume you meant "x*=3;alert(x)", not "a*=3;alert(a)") |
| 14:04 | <jgraham> | Yes |
| 14:05 | <zewt> | jgraham: looks like it's different in webkit and FF (which I suspect is why you're asking) |
| 14:06 | <zewt> | sync in webkit (or chrome, anyway), async in FF |
| 14:06 | <jgraham> | zewt: and IE and OPera |
| 14:06 | <jgraham> | Well IE matches WebKit and Gecko matches Opera |
| 14:08 | <zewt> | (also, in WebKit I can say "x", but in FF I have to say parent.x; probably obvious but I've never really done much with iframes) |
| 14:09 | <jgraham> | Yeah, well there is some weirdness there |
| 14:09 | <jgraham> | WebKit really doesn't like navigating frames with javascript URLs |
| 14:09 | <jgraham> | It always? tries to navigate the top level browsing context |
| 14:09 | <jgraham> | (maybe not if you set the location explicitly?) |
| 14:11 | <zewt> | i'm not even sure where the path in the spec is for executing javascript: during navigation |
| 14:12 | <zewt> | offhand it seems like it would be weird if a.click() caused the storage mutex to be released synchronously |
| 14:13 | <zewt> | (the first step in navigation) |
| 14:13 | <zewt> | so maybe it doesn't even go that way and I missed something earlier in click() |
| 14:16 | <jgraham> | zewt: I assume it happens in step 21 of the navigation algorithm |
| 14:19 | <jgraham> | It has to via navigation. Also I'm not really sure what effect releasing the storage mutex would have, so I haven't even thought about that |
| 14:19 | <jgraham> | Does anyone even implement it? |
| 14:19 | <zewt> | not for localStorage as far as I know, not sure about cookies or anything else |
| 14:20 | <zewt> | afaik the storage mutex is never intended to be released synchronously once it's taken |
| 14:20 | <zewt> | (not that I'm the right person to ask or anything) |
| 14:20 | <jgraham> | But navigate is never *called* asynchronously |
| 14:20 | <jgraham> | So that isn't just a problem in this case if it is a problem |
| 14:20 | <zewt> | oversight, maybe? dunno |
| 14:20 | <jgraham> | It is typically called synchronously and become asynchronous later |
| 14:22 | <matjas> | so if http://wiki.whatwg.org/wiki/Web_ECMAScript becomes a spec, should it be named “Web ECMAScript”, “Web JavaScript”, or something else? |
| 14:22 | <matjas> | (I kinda like “Web JavaScript”) |
| 14:22 | <zewt> | i suppose if we end up stuck with that ugly "oneTimeOnly" blob URL thing, then the time a URL is released would be the same as whenever a javascript: URL is evaluated |
| 14:23 | <zewt> | (such a broken API though, hopefully it'll go the autoRelease way) |
| 14:23 | <zewt> | (not Blob, the MS oneTimeOnly thing) |
| 14:27 | <jgraham> | I guess Hixie is still not around |
| 14:44 | <zewt> | google's "hide this site's results" hack is annoyingly broken |
| 14:45 | <zewt> | for some reason it won't let you hide a site without viewing it first, and then that doesn't work half the time |
| 14:45 | <zewt> | so instead of blocking a site, i just end up with "you've viewed this site 5 times" which will probably make it *more* likely to return that site |
| 14:53 | <Ms2ger> | MikeSmith, btw, if you like setting up repos for tests... I've got some typedarray and webgl tests that could use a better place than my bitbucket :) |
| 14:54 | <MikeSmith> | Ms2ger: I don't mind setting them up but I do like making sure we have a common place to share them |
| 14:54 | <MikeSmith> | so I am happy to create more repos |
| 14:55 | <MikeSmith> | I would rather not be in the setting-up-repos business |
| 14:55 | <Ms2ger> | Heh :) |
| 14:55 | <MikeSmith> | IMHO would be far better to use the W3C area at Github, and make it all self-serve there for anybody who we trust |
| 14:56 | <MikeSmith> | but some crazy people prefer Bitbucket over Github |
| 14:56 | <Ms2ger> | That would involve git, so I don't think that's better :) |
| 14:56 | <MikeSmith> | hah |
| 14:56 | <MikeSmith> | the W3C systems team also is not religious about having everything under the sun hosted at W3C and would be happy to have stuff at Github |
| 14:57 | <Ms2ger> | Well, I'm even less religious about hosting anything at the W3C :) |
| 14:57 | <MikeSmith> | heh |
| 14:57 | <Ms2ger> | But the automated rsync to w3c-test.org is nice |
| 14:58 | <MikeSmith> | well, some on the team seem are still a bit more wedded to everything needed to be hosted at W3C |
| 14:59 | <Ms2ger> | Given that they're paid to host things at the W3C, I can see their point :) |
| 15:01 | <Ms2ger> | matjas, if you care for the html testing tf meeting, that's now :) |
| 15:01 | <MikeSmith> | systems team is not paid for it really, and I think would be quite happy to have less stuff to have to maintain, and less people complaining to them and asking for enhancements |
| 15:02 | <jgraham> | I am quite in favour of just using github. I think the tooling there is poor, but better than what we have with hg |
| 15:03 | <MikeSmith> | yeah |
| 15:04 | <MikeSmith> | what we have with the current W3C hg/dvcs.w3.org is ... minimal |
| 15:18 | <jscheel> | hi everyone. Was there ever any progress made with disabling css cascading? http://lists.whatwg.org/pipermail/whatwg-whatwg.org/2006-December/008787.html |
| 15:19 | <jscheel> | it seems like this issue comes up every few years, but nobody ever has a good solution (outside of frames) |
| 15:19 | <jscheel> | *iframes |
| 15:23 | <zcorpan> | jscheel: the situation there hasn't changed afaik |
| 15:23 | <jscheel> | zcorpan: sad :( the suggestions in that proposal were great |
| 15:24 | <jscheel> | especially cascade:off |
| 15:25 | <jscheel> | I think another interesting, albeit potentially difficult, solution would be to allow link elements inside other elements in the body, and limit them to that node and it's children |
| 15:25 | <jscheel> | good luck getting that going in the next 20 years though, right :) |
| 15:26 | <zcorpan> | well we have scroped stylesheets already |
| 15:26 | <zcorpan> | so you can do <div><style scoped>@import "foo";</style>...</div> |
| 15:27 | <zcorpan> | but global styles still apply to that div as normal |
| 15:28 | <jscheel> | zcorpan: I didn't know that, interesting. But, globals still apply, huh? |
| 15:28 | <jscheel> | zcorpan: it's a huge issue when you allow user-submitted styles, and the first thing they do is import a global reset |
| 15:32 | <jscheel> | zcorpan: oh, is scoped actually supported anywerhe? |
| 15:32 | <jscheel> | nevermind |
| 15:32 | <jscheel> | no support at all |
| 15:32 | <jscheel> | lamesauce |
| 15:37 | <zcorpan> | jscheel: i see chrome has something in about:flags |
| 15:37 | <jscheel> | zcorpan: yeah, was just looking at that |
| 15:38 | <jscheel> | zcorpan: and a ployfill that seems to work ok |
| 15:38 | <jscheel> | https://github.com/thingsinjars/jQuery-Scoped-CSS-plugin |
| 15:38 | Ms2ger | hijacks w3c-test.org |
| 15:39 | <Ms2ger> | (When running tests in Gecko, at least) |
| 15:42 | <jgraham> | On your testing DNS? |
| 15:47 | <Ms2ger> | The proxy-something we have, yes |
| 15:48 | <matjas> | Ms2ger: damn, I was afk… also I don’t know in which channel those meetings take place |
| 15:52 | <Ms2ger> | #HTMLT on irc.w3.org |
| 15:56 | <zcorpan> | reading specs is hard, let's go shopping. https://bugs.webkit.org/show_bug.cgi?id=89313 |
| 16:02 | <jgraham> | That's quite the fail |
| 16:02 | <zcorpan> | http://bl.ocks.org/2941604 seems to mix v.nu errors and traditional validator errors, which can make it hard to conclude things |
| 16:04 | <zewt> | heh chuckling at the first |
| 16:04 | <zewt> | one of the sillier "requirements" of HTML |
| 16:05 | <zewt> | that and requiring <title>, which I don't get at all |
| 16:06 | <zewt> | (the only thing preventing simply "<!doctype html>" from validating) |
| 16:08 | <zcorpan> | missing <title> isn't a common error |
| 16:09 | <jgraham> | It is in helper files for my tests :) |
| 16:09 | <AryehGregor> | Or, say, iframe contents. |
| 16:09 | <AryehGregor> | Perfectly legitimate there. |
| 16:09 | <jgraham> | (often files were the whole content is "<!doctype html>1" or something) |
| 16:09 | <AryehGregor> | I think I filed a bug once, but it got WONTFIXed. |
| 16:10 | <jgraham> | (these are almost always loaded in iframes) |
| 16:11 | <zcorpan> | yeah well we shouldn't optimize validation for test cases. :-) but iframes in general may be a fair point |
| 16:27 | <zcorpan> | matjas: i think it was jgraham who wrote those |
| 16:29 | <matjas> | Ms2ger: ah, next time! |
| 16:30 | <Ms2ger> | You don't miss much :) |
| 16:30 | <Ms2ger> | Except for http://w3c-test.org/html/tests/submission/Ms2ger/events/event-handler-javascript.html maybe :) |
| 16:31 | <matjas> | mind = blown |
| 16:31 | <zcorpan> | matjas: did you see http://krijnhoetmer.nl/irc-logs/whatwg/20120619#l-640 ? |
| 16:31 | <matjas> | so that‘s how those things can work |
| 16:31 | zcorpan | needs to run |
| 16:32 | <matjas> | zcorpan, ah no, i had missed that one |
| 16:33 | <matjas> | gsnedders: my identifier tests are available here as well: mathias.html5.org/tests/javascript/identifiers/ works in any shell if you just replace text() with console.log() / print() |
| 16:34 | <matjas> | gsnedders: isn’t the whole idea of using testharness.js for all tests that you can replace testharness.js with something that works in your shell, copy the <script> content of the HTML test case, and run it anywhere? |
| 16:35 | <matjas> | (honest question; i assumed this was the case) |
| 16:39 | <gsnedders> | matjas: Nope. Most tests rely on the DOM and shells have no DOM, so it's certainly not true in general. |
| 16:40 | <matjas> | gsnedders: of course… but I don’t see how that is a problem for the web-ecmascript tests |
| 16:41 | <matjas> | gsnedders: what do you suggest? i can provide the tests in a different format / using a different harness if you prefer |
| 16:44 | <gsnedders> | matjas: I dunno, all of us JS engine people have our own fairly different test harnesses :P |
| 16:47 | <Ms2ger> | gsnedders, let's standardize on testharness.js :) |
| 16:50 | <gsnedders> | Ms2ger: I want something that works in JS shells, damn it! |
| 16:50 | <Ms2ger> | gsnedders, let's work on making that happen? |
| 16:50 | <gsnedders> | And the lack of any async stuff makes that impossible to fully-implement the API of. |
| 16:50 | <Ms2ger> | Can you reorder that sentence in a way I understand? |
| 16:52 | <gsnedders> | testharness.js has support for async tests. JS shells (or at least the normal ones) don't allow for multiple threads of execution, and hence don't allow anything to be async. As a result, you can't implement the current testharness.js API for a JS shell. |
| 16:53 | <Ms2ger> | Well, we don't need async tests in the shell, then, do we? |
| 16:53 | <gsnedders> | No. But given the sync tests are just impl as async tests, we may get in to edge-case differences. |
| 16:54 | <gsnedders> | I mean, I'm not saying this isn't doable… |
| 16:54 | <Ms2ger> | Ah, right |
| 16:55 | <gsnedders> | It might well be doable. But you may well end up with subtle API differences. |
| 16:57 | <gsnedders> | Anyone read The Art of Readable code? Worth getting? |
| 16:58 | <Ms2ger> | There's some setTimeouts for timeouts, but I think it's probably fine |
| 17:00 | gsnedders | has scarcely looked at testharness.js since jgraham first asked him to review it two years ago :) |
| 17:01 | <Ms2ger> | There's some things like assuming that `document` exists, but I don't think you'd need much architectural work |
| 17:01 | <Ms2ger> | Anyway, I don't care about shells, so if you want it... :) |
| 17:06 | <gsnedders> | This involves effort! :) |
| 17:06 | <matjas> | My kingdom for cross-spec reference support in http://pimpmyspec.net/ |
| 17:07 | <Ms2ger> | jgraham, interested in a kingdom? |
| 17:25 | <tantek> | what's the opinion in these circles about view-mode? http://www.w3.org/TR/2012/REC-view-mode-20120619/ |
| 17:25 | <tantek> | annevk - I notice it specifically has a fullscreen feature: http://www.w3.org/TR/2012/REC-view-mode-20120619/#dfn-fullscreen |
| 18:10 | <matjas> | is PSP file upload broken? the exact same source document works fine over HTTP |
| 18:11 | <matjas> | <type 'exceptions.ValueError'>: Input object has no element: lxml.etree._ElementTree // args = (u'Input object has no element: lxml.etree._ElementTree',) // message = u'Input object has no element: lxml.etree._ElementTree' |
| 19:42 | <erichynds_> | in CORS, can someone explain what an "origin-list-or-null" value is? http://tools.ietf.org/html/rfc6454#section-7.1 |
| 19:42 | <erichynds_> | trying to figure out the correct way to define multiple origin hosts |
| 19:42 | <erichynds_> | webkit doesn't like space separated, but it works in firefox |
| 19:42 | <Ms2ger> | origin-or-null = %x6E %x75 %x6C %x6C / serialized-origin |
| 19:43 | <erichynds_> | can you explain what that means? |
| 19:43 | <Ms2ger> | Oh, "origin-list-or-null" |
| 19:43 | <Ms2ger> | Don't do that |
| 19:43 | <Ms2ger> | http://lists.w3.org/Archives/Public/www-archive/2012Jun/0001.html |
| 19:43 | <Ms2ger> | annevk, if you read the logs, you may want to add to the wiki |
| 19:44 | <erichynds_> | ah ha. so annevk is saying that it's a single origin or null? |
| 19:44 | <Ms2ger> | I think so |
| 19:44 | <erichynds_> | got it. |
| 19:44 | <erichynds_> | thank you |
| 19:45 | <Ms2ger> | Np |
| 19:54 | <Ms2ger> | https://twitter.com/idiot/status/215143944953135106 |
| 20:16 | <jscheel> | does anyone know if ie9 will support the css3 "initial" property value? |
| 20:32 | <zcorpan> | matjas: how about "JavaScript"? |
| 20:54 | <annevk> | Ms2ger: add what to the logs? |
| 20:55 | <Ms2ger> | To the IETF wiki page |
| 20:55 | <annevk> | Ms2ger: seems you haven't read it |
| 20:55 | <Ms2ger> | No :) |
| 20:55 | <annevk> | curious though |
| 20:56 | <annevk> | I wonder what erichynds_ thought Firefox supports |
