| 00:07 | <Hixie> | are filenames always ltr? |
| 00:07 | <Hixie> | or are paths in hebrew systems (e.g.) rtl? |
| 00:07 | <Hixie> | or is it per-component? |
| 00:07 | <Hixie> | or what? |
| 00:07 | <Hixie> | anyone know? |
| 02:31 | <Yuhong> | OT, but XP is going out of support in one year. |
| 02:32 | <Yuhong> | I wonder how long Firefox and Chrome will support it afterwards. |
| 06:15 | <zcorpan> | should FormData be clone-able (in structured clone)? |
| 08:33 | <annevk5> | zcorpan: so nothing needs to be said for the document global environment? |
| 08:33 | <annevk5> | zcorpan: seems so weird |
| 08:33 | <zcorpan> | annevk: the default is that things are exposed in document global environment and not exposed in worker global environment |
| 08:34 | <annevk> | is that actually explicit somehow? |
| 08:35 | <zcorpan> | "Unless otherwise specified, each ECMAScript global environment exposes all interfaces and exceptions that the implementation supports." webidl |
| 08:35 | <annevk> | I see. Seems rather ugly. |
| 08:36 | <zcorpan> | workers spec says "The DOM APIs — specifically, all the interfaces defind in the DOM specification other than exceptions — must not be exposed if the JavaScript global environment is a worker environment. [DOM]" |
| 08:37 | <annevk> | Should it not say: "Unless stated otherwise, no interfaces must be exposed."? |
| 08:37 | <zcorpan> | though i don't understand why the langauge for constructors is different |
| 08:38 | <zcorpan> | yeah, that seems clearer |
| 08:52 | <annevk> | zcorpan: I think FormData, URL, and URLQuery should probably be able to pass the boundary |
| 08:52 | <annevk> | I guess I should make the worker requirement in URL too. And in Encoding. |
| 08:55 | <annevk> | So structured cloning is an example of which is not defined in a modular way. The ideal way of doing that is to say that objects can have an associated structured clone algorithm and that if they don't you throw a DataCloneError. |
| 08:57 | <zcorpan> | structured clone has an extension point though |
| 08:57 | <zcorpan> | If input is an object that another specification defines how to clone |
| 08:57 | <zcorpan> | Let output be a clone of the object as defined by the other specification. |
| 08:57 | <annevk> | I know |
| 08:58 | <annevk> | But you don't want to define it in that way because that makes matters inconsistent. |
| 08:58 | <jgraham> | You need to define how to write a structured clone algorithm for an object though |
| 08:59 | <jgraham> | And also restrictions on the output of such an algorithm, perhaps |
| 08:59 | <annevk> | That seems orthogonal to how to structure this though. |
| 09:00 | <jgraham> | Well, not quite |
| 09:00 | <jgraham> | I mean I agree that your organisation sounds better, but I don't quite see how the details would fit together |
| 09:00 | <annevk> | The way it's done now favors certain objects over others and makes it harder to understand what you need to define if you define a new object. |
| 09:01 | <annevk> | If all objects instead had an indication of how you'd obtain a structured clone (and whether or not they're exposed in workers) it'd be much clearer for the next person who comes along and defines an object. It also allows more modular development. |
| 09:16 | <matjas> | Hixie, typo in http://html5.org/tools/web-apps-tracker?from=7796&to=7797: “scriptingy” |
| 09:48 | <annevk> | Open up the HTTP WG list archives for the first time in a while. Open http://lists.w3.org/Archives/Public/ietf-http-wg/2013AprJun/0001.html Get trolled hard. |
| 09:59 | <Ms2ger> | "I respectively request" |
| 10:02 | <jgraham> | Is that technically trolling? I mean it's a rather direct homage to "A Modest Proposal", which is a rather tired trope these days. |
| 10:03 | <marcosc> | anyone got an invite to get on this "Inter-net"? |
| 10:03 | <marcosc> | is that like twitters? |
| 10:07 | <annevk> | Ok guys, now look at the date ;-) |
| 10:08 | <marcosc> | :) |
| 10:08 | <jgraham> | I don't see how it being an April Fool changes the fact that it's been done to death :) |
| 12:30 | <asmodai> | I find myself in this wonderful area of HTML, CSS, images and email. I want to slash my wrists. |
| 12:34 | <Ms2ger> | Does anyone implement <track> already? |
| 12:35 | <zcorpan> | Ms2ger: opera does |
| 12:36 | <Ms2ger> | Presto, you mean? |
| 12:36 | <zcorpan> | yes |
| 12:36 | <jgraham> | And also Weblinkit, I think? |
| 12:36 | <zcorpan> | webkit and IE i think |
| 12:36 | <Ms2ger> | I see |
| 12:36 | <zcorpan> | dunno if they've shipped |
| 12:37 | <zcorpan> | we submitted tests for track and webvtt |
| 12:38 | <zcorpan> | though the spec has changed a bit |
| 12:38 | <Ms2ger> | Hmm |
| 12:38 | <tomasf> | http://www.longtailvideo.com/html5/track/ |
| 12:38 | <zcorpan> | e.g. there's now a WebVTTCue constructor |
| 12:38 | <Ms2ger> | Is there any reason the first argument to HTMLMediaElement.addTextTrack isn't an enum? |
| 12:39 | <zcorpan> | no |
| 12:40 | <Ms2ger> | Lemme file that, then |
| 12:42 | <zcorpan> | interface TextTrack : EventTarget { |
| 12:42 | <zcorpan> | readonly attribute DOMString kind; |
| 12:42 | <zcorpan> | i guess that could also use enum? |
| 12:45 | <zcorpan> | did webvtt ever overhaul its rendering rules? |
| 12:46 | <Ms2ger> | Looks like it |
| 12:46 | <Ms2ger> | (enum, not rendering) |
| 12:48 | zcorpan | waves |
| 12:53 | marcosc | wishes that it was mandatory for F2F meeting hosts to provide a bed for when discussions drag on. Would benefit from a quick power nap. |
| 12:53 | <Ms2ger> | sysapps? |
| 12:53 | <marcosc> | yep |
| 12:55 | <Ms2ger> | Hrm |
| 12:55 | <Ms2ger> | Anyone who can find something about when track elements should load their source? |
| 12:57 | <darobin> | hmmm, ISTR there was a bug about that |
| 12:59 | <jgraham> | A bed, or one per participant? Because I'm not sure that the natural repression that comes from being British would allow me to be comfortable sharing a bed with the rest of a WG. |
| 13:01 | <marcosc> | jgraham: good point... it might cause cultural issues. I don't mind. |
| 13:25 | <karlcow> | marcosc, jgraham: hmm it really depends on who is in the bed with you. But I wonder if a consensus could be reached around something like this http://122.img.pp.sohu.com/images/blog/2008/1/20/15/29/1183327596f.jpg |
| 13:28 | <jgraham> | darobin: You know that the github API allows you to test push hooks without making an actual commit? |
| 13:29 | <marcosc> | karlcow: :) |
| 13:35 | <darobin> | jgraham: you know what, I found that option the very second I made the commit :) |
| 13:56 | <odinho> | Art has process questions for reviews of testsuites: https://github.com/w3c/web-platform-tests/pull/62#issuecomment-16171131 |
| 14:05 | <odinho> | darobin, jgraham, Ms2ger, tobie: So, getting w3c-test.org to also carry the submission branches. |
| 14:05 | <odinho> | WDYT? |
| 14:05 | <odinho> | (+)MikeSmith too |
| 14:06 | <Ms2ger> | sgtm |
| 14:06 | <odinho> | Right now there's master and CR, there could be a "submissions" folder where every submission branch is checked out as well. |
| 14:06 | <Ms2ger> | MikeSmith, darobin, tobie: set it up, pls |
| 14:08 | <darobin> | odinho: you write the code to make it happen, taking into account that anything called variations on .php must be excluded, and I'll happily set it up :) |
| 14:12 | <odinho> | Bah :P The webstandards team in Opera should hire me, with all the work I'm doing for them. |
| 14:13 | <odinho> | I'm very unsure if I'll be able to do that in a timely fashion, -- but step one would be to put out what you use now on GitHub or something. |
| 14:14 | <Ms2ger> | odinho, I believe Mozilla's hiring ;) |
| 14:15 | <odinho> | Ms2ger: Sneaky man ;-P |
| 14:15 | <Ms2ger> | :) |
| 14:19 | marcosc | found a picture of Ms2ger |
| 14:19 | <marcosc> | http://www.3gpp.org/IMG/png/businessman.png |
| 14:22 | <darobin> | odinho: what we use on GitHub? you mean to pull on a regular basis? |
| 14:23 | <darobin> | odinho: here's the full code, */10 * * * * cd /u/www.w3c-test.org/web-platform-tests/master && git pull --quiet" |
| 14:24 | <odinho> | Oh. |
| 14:25 | <odinho> | Some serious programming there. Better share with the world ;-) |
| 14:25 | <jgraham> | So it breaks forever if someone makes a non-ff push? :) |
| 14:25 | <darobin> | it probably does! |
| 14:26 | <darobin> | I'm right now improving that though |
| 14:26 | <darobin> | but that won't make it react to PRs |
| 14:26 | <jgraham> | Nope |
| 14:31 | <jgraham> | making it react in some way to pull requests is not hard |
| 14:32 | <jgraham> | Makign it do the right thing is a bit harder, perhaps |
| 14:33 | <jgraham> | https://gist.github.com/kennethreitz/3709868 is quite helpful though |
| 14:35 | <jgraham> | Basically it feels like you should be able to do the same as you are implementing for master, but with the pr branch, the checkout in a submissions/ folder, and apache configured to not run PHP under submissions/ |
| 14:36 | <darobin> | jgraham: I thought of apache not running anything in there, but it's less useful that way |
| 14:36 | <darobin> | you could have a legit submission interacting with legit *existing* php that hasn't been modified |
| 14:36 | <darobin> | but that might be too complex, at least at first |
| 14:37 | <darobin> | that's a useful gist, it's not all that we need but it's a good brick |
| 14:38 | <darobin> | in general I wouldn't mind a way of cleaning up closed PRs, too |
| 14:39 | <jgraham> | darobin: There is an event for closed |
| 14:40 | <jgraham> | Just use that to delete the directory |
| 14:40 | <jgraham> | Why wouldn't existing php run? |
| 14:40 | <jgraham> | If a test links to /foo/bar.php it will work fine. It's only a problem if it tries to use a relative URL |
| 14:41 | <darobin> | aren't those always relative? |
| 14:42 | <darobin> | any link to PHP outside the repo will work |
| 14:42 | <darobin> | links to PHP inside the repo though not |
| 14:42 | <jgraham> | Well inside/outside is a bit misleading |
| 14:43 | <jgraham> | If you link to /resources/useful.php things will work fine |
| 14:43 | <jgraham> | If you try to link to ../../resources/useful.php things will break |
| 14:43 | <jgraham> | Both of those are "inside the repo" in some sense |
| 14:43 | <darobin> | I hate it when you get philosophical |
| 14:44 | <darobin> | but yes, that's what I meant |
| 14:44 | <darobin> | anything in /resources is safe anyway since it's in another repo to boot |
| 14:45 | <odinho> | Anything outside the apache no-go-zone will run. |
| 14:45 | <darobin> | but I was thinking of stuff like JS tests side by side with PHP |
| 14:45 | <darobin> | odinho: I understand that |
| 14:45 | <odinho> | darobin: Yeah, -- I was creating a sentence. |
| 14:45 | <darobin> | but let's say we're testing the unicorn protocol |
| 14:45 | <darobin> | which is a well-known part of HTML |
| 14:45 | <odinho> | Like Server Sent Events, or CORS. |
| 14:45 | <darobin> | so we have html/cryptozoology/unicorn/ |
| 14:46 | <darobin> | in there there's protocol.php and api.html which has some JS |
| 14:46 | <darobin> | I want to submit some new tests: I'm only going to link to "protocol.php" not "/web-platform-tests/html/cryptozoology/unicorn/protocol.php" |
| 14:46 | <jgraham> | Well of course we could do something like git diff --name-only git `merge-base` HEAD | xargs 'chmod -x' |
| 14:46 | <darobin> | that wouldnt' run in a PR |
| 14:46 | <jgraham> | Or whatever the right syntax is |
| 14:47 | <darobin> | it's not a big deal, I was just pointing that out |
| 14:47 | <darobin> | I think we can actually live with turning processing off for PRs |
| 14:48 | <jgraham> | (does php actually need a executable bit set to run under mod_php? Maybe not. Something more compelx could be needed) |
| 14:48 | <odinho> | It doesn't need it no. But it could maybe be configured to work like that. |
| 15:52 | <dglazkov> | good morning, Whatwg! |
| 16:12 | <gsnedders> | https://github.com/html5lib/html5lib-tests/issues/2 in case anyone wants to argue over the tokenizer file format. |
| 17:18 | <gsnedders> | hsivonen, abarth: https://github.com/html5lib/html5lib-tests/issues/2 |
| 17:18 | <jwalden> | MikeSmith: that Kissinger quote from yesterday's pretty funny |
| 17:54 | <abarth> | gsnedders: will read in a bit |
| 18:34 | <zcorpan> | Ms2ger: http://www.whatwg.org/specs/web-apps/current-work/multipage/the-video-element.html#sourcing-out-of-band-text-tracks - "When a text track corresponding to a track element experiences any of the following circumstances, the user agent must start the track processing model for that text track and its track element:" |
| 20:21 | <smaug____> | oh, daily dglazkov bugmail spam :p |
| 20:22 | smaug____ | is perhaps following too many bugzilla components |
| 21:00 | dglazkov | is drowning in bugs |
| 21:11 | <smaug____> | aren't we all? |
