| 00:00 | <gsnedders> | you could try emailing the hotel? |
| 01:16 | <JakeA> | Yeah. Emailed w3, but probably quicker to go straight to the hotel. |
| 04:56 | <annevk> | https://github.com/w3c/html/issues/550 |
| 04:56 | <annevk> | Somehow I anticipated censor |
| 04:56 | <annevk> | https://irccloud.mozilla.com/file/Rkah5CcD/web-platform-wg-censoring.png |
| 06:41 | <mkwst> | annevk: I think Anolis sets a `title` attribute on `<dfn>` tags that matches their ID. Is that something we could change? I think Bikeshed is picking it up as the defined term (/cc TabAtkins, if you're still awake). |
| 06:43 | <mkwst> | That is, Fetch's generated HTML for `request` is `<dfn data-export="" id="concept-request" title="concept-request">request</dfn>`. Looking at the `anchors.data` file Bikeshed downloads, `concept-request` is being used as the defined term. |
| 06:49 | <annevk> | mkwst: in theory we could patch Anolis to use data-x I suppose and replace title with data-x |
| 06:52 | <mkwst> | Hrm. I guess I need to look at the code to understand what it's actually using the title for. If we're going to patch Anolis, I guess we could just teach it to produce some of this metadata (like `data-lt`) itself. |
| 06:55 | <annevk> | mkwst: the title is used to link term usage back to its definition |
| 06:55 | <annevk> | mkwst: and it's used to create the ID |
| 06:55 | <mkwst> | The title on the `<a>`, I understand. The title on the `<dfn>` is strange, given that we have the `id`. |
| 06:56 | <annevk> | mkwst: no, the ID is derived from the title |
| 06:56 | <annevk> | mkwst: but you could have a different ID if you want from the linking term |
| 06:58 | <annevk> | mkwst: I'm not sure what's more work, changing Anolis and the specs that use it, or convert the specs to bikeshed |
| 06:58 | <mkwst> | *sigh* This looked like such an easy change. :) |
| 07:12 | <Ms2ger> | Hahahaha |
| 08:00 | <annevk> | mkwst: you consider https://github.com/whatwg/html/pull/1619 done right? |
| 08:00 | annevk | is about to merge, but thought he'd check |
| 08:00 | <mkwst> | annevk: Yeah, if y'all are happy with it, I am too. |
| 08:01 | <mkwst> | Thanks for working through these patches with me. It feels like they've taken more iterations than usual, so I appreciate your patience. :) |
| 08:02 | <annevk> | No worries, am happy to help |
| 08:18 | <zcorpan> | annevk: mkwst: trying to keep Anolis up to date seems more painful to me. we should stop using it imo |
| 08:19 | <mkwst> | zcorpan: 1. I agree! 2. Porting things like Fetch over seems like a lot of work. :/ |
| 08:19 | <zcorpan> | yeah |
| 08:25 | <mkwst> | What else uses Anolis? XHR? |
| 08:25 | <Ms2ger> | encoding |
| 08:26 | <annevk> | fetch |
| 08:26 | <Ms2ger> | fullscreen |
| 08:26 | <annevk> | fullscreen |
| 08:26 | <annevk> | html-differences |
| 08:26 | <annevk> | mimesniff |
| 08:26 | <annevk> | xhr |
| 08:26 | <annevk> | That's all the directories I have with an Anolis Makefile |
| 08:28 | <mkwst> | More than I thought. |
| 08:29 | <zcorpan> | TabAtkins: would it be possible to develop a script to convert from anolis to bikeshed? ^ |
| 08:32 | <zcorpan> | anyone with Edge? http://software.hixie.ch/utilities/js/live-dom-viewer/saved/4393 |
| 08:42 | <annevk> | I created an issue for mutation events: https://github.com/whatwg/dom/issues/305 |
| 08:42 | <annevk> | zcorpan: sure |
| 08:43 | <mkwst> | zcorpan: What is it supposed to do? I can send you a screenshot? |
| 08:43 | <zcorpan> | I'm interested in what the log says |
| 08:43 | <zcorpan> | screenshot is ok |
| 08:44 | <mkwst> | https://www.irccloud.com/pastebin/cOoVRWdw/ |
| 08:44 | <zcorpan> | thanks! |
| 08:44 | <annevk> | so |
| 08:44 | <annevk> | fwiw |
| 08:44 | <annevk> | my results in Edge 14 are very different |
| 08:44 | <annevk> | but I need to take a screenshot since copy-and-paste doesn't work |
| 08:45 | <annevk> | https://irccloud.mozilla.com/file/6AvEOBd7/edge14.png |
| 08:45 | <annevk> | zcorpan: ^^ |
| 08:46 | <annevk> | mkwst: do you have Edge 13? |
| 08:47 | <zcorpan> | annevk: thanks. that matches the result of IE8..11 also |
| 08:47 | <mkwst> | *ahem* It's possible I ran that in Chrome. :( |
| 08:48 | <zcorpan> | haha |
| 08:48 | <mkwst> | https://www.irccloud.com/pastebin/8OjeXno4/ |
| 08:48 | <mkwst> | That's what I get in Edge 13. Looks pretty similar to Chrome? |
| 08:49 | <zcorpan> | thanks. chrome is different (empty string for most, and act like textContent for option.innerText sometimes) |
| 08:50 | <mkwst> | I see. Well, if Chrome is wrong, file a bug and I'll ask people who know things about things to help. :) |
| 08:50 | <zcorpan> | it seems IE and Edge behave like Gecko does with their latest fix, and chromium is non-sensical |
| 08:51 | <zcorpan> | mkwst: bug is https://bugs.chromium.org/p/chromium/issues/detail?id=636327 (via https://github.com/rocallahan/innerText-spec/issues/5 ) |
| 08:53 | <mkwst> | Ok. Kent is a good guy; I'm sure he'll be thrilled to take a look at your test. |
| 11:01 | <MikeSmith> | annevk: is tel:9052670719 a value URL |
| 11:03 | <annevk> | valid? I think so |
| 11:03 | <MikeSmith> | Galimatias seems to parse the 9052670719 as a port number |
| 11:04 | <annevk> | That seems like a bug |
| 11:04 | <annevk> | Maybe add it to wpt |
| 11:04 | <MikeSmith> | OK |
| 11:06 | <MikeSmith> | filed https://github.com/w3c/web-platform-tests/issues/3480 for now |
| 11:06 | <MikeSmith> | will add the test laster |
| 11:20 | <MikeSmith> | annevk: so I wi will still add a test for it if we do not have one already, but actually the bug I was seeing was caused by the document I was testing with |
| 11:21 | <annevk> | Ok |
| 11:21 | <MikeSmith> | which has href="https://tel:9052670719"; |
| 11:21 | <annevk> | Aaah |
| 11:21 | <annevk> | Yeah that would parse it as port |
| 11:21 | <MikeSmith> | yeah |
| 11:22 | <MikeSmith> | anyway it still was htting a bug in galimatias in that galimatias calls parseInt on that port number part |
| 11:22 | <MikeSmith> | but 9052670719 is bigger than 2^32 or whatever the limit it, so java throws |
| 11:22 | <MikeSmith> | *limit for an int is |
| 12:05 | <Domenic> | zcorpan: here are my Edge 14 results... |
| 12:05 | <Domenic> | https://www.irccloud.com/pastebin/wp99E4RY/ |
| 12:06 | <Domenic> | I thought it was a recent spec change that ports must be < 2^32 |
| 12:06 | <Domenic> | I think > 2^32 is an invalid URL |
| 12:06 | <zcorpan> | Domenic: thanks. are you certain this is not Chrome you tested? :-) |
| 12:07 | <Domenic> | zcorpan: yeah quite sure.. it is a bit strange it's different than annevk's |
| 12:08 | <annevk> | Weird, I use |
| 12:08 | <zcorpan> | http://www.whatsmyua.com/ |
| 12:08 | <zcorpan> | tkent also got "" for select.innerText in Edge per his comment in the chromium issue |
| 12:08 | <annevk> | BrowserStack |
| 12:08 | <Domenic> | Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/51.0.2704.79 Safari/537.36 Edge/14.14393 |
| 12:09 | <MikeSmith> | Domenic: oh about the port number I guess I need to check the current spec then |
| 12:10 | <annevk> | 14.14300 |
| 12:10 | <MikeSmith> | Galimatias seems to not be maintained any longer, and so for the HTML checker I’m using a fork of it with patches I have made for other open issues that have not be fixed upstream |
| 12:13 | <MikeSmith> | ah OK 2^16 yeah |
| 12:13 | <zcorpan> | annevk: thx |
| 12:46 | <zcorpan> | MikeSmith: opinion on "change the content model or rp to "text"?" ? https://github.com/whatwg/html/issues/1679 |
| 12:47 | MikeSmith | looks |
| 12:49 | <MikeSmith> | zcorpan: yeah I think restricting it text would help authors catch potential markup errors better |
| 12:50 | <MikeSmith> | I cannot think of any cases I’ve ever seen where anybody had markup in rp intentionally |
| 12:52 | <MikeSmith> | but I would not be surprised if the let’s-overengineer-all-ruby-things partisans have some they’ve dreamed up |
| 13:23 | <tobie> | Update on the WebIDL spec Bikeshed conversion for those of you interested by this: https://github.com/heycam/webidl/issues/147 |
| 13:30 | <annevk> | once ruby is deployed enough <rp> would basically not be used |
| 13:30 | <annevk> | I'm surprised it got added at all |
| 13:34 | <zcorpan> | maybe innerText can always add ( and ) before and after <rt> and ignore <rp> |
| 13:38 | <annevk> | tobie: \o/ |
| 13:39 | <annevk> | tobie: that is great news |
| 13:45 | <tobie> | annevk: yeah, I'm pretty excited to get people's feedback. |
| 13:52 | <zcorpan> | tobie: looks awesome |
| 13:52 | <annevk> | tobie: it looks great to me too |
| 13:53 | <annevk> | tobie: hopefully heycam|away can sign off on it, seems like you did all the legwork for review too by writing tools to compare the output and such |
| 14:03 | <tobie> | Thanks! |
| 14:03 | <tobie> | I need to change some of the tags so I can do an HTML diff. |
| 14:07 | <tobie> | The differ doesn't like <emu-…> that much. |
| 14:08 | <nox> | Is <dialog> supported, somewhere? |
| 14:09 | <nox> | Spec and WPT disagree. |
| 14:10 | <nox> | Spec says that dialog.close() should abort steps if there is no 'open' attribute; |
| 14:10 | <nox> | but WPT and Chrome both says it should throw an InvalidStateErr. |
| 14:10 | <nox> | So I guess the spec should change? |
| 14:11 | <MikeSmith> | I thought gecko supported dialog too |
| 14:11 | <MikeSmith> | now |
| 14:11 | <MikeSmith> | yay tobie |
| 14:11 | <nox> | Let me try on nightly. |
| 14:11 | <nox> | MikeSmith: Doesn't seem like it. |
| 14:11 | <MikeSmith> | oh |
| 14:12 | <MikeSmith> | hmm yeah https://bugzilla.mozilla.org/show_bug.cgi?id=840640 |
| 14:13 | <nox> | MikeSmith: So should it throw? Why does the spec not throw but WPT and Chrome do? |
| 14:14 | <nox> | Independent decision on Chrome side and they are the tests' authors, maybe? |
| 14:16 | <Ms2ger> | Doesn't Gecko impl it? |
| 14:17 | <MikeSmith> | nox: seems to me it should throw |
| 14:17 | <MikeSmith> | rightly I mean |
| 14:17 | <MikeSmith> | I dunno why the spec says otherwise or when it might have been changed |
| 14:18 | <nox> | Ms2ger: No. |
| 14:18 | <MikeSmith> | but if we have a WPT test for that expects it to throw I would think that must have matched the spec at the time the test was written |
| 14:18 | MikeSmith | wonders who wrote the test |
| 14:19 | <MikeSmith> | 98556c0c715f1406a2f4b5cf332e2fa4e621794a is the spec change |
| 14:19 | <Ms2ger> | nox, https://bugzilla.mozilla.org/show_bug.cgi?id=840640 |
| 14:19 | <MikeSmith> | https://www.w3.org/Bugs/Public/show_bug.cgi?id=25114 |
| 14:19 | <Ms2ger> | Huh |
| 14:20 | <MikeSmith> | zcorpan: ⬆ |
| 14:21 | <MikeSmith> | so the test needs to be fixed |
| 14:34 | <nox> | MikeSmith: Our Servo contributor will do that. |
| 14:34 | <MikeSmith> | super |
| 14:35 | <zcorpan> | nox: i have a fix locally |
| 14:35 | <nox> | zcorpan: Oh. |
| 14:35 | <zcorpan> | assuming it's only html/semantics/interactive-elements/the-dialog-element/dialog-close.html |
| 14:41 | <nox> | zcorpan: Our contributor fix should be better, |
| 14:41 | <nox> | zcorpan: he checked whether return value got altered or nt. |
| 14:41 | <nox> | not* |
| 14:42 | <zcorpan> | nox: ok, cool. let's close my pr then :-) |
| 14:42 | <nox> | :) |
| 14:44 | <zcorpan> | nox: do you know if there is a chromium issue filed already? |
| 14:44 | <nox> | zcorpan: Absolutely not. |
| 14:48 | <zcorpan> | filed https://bugs.chromium.org/p/chromium/issues/detail?id=638943 |
| 15:08 | <annevk> | Wow: https://bugzilla.mozilla.org/show_bug.cgi?id=1296306 |
| 21:06 | <zcorpan> | hmmmm. why is imageTag here, or why does it exist at all? https://cs.chromium.org/chromium/src/third_party/WebKit/Source/core/html/HTMLElement.cpp?sq=package:chromium&dr=CSs&rcl=1471276009&l=157 |
| 21:31 | <KevinMarks> | Gah. Why does subresource integrity use base64 encoding, which is full of non url-safe characters? |
| 21:32 | <KevinMarks> | If I try to pass the hash via a url I'm in escaping hell. |
| 21:57 | <terinjokes> | KevinMarks: it's my favorite thing about the JSON Web Token article on Wikipedia. "The tokens are designed to be compact, URL-safe and usable especially in web browser single sign-on (SSO) context." |
| 21:57 | <terinjokes> | except then they use Base 64 |
| 21:59 | <KevinMarks> | I've standardised on newbase60 for my own things, so was lulled into a false sense of security. |
| 22:02 | <KevinMarks> | Looking at https://github.com/hash-uri/hash-uri/blob/master/README.md but that has slashes in too |
| 22:07 | <terinjokes> | i saw something recently where they decided to use base50 or something |
| 22:54 | <Domenic> | cdumez continues to be the best https://github.com/WebKit/webkit/commit/f203fd07772753d2f73dcce865664a943dc6e4ff |
| 22:58 | <KevinMarks> | Also saw x-object-meta-sha1base36: 1d91dx0894wjewukeyxu56os5uhx4ph |
| 22:58 | <KevinMarks> | Which is a new one on me |
