| 01:32 | <JonathanNeal> | Hixie: tried digital ocean? |
| 01:59 | <Hixie> | JonathanNeal: what's digital ocean? |
| 01:59 | <Hixie> | oh, another provider |
| 02:00 | <Hixie> | JonathanNeal: i'm happy with the value for money i get at dreamhost |
| 02:17 | <JonathanNeal> | Hixie: I'm at an impasse. I love dreamhost support, but I love the control and performance of digital ocean. |
| 02:33 | <JonathanNeal> | I didn't know that the whatwg spec allows for multiple main elements. Am I getting that right? |
| 02:51 | <Hixie> | JonathanNeal: yup, it's like class=main basically |
| 02:51 | <Hixie> | or role=main in aria |
| 02:51 | <Hixie> | it gives the main content of something |
| 02:51 | <Hixie> | as opposed to the header, footer, nav, etc |
| 04:28 | <JonathanNeal> | Hixie: what did <main> solve? Does the absence of <main>, in theory, imply that all content of a document or sectioning element is the main content? |
| 04:28 | <Hixie> | <main> doesn't solve anything. |
| 04:28 | <Hixie> | it's pointless. |
| 04:36 | <JonathanNeal> | Then why is it in the WHATWG spec? |
| 04:44 | <Hixie> | browsers implement it, so it has to be in the parser |
| 04:44 | <Hixie> | once it's in the parser, *shrug*, it's like <samp> |
| 04:44 | <Hixie> | not much point to it, but no point making it non-conforming. |
| 04:44 | <JonathanNeal> | Wow, that's all it takes to force the issue? Ouch. |
| 05:03 | <Hixie> | JonathanNeal: well it's all it takes to force the issue on the browser conformance criteria, certainly |
| 05:03 | <Hixie> | JonathanNeal: on authoring conformance criteria, i just try to make the spec be the most helpful given what browsers do |
| 05:04 | <Hixie> | JonathanNeal: so if you want to characterise that as "force", sure |
| 05:05 | <JonathanNeal> | Is there a version of the spec that only shows "the good parts" and hides the conformance stuff? |
| 05:06 | <Hixie> | developer.whatwg.org ? |
| 05:06 | <Hixie> | developers.whatwg.org sorry |
| 05:10 | <GPHemsley> | In case anyone's wondering, I just committed a change to anolis that sorts attributes alphabetically |
| 05:12 | Hixie | tries to work out what he did that broke his pipeline |
| 05:12 | <Hixie> | i'm getting an error in my index preprocessor that has zero to do with anything i changed. wtf. |
| 05:40 | <GPHemsley> | Hixie: That's not related to my anolis change, is it? |
| 05:40 | <GPHemsley> | The mimesniff spec is once again open for business, BTW. |
| 05:40 | <Hixie> | no, it works fine with the earlier version of the html spec |
| 05:40 | <Hixie> | excellent to hear about mimesniff |
| 05:41 | <GPHemsley> | (Now that I've fixed anolis, I can get back to committing without worrying that all my attributes are gonna rearrange themselves randomly) |
| 05:41 | <GPHemsley> | Yeah, it seems to have finally attracted some notice... |
| 05:43 | <GPHemsley> | Hixie: I'm tracking that <video> bug, but I haven't seen anything that warrants any changes on my end. If something comes up, make sure I realize it. (Calling me out by name in the bug would be enough.) |
| 05:44 | <Hixie> | well, the relevant algorithm isn't finished |
| 05:44 | <Hixie> | other than that, i'm not aware of any needed changes yet |
| 05:44 | <GPHemsley> | ok |
| 05:44 | <GPHemsley> | feel free to point out the parts that need finishing |
| 05:45 | <GPHemsley> | in the meantime, though, I think I'm gonna go to bed |
| 05:46 | <Hixie> | the parts are in red :-) |
| 06:30 | <zcorpan> | cors's a rec? woot |
| 06:31 | <zcorpan> | annevk-cloud: second rec, is it? |
| 06:34 | <wirepair> | hsivonen: quick question, why does nu.validator.htmlparser.test.TokenPrinter go totally bonkers when encountering scripts that contain < >? |
| 06:35 | <wirepair> | http://pastie.org/private/cloijixbsxsaxh5h3uxfza <-- example |
| 06:35 | <wirepair> | do i need to tell the driver to ignore script data? |
| 06:37 | <wirepair> | (and if so, how)? |
| 06:38 | <JonathanNeal> | Here's a blog I'm working on addressing the subject of subheadings in W3C's HTML spec, http://www.jonathantneal.com/blog/introducing-subhead/ password: review |
| 06:40 | <Hixie> | JonathanNeal: you give the w3c way too much credit :-) |
| 06:40 | <Hixie> | JonathanNeal: most of what you describe as "you" was done in the whatwg, much of it long before the w3c came along |
| 06:41 | <Hixie> | JonathanNeal: not to mention that <hgroup> doesn't force a grouping pattern, so it works fine :-) |
| 06:43 | <JonathanNeal> | It does, on <h1-6>, as does the entire outline algorithm. |
| 06:44 | <Hixie> | the html4 outline algorithm forces grouping? what? |
| 06:44 | <Hixie> | i don't understand the problems you're trying to solve |
| 06:44 | <Hixie> | what's a page hgroup doesn't work on? |
| 06:45 | <JonathanNeal> | The new outline algorithm; <section>, <aside>, etc. And I would argue that <hgroup> is most "convoluted" example of it. |
| 06:45 | <Hixie> | the outlining of the new sectioning elements, if you ignore the html4 stuff, is literally just nested elements. |
| 06:45 | <Hixie> | there's not much to it. |
| 06:45 | <Hixie> | what's convoluted about <hgroup>? |
| 06:45 | Hixie | is baffled by this conversation |
| 06:47 | <JonathanNeal> | <hgroup> does not allow child elements like <a> or <span>, which is too strict. |
| 06:47 | <Hixie> | do you have a sample page where i can look at what you're trying to do? |
| 06:48 | <JonathanNeal> | Outside of <hgroup>, things like <h3>Lord of the Rings</h3><h2>The Two Towers</h2> mean something very different. |
| 06:48 | <Hixie> | yes, that's why we have hgroup... |
| 06:49 | <JonathanNeal> | Hixie: when I was at Liferay, our CMS let users dynamically add a heading and slogan, and when we used <hgroup> it meant developers couldn't stick presentational elements inside it without adding them one of the inner headings themselves. |
| 06:50 | <Hixie> | so, regardless of what we do with hgroup, you know that all the presentational elements are no longer conforming, right... |
| 06:50 | <JonathanNeal> | <span>s? |
| 06:51 | <Hixie> | can you paste an example of what you want to do? |
| 06:54 | <JonathanNeal> | https://gist.github.com/jonathantneal/8469464 |
| 06:56 | <Hixie> | that shouldn't be conforming even if you replace the <hgroup> with <div>s and the <hx>s with <p>s |
| 06:56 | <Hixie> | use CSS for that kind of thing |
| 07:06 | <JonathanNeal> | In the example I provided, I agree with you. To the more general point, <subhead> can be a child or sibling of a <h1-6>, which is more flexible. Sorry if I'm starting to repeat our last discussion. |
| 07:07 | <Hixie> | do you have an example that shows what you want to do with <hgroup> that you can't do that you agree you should be able to do? |
| 07:10 | <JonathanNeal> | Probably none that you haven't disputed already. |
| 07:11 | <Hixie> | i'm only looking for ones _you_ think are legit, not for ones _i_ think are legit |
| 07:15 | <JonathanNeal> | Any subheading that is best expressed as a child of the heading. <h1>Dr. Strangelove <subhead>or: ...</subhead></h1> |
| 07:16 | <Hixie> | in what sense is that "best expressed as a child"? seems entirely equivalent to <h1>Do Strangelove</h1> <h2>or: ...</h2> |
| 07:19 | <JonathanNeal> | That subheading is intrinsically inline. |
| 07:20 | <Hixie> | hm, actually, looks like it's not a subheading |
| 07:20 | <Hixie> | the title is just that long |
| 07:21 | <MikeSmith> | bingo |
| 07:22 | <MikeSmith> | in that case at least |
| 07:22 | <MikeSmith> | it's just a title with a colon in it |
| 07:22 | <JonathanNeal> | Hi Mike! |
| 07:22 | <MikeSmith> | hey man |
| 07:24 | <MikeSmith> | JonathanNeal: fwiw, to me subheadings have always seemed to be something that end users don't care much about or are not even aware of |
| 07:25 | <hsivonen> | wirepair: I don't know. I haven't look at TokenPrinter in years |
| 07:25 | <JonathanNeal> | MikeSmith: there's a dangerous, grey, probably swerving and overlapping line between clarity and verbosity. |
| 07:26 | <MikeSmith> | JonathanNeal: how so? To end users they're rendered the same regardless of how you mark them up |
| 07:27 | <MikeSmith> | I mean you can put a line break into the title wherever you choose |
| 07:28 | <JonathanNeal> | I could try to semantically distinguish bylines from headings (eg https://gist.github.com/jonathantneal/7818198#project-gutenberg ) but the visual styling will mean more to most users than the elements behind them. I might argue the world doesn't *need* <abbr> (it can be accomplished with a span and title="") or <h2-6> (could just have easily had <h id=""> |
| 07:28 | <JonathanNeal> | and <h for="">). |
| 07:29 | <MikeSmith> | or you can just use a colon if you want, which in practice has the meaning "the part after this colon is the subheading|-title" |
| 07:29 | <MikeSmith> | JonathanNeal: sure |
| 07:29 | <wirepair> | hsivonen: hrm, ok |
| 07:30 | <wirepair> | i'll look at ff source to figure out why they don't consider it a start of a new tag |
| 07:30 | <JonathanNeal> | And with ever advancing CSS selectors, it could be all the easier to style. However, it can be nice to be clear at when you intend those breaks. "Mission: Impossible" |
| 07:30 | <MikeSmith> | JonathanNeal: like all such things, I guess it comes down to how you choose to spend your authoring time |
| 07:31 | <MikeSmith> | and yeah there are always exceptions but the exceptions are almost always common sense |
| 07:32 | <JonathanNeal> | Right. To my original point, if W3C thinks <hgroup> is too restrictive (which I sympathize with as seen in our above discussion) then I thought to write that "letter to a friend" blog post. |
| 07:32 | <MikeSmith> | I mean, nobody thinks "IMpossible" in that example is a subheading |
| 07:32 | <JonathanNeal> | WHATWG saw the need (first) and contributed their solution. |
| 07:32 | <MikeSmith> | JonathanNeal: yeah fair enough |
| 07:33 | <Hixie> | JonathanNeal: personally i still don't really see needs that aren't met by hgroup. You could easily do hgroup > { display: inline } if you really needed it. |
| 07:33 | <Hixie> | JonathanNeal: i'm more worried about your letter causing further forking. |
| 07:38 | <JonathanNeal> | That's a valid concern, to which my must assert an opinion: <hgroup> died long before W3C killed it. Bruce Lawson complained about it back in 2010. It wasn't the cowpath that won over developers. As a result, there is a void, and <subhead> is a solution that addresses the specific reasons <hgroup> was rejected. To that end, can you think of reasons <subhead> |
| 07:38 | <JonathanNeal> | is worse than using <hgroup>? |
| 07:39 | <JonathanNeal> | "my must assert" ... seriously, my apologies for such poor writing. |
| 07:40 | <Hixie> | hgroup is widely used, widely implemented, and mentioned in many tutorials. how is it dead? |
| 07:41 | <Hixie> | the only thing that's slowed use of hgroup is the FUD over it causing people being confused about whether they can use it or not |
| 07:41 | <JonathanNeal> | I've found ONE so far on Google, and the bottom paragraph is rather telling http://webdesign.about.com/od/html5tutorials/a/use-hgroup-element.htm |
| 07:43 | <Hixie> | the "controversy" one? |
| 07:43 | <JonathanNeal> | I'm looking for one from the last twelve months that recommends it. |
| 07:43 | <Hixie> | in the last 12 months people will have avoided recommending it because of the nonsense about it being deprecated |
| 07:44 | <zcorpan> | i recommend that JonathanNeal uses <hgroup>! |
| 07:44 | <zcorpan> | there, mere seconds ago |
| 07:45 | <Hixie> | https://developer.mozilla.org/en-US/docs/Web/HTML/Element/hgroup was last updated in 2013 |
| 07:45 | <Hixie> | http://www.html5tutorial4u.com/hgroup-element.htm is (c) 2014, dunno when last updated |
| 07:46 | <JonathanNeal> | My requirements were probably loaded. Sorry. When I said "that recommends it" I think your example mentioning "refrain from using it" disqualifies it. |
| 07:46 | <JonathanNeal> | But who would write a tutorial and ignore the controversy, so I concede that it should be impossible to find a tutorial that recommends it since W3C removed it, if only because they removed it. |
| 07:47 | <Hixie> | well yeah, everyone who tries to be independent is going to say "but be wary kids, the w3c has dropped it!" |
| 07:48 | <Hixie> | http://developers.whatwg.org/ recommends it and was updated within the last month or so |
| 07:49 | <JonathanNeal> | My opinion™: Before it was officially removed, it was shrouded in controversy, and before that it was often complained about, loudly by (i'm guessing) Lawson and Falkner. |
| 07:49 | <Hixie> | i never heard of any controversy before it was dropped, other than from faulkner. |
| 07:49 | <Hixie> | at least, not any more than the same background complaining we get about everything |
| 07:50 | <JonathanNeal> | Tangent, "developers" is beautiful. I am so glad that was done. I love it, and just want to affirm it again. |
| 07:50 | <Hixie> | if you want to see controversy, you should see e.g. the complaints on appcache. |
| 07:50 | <Hixie> | now _that_ is controversy |
| 07:51 | <JonathanNeal> | what's appcache? you mean manifest? *runs* |
| 07:51 | <MikeSmith> | face punching |
| 07:51 | <Hixie> | appcache is the feature that includes manifest="", yes |
| 07:51 | <JonathanNeal> | If it wanted to be part of html5 so badly, why wasn't it written in XML? RSS knew better. |
| 07:52 | <JonathanNeal> | Sorry, I was running, but MikeSmith nailed me in the face. |
| 07:54 | <MikeSmith> | JonathanNeal: no I geuss the appcache hater guy did accidently |
| 07:54 | <MikeSmith> | or maybe he wants to punch subheadings in the face too |
| 07:55 | <MikeSmith> | I think he should really save his punches for the DOM |
| 07:55 | <zcorpan> | yeah careful with inventing new stuff, your face might become a target for punches |
| 07:56 | <zcorpan> | subpunch in the <subhead> |
| 07:57 | <MikeSmith> | I bet this guy probabably doesn't pay too much attention to CSS yet |
| 07:57 | <JonathanNeal> | oh... |
| 07:57 | JonathanNeal | puts aside the suggestion to drop <, >, and allow css selectors as elements. |
| 07:57 | <MikeSmith> | or maybe he tried to but then as soon as he started he keeled over and died right away from anger |
| 07:58 | <MikeSmith> | before he could unleash the punches |
| 07:59 | <MikeSmith> | anyway in other news I guess Domenic_ is afk by now |
| 07:59 | <MikeSmith> | would really like to chat with him some about Streams stuff |
| 08:00 | <MikeSmith> | and/or marcosc |
| 08:06 | MikeSmith | wonders if foolip's full-time job is now demolition |
| 08:10 | <MikeSmith> | Hixie if you're still awake didn't Philip`s multipage script used to correct fragment references so that they get rewritten to the right URL? |
| 08:10 | <MikeSmith> | redirected |
| 08:10 | <Hixie> | yeah is it not working? |
| 08:11 | <MikeSmith> | Hixie: unless I'm misunderstanding it doesn't seem to be working now |
| 08:11 | <MikeSmith> | e.g., http://www.whatwg.org/htm/#event-0 |
| 08:11 | <Hixie> | damnit |
| 08:11 | <Ms2ger> | MikeSmith, html? |
| 08:11 | <MikeSmith> | yeah |
| 08:11 | <MikeSmith> | meant html |
| 08:11 | <Ms2ger> | That used to work for me... |
| 08:11 | <MikeSmith> | http://www.whatwg.org/html/#event-0 |
| 08:12 | <MikeSmith> | http://www.whatwg.org/specs/web-apps/current-work/multipage/#event-0 |
| 08:12 | <MikeSmith> | should go to http://www.whatwg.org/specs/web-apps/current-work/multipage/section-index.html#events-0 |
| 08:12 | <MikeSmith> | but doesn't |
| 08:13 | <Hixie> | it seems the multipage script on anne's server occasionally returns incomplete data |
| 08:13 | <MikeSmith> | oh |
| 08:14 | <Ms2ger> | http://www.whatwg.org/html/#dom-blur wfm |
| 08:14 | <MikeSmith> | hmm yeah http://www.whatwg.org/html/#event for example works |
| 08:14 | <MikeSmith> | Ms2ger: yeah it's just some that don't |
| 08:14 | <Hixie> | i just regenned |
| 08:14 | <MikeSmith> | maybe newer ones |
| 08:16 | <MikeSmith> | Hixie: http://www.whatwg.org/specs/web-apps/current-work/multipage/#event-0 still not working afaikt |
| 08:16 | <Hixie> | is there an id="event-0" anywhere? |
| 08:17 | <MikeSmith> | yeah |
| 08:17 | <MikeSmith> | oh shit |
| 08:17 | <MikeSmith> | sorry |
| 08:17 | <MikeSmith> | events-0 |
| 08:18 | <Ms2ger> | Well, that seems like a good reason for it not to work :) |
| 08:18 | <MikeSmith> | somebody should punch me in the face |
| 08:19 | <MikeSmith> | sorry Hixie |
| 08:19 | <MikeSmith> | for the noise |
| 08:19 | Ms2ger | gives MikeSmith a beer instead |
| 08:19 | <MikeSmith> | heh |
| 08:21 | <Hixie> | nn |
| 08:23 | <Ms2ger> | nn |
| 08:24 | <Ms2ger> | TabAtkins, have you filed a bug for the textContent special case? I think we discussed in in Gecko but decided against |
| 08:32 | <foolip> | MikeSmith: not really, but it's hard to stop! |
| 08:33 | <Ms2ger> | foolip! |
| 08:33 | <Ms2ger> | I had something for you to demolish |
| 08:33 | <Ms2ger> | It's an outstanding test review |
| 08:34 | <foolip> | Ms2ger: show it to me! |
| 08:34 | <Ms2ger> | But I don't remember which |
| 08:34 | <Ms2ger> | So do them all :) |
| 08:34 | <zcorpan> | Hixie: you might be able to recognize that something in-band is a text track without supporting it, and expose the raw data to JS, or some such |
| 08:34 | <sangwhan> | I'm assuming no - but has Window.close() / Window.open() ever been standardized? |
| 08:35 | <MikeSmith> | foolip: cool to see all that stuff getting cleaned up |
| 08:35 | <Ms2ger> | Yes? |
| 08:35 | <Ms2ger> | They're in HTML |
| 08:35 | <foolip> | Ms2ger: can I see which ones are assigned to me? I assume it had something to do with video or track? |
| 08:35 | <Ms2ger> | foolip, look if https://critic.hoppipolla.co.uk/dashboard has anything? |
| 08:36 | <Ms2ger> | But yes, probably along those lines |
| 08:36 | <foolip> | Ms2ger: sure, I don't often look there :) |
| 08:36 | <foolip> | Ms2ger: all I can find is https://critic.hoppipolla.co.uk/r/426 |
| 08:36 | <foolip> | but that's my review, not something for me to review... |
| 08:36 | <Ms2ger> | Oh, right, that was it, actually |
| 08:37 | Ms2ger | blames his memory |
| 08:37 | <sangwhan> | Ms2ger: Sorry, think I got the wording wrong - "implementation wise standardized" |
| 08:37 | <Ms2ger> | Is the word you're looking for "interoperable"? |
| 08:38 | <foolip> | Ms2ger: so, I was waiting for zcorpan, but he doesn't remember really |
| 08:38 | <foolip> | I'm going to drop that review and revisit the next time I try to get it working in Blink I think |
| 08:38 | <sangwhan> | Ms2ger: Yes. |
| 08:39 | <Ms2ger> | foolip, wfm |
| 08:40 | <Ms2ger> | foolip, and feel free to look at https://critic.hoppipolla.co.uk/r/74 if you have time too :) |
| 08:40 | <Ms2ger> | And / or add a filter for the media dirs |
| 08:40 | <foolip> | Ms2ger: ah yes, the one I'm waiting for. what can I do to move it along? |
| 08:41 | <foolip> | looks like people are waiting for ... something? |
| 08:41 | <Ms2ger> | Someone to review it :) |
| 08:41 | <Ms2ger> | Unless you wrote those tests, I guess |
| 08:41 | <foolip> | only a few of them |
| 08:42 | <foolip> | I'll have a look then |
| 08:42 | <Ms2ger> | Great, thank you! |
| 08:43 | <davve> | Ms2ger: Do you have a reference to where you decided against the textContent special case? |
| 08:47 | <Ms2ger> | Probably https://bugzilla.mozilla.org/show_bug.cgi?id=725221 or https://bugzilla.mozilla.org/show_bug.cgi?id=404385 |
| 08:52 | <foolip> | Ms2ger: do we really want to move and fix the tests at the same time? seems like that'll be suboptimal for history digging |
| 08:54 | <Ms2ger> | Mm |
| 08:54 | <Ms2ger> | The moving is the only real opportunity to fix them, though |
| 08:55 | <Ms2ger> | And I guess it won't be such a problem if we don't squash? |
| 08:55 | <sangwhan> | I'm a bit curious what the rationale for "or if it is a browsing context whose session history contains only one Document." in window.close() is |
| 08:56 | <sangwhan> | Now that everyone and his dog has a WebUI landing page, only contexts with the page that contains window.close() opened by the default protocol handler is closable due to this, in case the implementation is compliant |
| 08:56 | <marcosc> | MikeSmith: what do you want to know about streams? |
| 08:56 | <Ms2ger> | Hi marcosc |
| 08:57 | <sangwhan> | ...and Gecko ships with dom.allow_scripts_to_close_windows as false |
| 08:57 | <marcosc> | Ms2ger: hallo! |
| 08:57 | <Ms2ger> | I bet there's test reviews you could do :) |
| 08:57 | <marcosc> | Ms2ger, maybe? |
| 08:57 | <sangwhan> | Safari ignores it, period. IE seems to return something when window.close() is called unlike everyone else? |
| 08:57 | <foolip> | Ms2ger: ah right, of course, just keep the fixes separate |
| 08:58 | <foolip> | Ms2ger: so what's the scope of fixing to be done? conversion to testharness.js I guess, what else? |
| 08:59 | <Ms2ger> | foolip, wait, do any not use testharness.js? |
| 08:59 | <foolip> | Ms2ger: I think lots of them don't, but I haven't checked just now |
| 09:00 | <foolip> | but if they don't, that needs fixing, yes? |
| 09:00 | <zcorpan> | foolip: that PR is basically just the <track> tests, not all of opera's media tests |
| 09:01 | <zcorpan> | so they're either testharness or reftests |
| 09:01 | <Ms2ger> | Yes, if they don't, they do need fixing, but it looks like zcorpan is correct |
| 09:02 | <zcorpan> | dumping all of the media tests hasn't happened yet i think |
| 09:03 | Ms2ger | heads off for a bit |
| 09:03 | <foolip> | zcorpan: oh |
| 09:04 | <foolip> | zcorpan: can you rebase the move so that it can be merged cleanly to master? I'd like to have it merged locally and try actually running the tests as I review them |
| 09:05 | <foolip> | zcorpan: did your submission/Opera/media-resource-selection branch ever get anywhere? |
| 09:05 | <zcorpan> | foolip: i see some .htaccess which needs to be converted to wpt-serve equivalent (seems like it just sets no-cache, dunno if that's necessary with wpt-serve) |
| 09:06 | <zcorpan> | foolip: is it a matter of clicking "prepare rebase" in critic? |
| 09:06 | <foolip> | zcorpan: no, you have to actually resolve the conflicts as well :) |
| 09:07 | <zcorpan> | bah :-P |
| 09:10 | <zcorpan> | is http://robots.thoughtbot.com/keeping-a-github-fork-updated a set of steps i should follow? |
| 09:14 | <Ms2ger> | zcorpan, no, the prepare rebase button doesn't work here |
| 09:14 | <Ms2ger> | You need to rebase to master without making other changes, push, and then a new button appears |
| 09:14 | <Ms2ger> | Or ask jgraham :) |
| 09:14 | <sangwhan> | Ms2ger: Or just manually create a new review if all fails |
| 09:15 | <Ms2ger> | I'd prefer not :) |
| 09:15 | <zcorpan> | jgraham: i need to do something in git, please help |
| 09:17 | <SteveF> | faulkner says: if you wanna smoke hixie's weed pop a <hgroup> in your pipe |
| 09:18 | Ms2ger | says: I'm bored of hgroup talk already |
| 09:18 | <SteveF> | Ms2ger: better if you smoke it |
| 09:18 | <Ms2ger> | Feel free to |
| 09:19 | <SteveF> | FUDlicious |
| 09:20 | <zcorpan> | it might be dangerous to smoke a pipe in case you get punched in the face |
| 09:21 | <SteveF> | might be dangerous to punch someone smoking a knife pipe |
| 09:23 | zcorpan | successfully followed these steps https://help.github.com/articles/syncing-a-fork |
| 09:24 | <zcorpan> | now rebase the branch... |
| 09:37 | <sangwhan> | Hixie: Is the rationale behind window.close() only working when there is one document in the session due to abuse concerns? |
| 09:38 | <MikeSmith> | marcosc: for starters it would nice to have a short description of what the fundamental difference is between Domenic_ current draft and the WebApps draft |
| 09:39 | <MikeSmith> | would be helpful to me personally for understanding and I would think helpful to others who are trying to follow along at home |
| 09:39 | <zcorpan> | someone(TM) should go through MDN and fix all "DOM Level 0. Not part of any standard." |
| 09:40 | sangwhan | mysqldump and sed on MDN time, what could possibly go wrong |
| 09:41 | <MikeSmith> | sangwhan: Hixie is afk |
| 09:41 | <sangwhan> | MikeSmith: Aha |
| 09:42 | <MikeSmith> | sangwhan: you should ping him later if you're up late, or just wait til morning your time |
| 09:42 | <sangwhan> | MikeSmith: I'll try either of those, depending on how long I end up throwing around IPC messages |
| 09:49 | <MikeSmith> | zcorpan: if you're rebasing a branch that critic is tracking, critic is not going to be happy |
| 09:49 | <MikeSmith> | nor jgraham |
| 09:49 | <MikeSmith> | since he'll have to un-fugg critic after |
| 09:49 | <MikeSmith> | as far as in my experience at least |
| 09:50 | <zcorpan> | but the PR can't be merged cleanly now anyway |
| 09:50 | <zcorpan> | because some files moved and wpt-serve changed some things |
| 09:51 | <zcorpan> | but i'll wait for jgraham's advice |
| 09:55 | <marcosc> | MikeSmith: Domenic_, Feras, Takeshi and I are having a call to converge the specs next week |
| 09:55 | <marcosc> | if all goes to plan, we should only have a single spec |
| 10:01 | <Ms2ger> | And then a fork, amirite |
| 10:02 | <marcosc> | youisrite |
| 10:02 | <jgraham> | I can unfugg critic |
| 10:02 | <jgraham> | And indeed you can too, although you need to be a little bit careful |
| 10:03 | <Ms2ger> | zcorpan, the git master has arrived :) |
| 10:03 | <zcorpan> | seems like foolip is doing the heavy lifting for me |
| 10:04 | <jgraham> | On critic? |
| 10:04 | <foolip> | jgraham: I'm rebasingand resolving the conflicts |
| 10:06 | <jgraham> | foolip: OK. Do you know how to tell critic about the rebase for a tracking branch? |
| 10:12 | <foolip> | jgraham: no, and I think you're going to have to tell zcorpan, I just prepared the rebased branch: https://github.com/foolip/web-platform-tests/tree/submission/Opera/media |
| 10:12 | <foolip> | jgraham: I guess only simon can add commits to his review, yes? |
| 10:14 | <jgraham> | Yeah, due to the "magic of github" this simple process is really difficult |
| 10:16 | <jgraham> | Anyway, it is not too hard |
| 10:16 | <jgraham> | After pushing you press "rebase branch |
| 10:17 | <jgraham> | " |
| 10:17 | <jgraham> | And then in "upstream" where it says "refs/heads/master" you put the commit that you actually rebased onto |
| 10:17 | <jgraham> | Then you press fetch branch |
| 10:17 | <jgraham> | and hopefully it's all OK |
| 10:17 | <foolip> | zcorpan: when reviewing, should I raise issues for failing tests, or can we add failing tests? |
| 10:18 | <zcorpan> | foolip: we can add failing tests |
| 10:18 | <jgraham> | foolip: From a general repo policy failing tests are OK |
| 10:18 | <jgraham> | We only demand correctness |
| 10:18 | <jgraham> | Working out how to deal with actually running the tests is an implementor problem |
| 10:19 | <foolip> | zcorpan: so as long as they run and the failures aren't mysterious to me I'll mark it as reviewed, and try to fix them when importing to Blink, sound good? |
| 10:19 | <zcorpan> | foolip: sounds good |
| 10:19 | <zcorpan> | jgraham: were those instructions for me? is that a button in critic or github? |
| 10:20 | <Ms2ger> | Critic |
| 10:20 | <zcorpan> | i don't see "rebase branch" in critic (i see "rebase review" and "prepare rebase") |
| 10:21 | <jgraham> | s/branch/review/ then |
| 10:21 | <Ms2ger> | jgraham, rebase branch, I guess? |
| 10:22 | <jgraham> | "prepare rebase" is the one that doesn't work here |
| 10:22 | <jgraham> | It really can't be hard to hide that button… |
| 10:24 | <zcorpan> | ok so what do i put in "upstream"? |
| 10:24 | <jgraham> | 10:17 < jgraham> And then in "upstream" where it says "refs/heads/master" you put the commit that you actually rebased onto |
| 10:25 | <jgraham> | So whatever the first commit that isn't part of the branch is according to, say, git log |
| 10:27 | <zcorpan> | cfcfc6d5194040a9addb1c1004613db9af6d8075 ? |
| 10:28 | <jgraham> | foolip: ^ |
| 10:29 | <foolip> | zcorpan: yes! |
| 10:30 | <zcorpan> | how does critic become aware of https://github.com/foolip/web-platform-tests/tree/submission/Opera/media ? is that the next step? |
| 10:31 | <foolip> | zcorpan: you need to update your own branch to match mine first I gather |
| 10:31 | <foolip> | it's time for weekend for me now, I hope zcorpan and jgraham prevail! |
| 10:31 | <zcorpan> | have a nice one, and thanks for the help :-) |
| 10:32 | <foolip> | no problem, I want these tests! |
| 10:32 | <zcorpan> | so first i cherry-pick his commit and push, and then "rebase review" in critic? |
| 10:33 | <foolip> | zcorpan: first check out your own branch |
| 10:33 | <jgraham> | Pull his whole branch and push that, I guess |
| 10:33 | <foolip> | then git fetch https://github.com/foolip/web-platform-tests.git submission/Opera/media |
| 10:33 | <foolip> | then git reset --hard FETCH_HEAD |
| 10:33 | <foolip> | and then push -f |
| 10:33 | <foolip> | is my best guess |
| 10:34 | <foolip> | put you can of course just push my branch directly, but then your local branch willl be out of sync |
| 10:34 | <foolip> | that's it, I'm really going |
| 10:34 | <foolip> | poof |
| 10:43 | <zcorpan> | i have no upstream set on this branch. git push -f --set-upstream origin submission/Opera/media ? |
| 10:43 | <Ms2ger> | jgraham? |
| 10:47 | <jgraham> | zcorpan: The upstream in this case is just whatever the base commit of the branch is i.e. the first commit in history that isn't part of the branch (assuming no merges) So if you have a history like -M1-M2-M3-B1-B2-B3 where the M commits are part of master and the B commits are on the branch, it would be M3 |
| 10:49 | <zcorpan> | jgraham: are you talking about the "upstream" input box in critic or the error i got when trying `git push -f` (fatal: The current branch submission/Opera/media has no upstream branch. ) ? |
| 10:50 | <jgraham> | Oh, I was talking about the one on critic |
| 10:51 | <jgraham> | In this case you just want to git push -f --set-upstream} {name of your gh remote} HEAD:{name of the branch on github} |
| 10:52 | <jgraham> | {name of the branch on github} === submission/Opera/media |
| 10:53 | <zcorpan> | {name of your gh remote} = ? |
| 10:54 | <Ms2ger> | Probably origin |
| 10:54 | <zcorpan> | ah |
| 10:55 | <jgraham> | git remote -vv and look for the one that points to your GH remote |
| 10:56 | <zcorpan> | origin https://github.com/zcorpan/web-platform-tests.git (fetch) |
| 10:56 | <zcorpan> | origin https://github.com/zcorpan/web-platform-tests.git (push) |
| 10:56 | <zcorpan> | upstream https://github.com/w3c/web-platform-tests.git (fetch) |
| 10:56 | <zcorpan> | upstream https://github.com/w3c/web-platform-tests.git (push) |
| 10:56 | <annevk> | GPHemsley: if you're around, is there anything still unclear with respect to the image/svg+xml thing? |
| 10:56 | <Ms2ger> | So indeed origin |
| 10:57 | <zcorpan> | git push -f --set-upstream origin HEAD:submission/Opera/media |
| 11:00 | <jgraham> | Yes |
| 11:00 | <zcorpan> | ok done |
| 11:01 | <jgraham> | Do you want me to poke critic? |
| 11:01 | <zcorpan> | now rebase review in critic with cfcfc6d5194040a9addb1c1004613db9af6d8075 as "upstream" right? |
| 11:03 | <jgraham> | Yes |
| 11:04 | <zcorpan> | ok done. now "enable tracking"? |
| 11:04 | <jgraham> | Yes |
| 11:05 | <zcorpan> | hmm. done but it didn't bite it seems |
| 11:05 | <jgraham> | It did, you need to force refresh |
| 11:05 | <jgraham> | I don't know why |
| 11:07 | <jgraham> | OK, going to be afk (or afi) for a bit |
| 11:07 | <zcorpan> | ah, how silly. don't even know how to force refresh in new opera |
| 11:07 | <zcorpan> | thanks for the hlep |
| 11:50 | <Ms2ger> | dglazkov, ping |
| 12:10 | <zcorpan> | annevk: Hixie: can't you just get along? :-P |
| 12:25 | <Ms2ger> | zcorpan, foolip, where's cors-tester.py? |
| 12:26 | <Ms2ger> | Oh, looks like it's in the repo |
| 12:28 | <jgraham> | Seriously hg, where is something like git reset --hard? |
| 12:29 | <Ms2ger> | What does that do? |
| 12:30 | <davve> | jgraham: Tried 'hg revert --all' ? |
| 12:32 | <jgraham> | davve: That only works on the working copy, not the actual repo |
| 12:32 | <jgraham> | I was trying to get rid of a whole bunch of commits that I managed to pull in by accident trying to transfer stuff from one repo to another |
| 12:33 | <Ms2ger> | hg strip first-rev-to-get-rid-of |
| 12:33 | <Ms2ger> | Note: destroys history, use with care |
| 12:34 | <jgraham> | Ms2ger: Yeah, so I ended up using hg strip, it it wasn't really clear what the first rev to get rid of was |
| 12:34 | <jgraham> | Possibly because I also don't know what the equivalent of merge-base is |
| 12:35 | <Ms2ger> | Maybe hg strip -r "outgoing(inbound)"? |
| 12:35 | <jgraham> | Perhaps? |
| 12:35 | <jgraham> | I did 'hg outgoing --template="{node}\n" -q | xargs hg strip' |
| 12:35 | <Ms2ger> | Sounds like it'd do the same |
| 12:36 | <jgraham> | Either way, git reset --hard origin/master would have been easier |
| 12:37 | <Ms2ger> | Because you know that one :) |
| 12:37 | <jgraham> | Well yes and the internet knows that one |
| 12:37 | <jgraham> | Google is not very helpful for hg |
| 12:38 | <jgraham> | It is usually people on stack overflow that know it about as well as I do and MDN |
| 12:38 | <Ms2ger> | http://stackoverflow.com/questions/2139165/mercurial-delete-all-local-changesets-revert-to-tree |
| 12:38 | <Ms2ger> | Is the first hit for "remove local changesets hg" |
| 12:38 | <Ms2ger> | Accepted answer gives hg strip 'roots(outgoing())', which is what I meant |
| 12:39 | <jgraham> | Well when I searched I found a question where the answer was roughly "uh, I dunno, reclone? Or use strip?" |
| 12:39 | <jgraham> | Which I had already worked out... |
| 12:40 | Ms2ger | kicks critic |
| 12:40 | <Ms2ger> | JavaScript error: https://critic.hoppipolla.co.uk/static-resource/changeset.js?mxynzt, line 822: files[file_id] is undefined |
| 12:43 | <Ms2ger> | STR: https://critic.hoppipolla.co.uk/6bf4020f?review=74 click on one of the -4's for the htaccess removal, click fetch deleted lines |
| 12:48 | <annevk> | zcorpan: Hixie and I don't get along? |
| 12:49 | <zcorpan> | annevk: you keep disagreeing about url query encoding |
| 12:50 | <Ms2ger> | Is there a way to have github tell you the PR of some commit? |
| 12:51 | <annevk> | zcorpan: ah yes we do |
| 12:52 | <annevk> | zcorpan: we both have to care for all the other people that most likely do not care at all |
| 12:54 | <zcorpan> | annevk: if you both care but don't come to a conclusion then we're not making progress |
| 12:58 | <annevk> | zcorpan: so it sounds like you care too |
| 12:58 | <annevk> | zcorpan: I added another comment |
| 13:01 | <zcorpan> | i care about coming to a conclusion, less which it is :-) |
| 13:41 | <annevk> | I always attributed 'The answer to "how do I mark up X" is "mu"' to Mark Pilgrim, but he quoted from http://www.propylon.com/news/ctoarticles/Zen_and_the_Art_of_Motorcycle_Manuals_20020822.html which might have quoted from another site I can't access |
| 13:46 | <jgraham> | Zen and the Art of Motorcycle Maintainance is a book, so presumbably this is a reference to that |
| 13:46 | <darobin_> | yeah, I doubt ZAMM would be quoting from a site :) |
| 14:09 | <zcorpan> | i tried the link in web.archive but it just said "ad expired" or some such |
| 14:10 | <annevk> | darobin_: that article has a [1] directly after the quote which is a dead link |
| 14:11 | <darobin_> | annevk: it's been a very long time since I read ZAMM, but it's probably referring to http://en.wikipedia.org/wiki/Mu_%28negative%29 |
| 14:12 | <annevk> | darobin_: well that much is clear to me |
| 14:12 | <darobin_> | annevk: I would bet that the itw.etc site just had a definition of that, from before everything was on Wikipedia |
| 14:14 | <annevk> | darobin_: ait |
| 15:11 | <jgraham> | Where does the spec set document.contentType for parsed documents? |
| 15:13 | <GPHemsley> | annevk: I'm still not seeing the problem in the algorithm. image/svg+xml is always assumed to be the correct type. No sniffing (in the traditional sense) occurs. |
| 15:14 | <GPHemsley> | Ms2ger: I committed a change to anolis last night; I hope that's OK. |
| 15:33 | <annevk> | GPHemsley: oh you generalized from SVG to XML type? http://mimesniff.spec.whatwg.org/#sniffing-in-an-image-context |
| 15:33 | <annevk> | GPHemsley: basically in an image context only image/svg+xml should be recognized as a type, otherwise you must always sniff, iirc |
| 15:34 | <annevk> | GPHemsley: unless nosniff maybe |
| 17:23 | <Hixie> | sangwhan: iirc it's just what browsers do, not sure of motivation but annoyance blocking is likely |
| 17:26 | <jgraham> | Hixie: Do you know where document.contentType is set? |
| 17:26 | <jgraham> | (when parsing) |
| 17:27 | <Hixie> | i would assume it's not set during parsing, but interesting question |
| 17:27 | <jgraham> | Well I mean when constructing a document for the parser to work on |
| 17:27 | <Hixie> | found it |
| 17:28 | <Hixie> | it's set in the "Page load processing model for..." sections |
| 17:28 | <Hixie> | e.g. "When an HTML document is to be loaded in a browsing context, the user agent must queue a task to create a Document object, mark it as being an HTML document, set its content type to "text/html", create an HTML parser, and associate it with the document." |
| 17:28 | <dglazkov> | good morning, Whatwg! |
| 17:28 | <Hixie> | found by searching for the Dependencies section in HTML, DOM subsection, finding "content type", and clicking on that |
| 17:30 | <jgraham> | So that doesn't say what happens for XML files |
| 17:30 | <Hixie> | oh, so it doesn't |
| 17:30 | <Hixie> | how odd |
| 17:30 | <Hixie> | looks like that one section omits it |
| 17:31 | <jgraham> | I started here |
| 17:31 | <jgraham> | http://www.whatwg.org/specs/web-apps/current-work/multipage/the-xhtml-syntax.html#parsing-xhtml-documents |
| 17:31 | <jgraham> | Which didn't really help |
| 17:31 | <Hixie> | oh, i see why |
| 17:31 | <jgraham> | But it does make me wonder when the XML parser can implicitly create a Document and what happens in that case |
| 17:31 | <Hixie> | i don't actually create the document for XML documents, nor parse it |
| 17:31 | <Hixie> | because that's supposed to be defined in the XML specs |
| 17:31 | <Hixie> | "When faced with displaying an XML file inline, user agents must first create a Document object, following the requirements of the XML and Namespaces in XML recommendations, RFC 3023, DOM, and other relevant specifications. [XML] [XMLNS] [RFC3023] [DOM]" |
| 17:32 | <Hixie> | hence later statements like "If the root element, as parsed according to the XML specifications cited above..." |
| 17:32 | <Hixie> | i can add a big warning there about how this isn't actually defined yet |
| 17:32 | <Hixie> | if you like |
| 17:33 | <Hixie> | or we can just ignore it on the assumption that xml is going to die soon enough anyway (that's at least what abarth and crew keep threatening) |
| 17:34 | <jgraham> | Well given the current text in DOM, you would be fogiven for thinking that would create an application/xml contentType always |
| 17:34 | <jgraham> | *forgiven |
| 17:34 | <Hixie> | it's not just contentType that doesn't get initialised here |
| 17:34 | <Hixie> | there's literally no requirement to invoke the XML parser |
| 17:34 | <Hixie> | (also, no definition of what an XML parser is) |
| 17:34 | <Hixie> | (nor how the XML spec maps to DOM) |
| 17:36 | <jgraham> | Interestingly "create a Document Object" doesn't seem to actually create a document object |
| 17:36 | <Hixie> | right |
| 17:36 | <Hixie> | that's just saying what should happen when the XML spec creates the object |
| 17:36 | <Hixie> | like i said, that section is assuming the XML specs define all this |
| 17:37 | <jgraham> | Well what actually create a HTML document then? |
| 17:37 | <jgraham> | i.e. a text/html Document |
| 17:38 | <Hixie> | the mythical "XML parser and DOM" spec |
| 17:38 | <Hixie> | oh for HTML |
| 17:38 | <Hixie> | well for HTML it doesn't say "create it as defined elsewhere" |
| 17:38 | <Hixie> | it just says "create it" |
| 17:39 | <jgraham> | I see "the user agent must queue a task to _create a Document object_", but following the link we get to an algorithm that assumes that the document has already been created |
| 17:39 | <Hixie> | "creade a Document object" is just a literal statement |
| 17:40 | <Hixie> | i suppose i could make the cross-references less confusing |
| 17:41 | <jgraham> | Well, ideally there would be a step that says "let /document/ be a new Document object" |
| 17:41 | <Hixie> | that's what the four words "create a Document object" mean |
| 17:41 | <Hixie> | it's just confusing because it hyperlinks to the implications thereof |
| 17:41 | <jgraham> | As a hyperlink they mean "do whatever the algorithm at the other end of the link says" |
| 17:41 | <Hixie> | right |
| 17:42 | <jgraham> | Which doesn't include actually creating a document object :) |
| 17:46 | <Hixie> | jgraham: k, let me poke at this |
| 18:00 | <Hixie> | jgraham: alright, look now (singlepage only) |
| 18:01 | <Ms2ger> | GPHemsley, yes, of course, thank you, and sorry for dropping the ball there |
| 18:01 | <Hixie> | jgraham: (uh, one sec) |
| 18:01 | <Ms2ger> | It's always at the end of my todo list :/ |
| 18:01 | <Ms2ger> | dglazkov, ping |
| 18:01 | <Hixie> | jgraham: (ok, ready) |
| 18:05 | <jgraham> | Hixie: lgtm |
| 18:05 | <jgraham> | Thanks |
| 18:05 | <Hixie> | thank _you_! |
| 18:21 | <Ms2ger> | "Add a couple of ImageData() constructors, and (in order to make that feasible) drop all the HD stuff on <canvas>." |
| 18:21 | <Ms2ger> | Woop |
| 18:26 | <Ms2ger> | Hixie, ping |
| 18:27 | <Hixie> | po |
| 18:27 | <Hixie> | er |
| 18:27 | <Hixie> | yo |
| 18:27 | <Hixie> | or pong |
| 18:27 | <Hixie> | whatever |
| 18:27 | <Ms2ger> | Heh |
| 18:27 | <Ms2ger> | https://www.w3.org/Bugs/Public/show_bug.cgi?id=24224 |
| 18:28 | <Ms2ger> | Looks like that changed the behaviour of document.title = "" to not create a text node where it did before |
| 18:28 | <Ms2ger> | Was that intentional? |
| 18:30 | <Hixie> | Ms2ger: it was not |
| 18:31 | <Hixie> | Ms2ger: i was trying to make the mutation observer stuff work |
| 18:31 | <Hixie> | Ms2ger: if browsers do create an empty node, can you reopen the bug and try to explain how i should be doing this? |
| 18:31 | <Hixie> | Ms2ger: i don't fully follow the mutation observer stuff |
| 18:32 | Ms2ger | looks where his test is |
| 18:33 | <Ms2ger> | var title = document.documentElement.firstChild.lastChild.firstChild; |
| 18:33 | <Ms2ger> | assert_true(title, "Need a node."); |
| 18:33 | <Ms2ger> | >.< |
| 18:35 | <Ms2ger> | http://w3c-test.org/web-platform-tests/master/html/dom/documents/dom-tree-accessors/document.title-06.html |
| 18:35 | <Ms2ger> | Gecko returns null |
| 18:35 | <Ms2ger> | Chrome returns a text node |
| 18:36 | Ms2ger | pulls up IE |
| 18:38 | <dglazkov> | Ms2ger: yo |
| 18:38 | <Hixie> | Ms2ger: safari says "assert_true: Need a node. expected true got Text node ""(...)" |
| 18:39 | <Ms2ger> | Yeah, that's because I write buggy tests :) |
| 18:39 | <Ms2ger> | So that matches Chrome and the old spec |
| 18:39 | <Ms2ger> | dglazkov, hey, there's a bunch of outstanding PRs on web-platform-tests for componentsy things, do you know anybody who could review? |
| 18:39 | <Hixie> | Ms2ger: what does IE do? |
| 18:39 | <dglazkov> | Ms2ger: yes, just send them my way. I may have lost track. |
| 18:40 | <dglazkov> | "may have" -> "definitely" |
| 18:40 | <dglazkov> | long vacations are great |
| 18:41 | <Ms2ger> | Hixie, looking |
| 18:41 | <Ms2ger> | IE doesn't load my test |
| 18:42 | <Ms2ger> | (And clocked content due to certificate errors... On msn.com, the default homepage) |
| 18:44 | <Ms2ger> | And it can't load the live dom viewer either |
| 18:44 | <Hixie> | o_O |
| 18:44 | <Ms2ger> | Anyone got IE? :) |
| 18:44 | <Hixie> | i have the worst luck testing IE |
| 18:47 | <Ms2ger> | Getting browserstack results from bz |
| 18:48 | <jgraham> | Pretty sure luck has nothing to do with it |
| 18:48 | <Ms2ger> | IE matches the new spec |
| 18:48 | <jgraham> | If I was at Microsoft I would totally do if username() == "Hixie" {do_crazy_shit()} |
| 18:49 | <Ms2ger> | Hixie, so I guess I'm fine with the change |
| 18:49 | <Hixie> | Ms2ger: k |
| 19:07 | <Ms2ger> | abarth, ping |
| 19:13 | <Ms2ger> | dglazkov, you should add a filter on critic |
| 19:21 | <dglazkov> | Ms2ger: teach me? |
| 19:23 | <Ms2ger> | Go to https://critic.hoppipolla.co.uk/home |
| 19:23 | <jgraham> | Click "Add Filter" |
| 19:24 | <jgraham> | Select the web-platform-tests repo |
| 19:24 | <jgraham> | Enter the path to web-components |
| 19:24 | <jgraham> | After you are done, make sure that you have entered your email address so that you get mail |
| 19:26 | <Ms2ger> | And the shadow-dom path |
| 19:26 | <Ms2ger> | dglazkov, ^ |
| 19:27 | <Ms2ger> | (Log in with github credentials) |
| 19:28 | <Ms2ger> | foolip, <3 |
| 19:29 | <jgraham> | Woah |
| 19:29 | <jgraham> | I thought foolip was having a weekend |
| 19:29 | <dglazkov> | yay, thanks! |
| 19:31 | <Ms2ger> | dglazkov, does your dashboard have things now? |
| 19:31 | <jgraham> | https://critic.hoppipolla.co.uk/dashboard |
| 19:33 | <dglazkov> | yes, they say "active", but all of them have already been closed, I think? Still clicking... |
| 19:33 | <dglazkov> | I don't get critic I guess |
| 19:33 | <jgraham> | dglazkov: There might be some that are out of sync due to historical bugs |
| 19:35 | <Ms2ger> | Yeah |
| 19:35 | <Ms2ger> | There are still open ones, though... |
| 19:36 | <Ms2ger> | Like https://github.com/w3c/web-platform-tests/pull/173 which doesn't have a critic review |
| 19:36 | <Ms2ger> | jgraham, ^ |
| 19:36 | <Ms2ger> | And https://github.com/w3c/web-platform-tests/pull/178 |
| 19:39 | <jgraham> | dglazkov: I dropped all the ones that were closed |
| 19:42 | <jgraham> | Uh, 173 did have a review until I dropped it just now :) |
| 19:42 | <jgraham> | No comment from critic though |
| 19:43 | <jgraham> | https://critic.hoppipolla.co.uk/r/157 |
| 19:43 | <Ms2ger> | Aha |
| 19:44 | <jgraham> | dglazkov: 5 reviews that would benefit from your attention unless I am miscounting |
| 19:44 | Ms2ger | sees 6 on gh |
| 19:45 | <Ms2ger> | Or 5 |
| 19:45 | Ms2ger | isn't sure he can count |
| 19:45 | <jgraham> | I see 4 now |
| 19:46 | <jgraham> | PR #s 173, 221, 278, 29 |
| 19:46 | <jgraham> | 8 |
| 19:46 | <Ms2ger> | 173 178 184 194 221 278 |
| 19:47 | <jgraham> | Uh, so that's not much overlap :) |
| 19:47 | <jgraham> | Ah, right 178 is the one I just fixed |
| 19:48 | <Ms2ger> | Looks like I counted imports (which should be html-imports) too |
| 19:49 | <jgraham> | OK, all those are in critic |
| 19:49 | <jgraham> | dglazkov: You also need a filter for imports/ |
| 19:50 | <jgraham> | Or maybe not? |
| 19:50 | <jgraham> | Yeah, that doesn't exist |
| 19:50 | <Ms2ger> | Not in master |
| 19:51 | <jgraham> | Oh, I see |
| 19:51 | <jgraham> | Which PR adds it? |
| 19:51 | <Ms2ger> | 178 |
| 19:52 | jgraham | adds dglazkov to the list of reviewers for that review |
| 19:52 | <jgraham> | Ms2ger: Please comment on the incorrect directory name |
| 19:53 | <Ms2ger> | jgraham, tobie did: https://github.com/w3c/web-platform-tests/pull/178#issuecomment-21849903 |
| 19:55 | <jgraham> | OK |
| 19:56 | <Ms2ger> | Does https://github.com/w3c/web-platform-tests/pull/188 have a review? |
| 20:04 | <Ms2ger> | gsnedders, want to write a test for https://www.w3.org/Bugs/Public/show_bug.cgi?id=24233? |
| 20:05 | <gsnedders> | Ms2ger: File a bug on html5lib-tests? |
| 20:06 | <Ms2ger> | gsnedders, want to file a bug on html5lib-tests? |
| 20:06 | <gsnedders> | Ms2ger: no |
| 20:06 | <Ms2ger> | :( |
| 20:12 | <Ms2ger> | gsnedders, do I just add it to the tests##.dat with the highest number? |
| 20:13 | <gsnedders> | Ms2ger: You look for any vaguely similar tests and add it there. |
| 20:14 | <Ms2ger> | gsnedders, do I need to get #errors right or will some fairy do that for me? |
| 20:15 | <gsnedders> | Ms2ger: I typically just throw in whatever html5lib gives. Which given that isn't up to date, and has never been perfect with the errors, probably isn't quite right. |
| 20:16 | <Ms2ger> | How do I get the errors out of html5lib? |
| 20:26 | <Ms2ger> | gsnedders, or r? https://critic.hoppipolla.co.uk/r/577 |
| 20:29 | <Ms2ger> | function successCallback(position) { |
| 20:29 | <Ms2ger> | test(function() {assert_true(true)},"Success Callback called"); |
| 20:40 | <gsnedders> | Ms2ger: Nah, that means reading the spec. :) |
| 20:41 | <gsnedders> | (I really don't feel that well at the moment, so anything that involves more than possibly witty responses on IRC is overdoing it.) |
| 20:43 | <Ms2ger> | Alright |
| 20:45 | <Ms2ger> | "Ms2ger reopened the pull request in a few seconds" |
| 20:45 | <Ms2ger> | Whoa |
| 20:48 | <Ms2ger> | jgraham, https://github.com/w3c/web-platform-tests/pull/160 doesn't seem to get a review |
| 20:50 | <gsnedders> | Ms2ger: Someone badly needs to go through all the spec changes in god-knows-how-long as make sure html5lib has tests to assert the currently correct behaviour :( |
| 20:50 | <Ms2ger> | Not just for html5lib :) |
| 20:50 | <Ms2ger> | I'd like a way to attach tests to svn revisions |
| 20:55 | <gsnedders> | I love Chromium's new issue wizard. |
| 20:56 | <gsnedders> | I'm not sure what the steps to reproduce testsuite updates are. |
| 20:56 | <gsnedders> | 1. Go to the Chromium repo, notice they're different from upstream; 2. Observe the upstream version is newer |
| 21:06 | <gsnedders> | https://code.google.com/p/chromium/issues/detail?id=335691 — you want steps? you get steps! |
| 21:27 | <jamesr__> | gsnedders: i cc'd some folks for you |
| 21:28 | <jamesr__> | not sure what you mean by the LGPL comment |
| 21:28 | <Ms2ger> | jamesr__, everything in the repo is lgpl by default? |
| 21:29 | <jamesr__> | it is? why? |
| 21:29 | <jamesr__> | there is some LGPL code in blink but it's the vast minority |
| 21:29 | <Ms2ger> | I would guess that was gsnedders's assumption |
| 21:30 | <gsnedders> | jamesr__: follow parent directories until you reach a license file. |
| 21:30 | <jamesr__> | gsnedders: what's the one you are hitting? |
| 21:31 | <gsnedders> | jamesr__: Oh, this has changed from last time I did this. It's now nothing until chromium/src/LICENSE, which is 3-clause BSD |
| 21:40 | <gsnedders> | jamesr__: But thanks for CC'ing people |
| 21:47 | <GPHemsley> | annevk: Any other XML-based image type would be treated just the same. You would never sniff an XML type. |
| 21:48 | <GPHemsley> | Ms2ger: That's alright, I dropped it too. But I needed it fixed in order to continue with mimesniff. And it turned out to be a lot easier than I thought. |
| 21:51 | <Hixie> | so https://www.w3.org/Bugs/Public/show_bug.cgi?id=20701#c119 makes every Window and Location have multiple instances |
| 21:51 | <Hixie> | so i can't refer to "the Window object" any more |
| 21:52 | <Hixie> | anyone got any bright ideas on how i do that without having to modify all 481 occurrences of Window in the spec? |
| 21:55 | <GPHemsley> | can you talk about "a non-native Window object" as distinct from "the Window object"? |
| 21:55 | <GPHemsley> | or a "Window-derived object" |
| 21:55 | <GPHemsley> | or what have you |
| 21:55 | <GPHemsley> | (I'm not clear on the full context) |
| 22:00 | <Hixie> | maybe... |
| 22:06 | <Hixie> | though i'm sure a lot of the spec talks about "the Window object of..." which will always have to be special-cased which would be a huge pain |
| 22:09 | <Hixie> | man the assumption that there's just one Window is really hard-baked into the spec |
| 22:19 | <jamesr__> | gsnedders: (i don't know the exact licensing details of these files, but i'd doubt it's LGPL. hopefully one of the cc'd folks can say for real) |
| 22:21 | <jamesr__> | gsnedders: also what is 3-clause BSD incompatible with? |
| 22:29 | <Ms2ger> | html5lib-tests is MIT, fwiw |
| 22:33 | <jamesr__> | right - are those incompatible with each other? |
| 22:34 | <Hixie> | mixing MIT and BSD is fine |
| 22:35 | <Hixie> | so long as you follow the respective licenses |
| 22:35 | <Hixie> | (i.e. include the copyright notice in redistributed stuff, mainly) |
| 22:53 | <jamesr__> | yeah need the notice |
| 23:38 | <jgraham> | If it is 3-clause BSD it would be nice to get it relicensed as MIT for html5lib |
| 23:38 | <jgraham> | Better than having multiple licenses in that repository |