2016-04-01 [20:24:52.0000] krijn: I wonder if we could get the Web History Project to archive the logs [20:25:14.0000] krijn: and maybe also provide some funds, however modest [20:25:40.0000] krijn: we have 10 years of important history there [20:26:44.0000] I mean other than just whatever is at archive.org [21:58:36.0000] MikeSmith: if I can do anything to help, let me know! [22:18:39.0000] krijn: FYI https://lists.w3.org/Archives/Public/public-webhistory/2016Apr/0000.html [22:30:32.0000] Thanks The Mike Smith :) [22:31:31.0000] I can easily make a static dump of all files and add a redirect to a new location [22:33:15.0000] yeah I am partly interested in just raising awareness about the logs among web-archaeologist people who could benefit from knowing about them [22:33:24.0000] heh yeah I don’t know where that “the” in “the Krijn Hoetmer” crept in from [02:20:59.0000] Lovely: https://twitter.com/LeonieWatson/status/715788733040168960 [02:21:26.0000] TabAtkins, you know why google is wasting money on the fork? [02:22:17.0000] There still isn't a spec which can give a canonical answer to whether a key should fire 'keypress' events or not, right? [02:25:18.0000] hallvors: nope [02:26:35.0000] I'm pretty sure I wrote several tests that could be applied.. [02:26:43.0000] ..to create such a spec [02:27:21.0000] s/applied/useful to sound less like a chatbot or something [02:28:56.0000] hallvors: in theory https://www.w3.org/TR/uievents/ should do it, but they have been opting out of most hard things it seems [02:30:28.0000] Right. I did scan through it.. [02:31:14.0000] Most spec editors seem to fancy adding cruft, specifying existing cruft isn't much loved [02:32:02.0000] That nobody seems to learn that the existing cruft influences the new cruft is a little worrisome [03:26:05.0000] hallvors: You don't sound nearly nazi enough to be a chatbot [03:32:28.0000] jgraham: thanks :) [04:56:47.0000] krijn: btw, krijn.html5.org is still a thing if you need space [05:15:49.0000] annevk: https://github.com/whatwg/url/commit/ae7a0f6f4f0314fa1af163844c33cfdefd83334f Is it intended that URL.href's setter doesn't take into account the URL's base anymore? [05:17:33.0000] Yeah I think so [05:17:58.0000] annevk: Why? [05:18:06.0000] URL.prototype.href should probably just have been a getter [05:18:26.0000] annevk: Oh I see, URL doesn't have a base anymore anyway? [05:18:31.0000] nox: yeah [05:18:38.0000] annevk: Makes sense. :) [06:34:53.0000] annevk: oh right! I could move big part of the archive there of course, and then start logging new stuff. Now remember at some point I put http://krijn.html5.org/irc-logs/ up there already.. [06:38:43.0000] It has all the space you want... [06:44:46.0000] annevk: has it been discussed somewhere how shadow scripts should be executed? [06:45:18.0000] or have we always just think of "like normal script" [06:47:47.0000] That was the conclusion based on some discussion [06:49:49.0000] annevk: will do, thanks! [06:50:12.0000] And then 301 a few years over there [06:51:34.0000] Wouldn't something like irc-logs.html5.org be better by the way? No need to keep my personal branding in there (even though I'm a sucker for all the extra attention throughout the years! ;) [06:52:15.0000] krijn: sure if you want [07:21:56.0000] krijn: can set that up later [07:22:36.0000] krijn: also happy to provide you with money for a new hard drive or some such, or something else you need [07:29:01.0000] annevk: do you think we will ever expose the URLs in the Response url list to script? I'm trying to decide if its worth the storage costs of saving the entire URL list or if we can just save a boolean "redirected" flag [07:31:08.0000] annevk: nm, I see spec note now that it won;t be allowed due "atomic HTTP redirect handling" [07:31:27.0000] wanderview: ever is hard, but not until we have CORS-like protocol for redirects [07:31:39.0000] wanderview: right [07:32:12.0000] annevk: we can start tracking URLs at that point [07:32:16.0000] if its added [07:32:47.0000] wanderview: yeah, makes sense, would require quite some work anyway [07:33:22.0000] annevk: are their other algorithms that look at older URLs in the list? like internal stuff that says "if ever redirected through origin X, do Y"? [07:33:47.0000] my impression is there are not [07:39:20.0000] wanderview: CSP and MIX maybe [07:40:02.0000] wanderview: if main page CSP wants to act on SW response [07:40:22.0000] wanderview: not sure if mkwst has defined the details yet [07:41:35.0000] I guess that would be good to know if its going to be needed [07:41:39.0000] mkwst: any thoughts? [07:42:31.0000] I think that's the reason the spec keeps track [07:45:35.0000] Would anyone know why all elements appear block when you put display: flex on a button in Firefox? [07:45:46.0000] s/elements/child elements [07:47:31.0000] JonathanNeal: #layout at irc.mozilla.org may be able to help more if its firefox specific behavior [07:48:18.0000] Thanks @wanderview, I didn’t know if that’s how flex is supposed to lay things out. [07:48:37.0000] I don't know if it is either, but if its a firefox bug the people in #layout would probably be the ones to talk to about it [07:52:27.0000] annevk: I guess we will store all the URLs in the list... it should be a minor hit given most Responses will still only have one URL [08:22:57.0000] That is a seriously quiet room, @wanderview, but thanks all the same :) [08:23:18.0000] JonathanNeal: ah, sorry... they might be on west coast US time... just waking up [08:23:40.0000] Fair enough. [08:58:30.0000] I think the CSP and MIX integrations are ~clearly defined. [08:58:37.0000] annevk, wanderview ^^ [08:58:55.0000] I haven't looked at MIX in a while, though, so I won't swear on it. [08:59:15.0000] But CSP certainly has a hook at the bottom of Fetch to deal with Responses (and passes in the Request for context). [08:59:28.0000] I haven't implemented that in Chrome yet. It's on a list somewhere. [09:58:13.0000] jyasskin: I was thinking, most permissions probably do need to deal with revocation [09:58:35.0000] jyasskin: e.g., if I revoke "notifications", existing notifications should probably be cleared [09:59:47.0000] jyasskin: and for "persistent-storage", the box mode should be changed and such (I don't want to change the box mode when permission is granted though, I think that should still go through the persist() method) [10:18:36.0000] annevk: So then I wouldn't bother defining "boolean permission", and would just have each permission define its revocation algorithm. Boolean ones would just make that the only algorithm they define. [10:19:33.0000] jyasskin: yeah, I guess null or algorithm makes sense for some of them [10:19:55.0000] annevk: For now, nothing defines its revocation behavior, and I think some permissions need to take the argument into account, but I'll plan for almost every permission needing to do something. [10:20:40.0000] jyasskin: yeah, it seems like revocation should mostly have an effect, e.g., stop watchPosition for "geolocation" [10:40:02.0000] Ooooh https://twitter.com/FxSiteCompat/status/715908743351967744 [10:40:05.0000] No more AppCache API [10:43:37.0000] I hope that works out, saves me a bunch of work fixing the existing stuff [11:02:53.0000] annevk: I think there is some question about the use counter analysis... they are still looking at it [16:14:02.0000] hi, my interpretation of https://fetch.spec.whatwg.org/#http-redirect-fetch is that the fetch spec intends to preserve custom request headers across redirects, especially for same-origin redirects; however, https://bugzilla.mozilla.org/show_bug.cgi?id=216828#c4 implies that this is not the case. Is anybody here familiar with this topic? 2016-04-02 [17:24:06.0000] gwicke: annevk [18:41:06.0000] MikeSmith: thanks, might try again next week [21:36:07.0000] gwicke: bug in Gecko then [04:47:47.0000] gwicke: I reopened the bug [08:00:40.0000] Domenic: I noticed something interesting - `filter` is much harder to implement than `map` with async iterators. [08:01:38.0000] I have all sorts of pyramids, anywhere I can check an existing implementation? [09:49:26.0000] gwicke: turns out wanderview already fixed it elsewhere, such a great person [10:27:17.0000] benjamingr__: async function* filter(ai, cb) { for await (const e of ai) { if (await cb(e)) { yield e; } } } [10:27:42.0000] benjamingr__: can omit the await before cb(e) if you only want sync predicates [10:27:55.0000] do JavaScript-engine implementations actually ever need to have code that directly corresponds to so ES slots, or are slots just a conceptual feature of the spec [10:28:17.0000] and of other specs that define things in terms of ES slots [10:28:31.0000] Domenic: I meant without having async generators - also, I noticed async generators aren't working in babel and they used to - any idea why? [10:28:48.0000] benjamingr__: sure, if you only use half the proposal, using the proposal is hard. [10:28:52.0000] benjamingr__: no idea [10:29:25.0000] MikeSmith: they need to produce all the observable features of the slots existing, so usually the easiest way to implement that is with some kind of slot-like thing (which I know for a fact V8 and JSC have) [10:29:54.0000] Domenic: well, since I'm doing it to understand the proposal - I think it's fair to place such a constraint on myself. [10:30:12.0000] Just like implementing map/filter on iterators without having generators. [10:32:05.0000] benjamingr__: maybe Babel originally implemented the now-withdrawn observable version of async generators and they never got around to implementing the stage 2 version. [10:36:21.0000] Domenic: OK, thanks. And the slot stuff came along in ES6 and after, right? I don’t remember anything about it being in ES5.1 or before. Or maybe there was another name previously? hmm, but if so I don’t know what it would have been. [10:36:48.0000] MikeSmith: no, it was in ES5, just with a more-confusing name "internal properties" I believe. [10:36:58.0000] aha [10:37:03.0000] geez [10:37:04.0000] OK [10:37:17.0000] well clearly I was overthinking this [10:37:19.0000] thanks [10:37:29.0000] (It was confusing since they did not behave like properties at all, e.g. no prototypal inheritance.) [10:37:33.0000] anytime! [10:37:43.0000] yup agreed about the old name [10:38:15.0000] and now that I understand “slots” seems very much the apt name [10:39:15.0000] /me now goes back to reading the spec he was reading before he distracted himself [11:38:46.0000] Domenic: ah, do you have any presentation about that? [11:39:40.0000] I was asked to talk about async iterators and to be frank I'm clueless about where things stand in ES, I can give an hour long talk about async iterators in Python but that'd be mega-lame. I want to show off async generators as an alternative to observables for many use cases and as a concurrency primitive on its own. [12:33:37.0000] I unchecked "allow merge commits" for html [12:42:32.0000] zcorpan: great, we should do that everywhere [12:44:12.0000] i can do that [12:49:07.0000] dnoe [15:18:49.0000] zcorpan++ [15:22:33.0000] annevk: I wonder if we should now update the TEAM.md to drop the instructions about how to merge a PR branch manually [15:23:23.0000] anyway, for now let’s see if this thing actually works [15:23:41.0000] /me pushes the green button on libpackage-deprecationmanager-perl [15:23:46.0000] oofs [15:23:53.0000] /me pushes the green button on libpackage-deprecationmanager-perl [15:23:57.0000] christ [15:25:46.0000] hot diggety dog https://github.com/whatwg/html/commit/78806b0a466dd8f1f04fdcd5bc51ca002ccbf31f 2016-04-03 [19:43:11.0000] https://w3c.github.io/browser-payment-api/specs/paymentrequest.html#user-agent-delegates-payment-request-algorithm has the statement “A user agent may not always be able to abort a request.” which could be misread as RFC 2219 language [19:43:36.0000] any suggestions for language to replace it with? [19:43:56.0000] or just leave it as is? [19:45:35.0000] maybe just “might” [04:00:17.0000] MikeSmith: yeah, a lot can be simplified now [04:00:32.0000] MikeSmith: great that it works 2016-04-04 [00:13:53.0000] MikeSmith: so I just did it as well, it seems a little weird that the option that we haven't enabled (and creates a merge commit) is still selected by default [00:14:15.0000] MikeSmith: probably something GitHub will fix and I'm not the first to notice I guess [00:15:31.0000] yeah [00:15:50.0000] I ran into that too and reckoned the same thing [00:16:03.0000] minor wrinkle that I guess they will get ironed out soon [01:10:11.0000] TabAtkins: another issue that's been awaiting a reply from you: https://www.w3.org/Bugs/Public/show_bug.cgi?id=23893 [01:10:48.0000] TabAtkins: there's quite a few, perhaps you should schedule some time to work on HTML rendering 😊 [03:48:21.0000] What was the way to become op on this channel again? [04:53:14.0000] hallvors: I don't think https://github.com/w3c/clipboard-apis/issues/26 can be left to them [04:53:47.0000] hallvors: in particular, I believe Michiel works on the W3C HTML fork and that's unlikely to be actively maintained for the long term [04:54:14.0000] Heck, it isn't even actively maintained now [04:54:23.0000] annevk: it's more in Selection or Editing land than in HTML though [04:54:35.0000] hallvors: focus isn't [04:54:42.0000] hallvors: and Clipboard isn't either, last I looked [04:55:15.0000] Not? I don't see any particular requirements I need to follow up that don't seem to be better spec'ed as Editing requirements.. [04:55:20.0000] Sorry, DataTransfer [04:55:50.0000] To clarify: I'm not talking about the existing state of things [04:55:53.0000] hallvors: e.g. the bits around focus transfer is very much a thing nothing but HTML defines [04:55:59.0000] but more conceptually what belongs where [04:56:37.0000] hallvors: I'm having a hard time believing that the focus stuff can migrate from HTML to elsewhere anytime soon [04:56:55.0000] why should the Clipboard API spec have to deal with focus transfer stuff? [04:57:17.0000] hallvors: "I have one requirement about focus (as far as I can remember) which is if JS during paste event processing moves focus, the pasting must target the control that gained focus." [04:57:50.0000] hallvors: afaict that requires using some kind of hook to clearly indicate what control you're talking about [04:58:56.0000] hm.. crap, why doesn't ReSpec generate a ToC for the spec now?? [04:59:19.0000] OK, now it did [04:59:22.0000] just slow.. [04:59:51.0000] What it says about focus is this one sentence: https://w3c.github.io/clipboard-apis/#event-listeners-that-modify-selection-or-focus [05:00:16.0000] hallvors: that link doesn't work [05:00:37.0000] But I'm also not sure why it matters that it's a single sentence [05:00:52.0000] If it's normative it should be crystal clear how it fits into the bigger picture [05:01:14.0000] because Respec is weird now.. .-p so this needs to hook into some definitions of Focus and Selection to be useful? [05:01:41.0000] (just search for focus, you'll find it fast) [05:04:00.0000] hallvors: yeah, all occurrences of focus require some link [05:04:07.0000] annevk: could you add some specific suggestions for that paragraph to issue 26? [05:04:24.0000] hallvors: I don't really have good suggestions at the moment, only that we'd want to define it [05:04:51.0000] hallvors: and once we do "If the event listener modifies the selection or focus, the clipboard action must be completed on the modified selection." can probably be removed (or turned into a note) since it would follow from the requirements [05:05:22.0000] hallvors: whenever you have something out-of-band like that requiring something, it's a good idea to figure out how you can turn it into a note, since it's likely some integration is missing [05:09:59.0000] So..you're saying there is no authoritative definition of Focus to link to at the moment? [05:11:30.0000] But once there is one, I should hook things up [05:12:03.0000] hallvors: HTML is the authoritative definition [05:12:24.0000] hallvors: I'm just saying that I don't really know the specifics of focus since I haven't studied that in detail [05:12:42.0000] hallvors: that's why I suggested you file some issues (also for DataTransfer) [05:23:26.0000] https://w3c.github.io/html/editing.html#focused and https://w3c.github.io/html/editing.html#focusable might define the concepts we should reference. [05:25:02.0000] hallvors: you should definitely not reference the fork... [05:25:50.0000] s/w3c.github/specs.whatwg or something [05:25:57.0000] hallvors: but e.g., https://html.spec.whatwg.org/multipage/interaction.html#focused seems reasonable [05:26:01.0000] /me doesn't like politics [05:26:34.0000] /me actually likes politics when it is about real issues [05:26:35.0000] hallvors: just think of it as the wrong link then [05:26:51.0000] hallvors: since it'll lead to the wrong thing being implemented [05:27:04.0000] it's not really a political thing at that point [06:01:34.0000] Is there a bug in github's UI? Twice now while navigating the UI, the green button says "Confirm squash and merge" but I have NOT clicked "Merge pull request" [06:02:52.0000] e.g. reloading the page is enough [06:06:03.0000] zcorpan: dunno, just got the same [06:34:54.0000] Domenic: where should https://www.w3.org/Bugs/Public/show_bug.cgi?id=26182 go? [06:35:38.0000] annevk: https://github.com/whatwg/console/issues [06:38:52.0000] Domenic: thanks, moved there [06:43:21.0000] The commit message also has doubled linebreaks in the textarea... Such confidence [09:55:04.0000] https://developer.microsoft.com/en-us/microsoft-edge/platform/issues/ [09:55:08.0000] new it seems [10:50:22.0000] GPHemsley: in https://mimesniff.spec.whatwg.org/#rules-for-identifying-an-unknown-mime-type 7.1.1 ends with "What about feeds?" question. I wonder, was there any discussion on this topic? [10:50:49.0000] also, what about SVG? AFAIU, according to this section it should end up as text/xml [11:55:37.0000] This is amazing https://developer.microsoft.com/en-us/microsoft-edge/platform/catalog/ [11:55:52.0000] Scrapes current specs and compares against all browser IDLs [11:56:13.0000] Using open source data for everyone but Edge, and internal data for Edge. [12:23:06.0000] nice! [12:23:49.0000] DOM2 Style but not CSSOM [12:29:23.0000] the visualization looks like a Venn Diagram but it looks like it only changes the size of the balloons, and the intersection is the same? [12:49:09.0000] annevk: HTMLHyperlinkElementUtils doesn't provide a searchParams getter? is that intentional? [13:30:37.0000] annevk: custom elements DOM+ parts is ready for another look after your last round of excellent feedback. It's ready for upstreaming IMO, but you might find new bugs. 2016-04-05 [18:05:02.0000] @Domenic: am i right in thinking that encodingOverride doesn't actually do anything yet in https://github.com/jsdom/whatwg-url ? [18:05:38.0000] in general, how finished is https://url.spec.whatwg.org/ and are there any complete reference implementations? [18:08:03.0000] @Sebmaster: ^ [18:12:30.0000] jhiesey: https://url.spec.whatwg.org/ intends to be finished and has browser implementations that conform [18:12:41.0000] annevk would be the person to ask for more details [18:13:32.0000] and fwiw there is java implementation at https://github.com/smola/galimatias that was written from scratch to match the requirements in teh spec [18:14:05.0000] jhiesey: yeah, encodingOverride isn't actually implemented [18:14:51.0000] I'm not sure it actually matters because it produce queryParams yet, which is the only point where it's used if i remember correctly [18:17:20.0000] s/because it produce/because it doesn't produce/ [18:33:52.0000] Is anyone here familiar with multipart streams? Forwarding on an inquiry from another channel "can one file be split between two parts of the same multipart stream" [20:44:56.0000] Sebmaster: you need it for correctly parsing the query state [20:45:27.0000] Sebmaster: URLSearchParams doesn't really use it [00:34:15.0000] annevk: If I add crytographic nonce metadata to requests, would you like me to expose it via Request/RequestInit? (re: https://github.com/whatwg/fetch/issues/269) [00:34:53.0000] mkwst: is there a use case? [00:35:10.0000] Beyond "explain the platform", I don't have one. [00:35:20.0000] mkwst: probably not for v0 then [00:35:34.0000] mkwst: it's rather tied to the CSP header, no? [00:36:01.0000] annevk: Yes. I don't know of anything else using the `nonce` attribute for anything. [00:36:01.0000] mkwst: that is, at the moment you also can't fake type or destination [00:36:19.0000] mkwst: so if nonce only works when type is script, it doesn't add much value [00:36:34.0000] mkwst: you might want to add nonce support to workers though... [00:36:38.0000] annevk: style too. [00:37:29.0000] https://github.com/w3c/webappsec-csp/issues/15 <-- worker nonces. [00:38:21.0000] Seems Nadoedalo is ahead of the curve [00:38:38.0000] it's not tough to be ahead of me... [00:39:39.0000] Do you happen to know how the HTML bits of xref work? [00:39:49.0000] Do I just run `html.py` to regenerate the json files? [00:40:21.0000] mkwst: you need to update html.json unfortunately [00:40:26.0000] mkwst: and then run html.py [00:40:54.0000] um. ok? what does html.py do, then? [00:41:39.0000] mkwst: cleanup and multipage references [00:42:16.0000] mkwst: so you don't need to stick to alphabetical order when inserting stuff in html.json [00:42:37.0000] Ok. Well, it's throwing exceptions. :) [00:42:46.0000] Is that still being used? [00:43:15.0000] Will you be terribly sad if I just add references without running html.py? :) [00:45:32.0000] Ms2ger: yeah [00:45:43.0000] mkwst: I guess I can try to fix it [00:46:57.0000] You could try to argue this is all my code, I suppose [00:49:47.0000] I'm mostly stuck with it since bikeshed has these odd things [00:50:05.0000] I guess I should really try to get bikeshed not to do the things I dislike, that's probably more productive [00:57:03.0000] annevk: HTMLHyperlinkElementUtils doesn't provide a searchParams getter? is that intentional? [00:57:24.0000] zcorpan: yeah [00:57:43.0000] annevk: ok. what was the rationale? [00:57:53.0000] zcorpan: since it can't work for Location and WorkerLocation it seemed better to just have it for URL [00:58:46.0000] zcorpan: minor post-decision-reason could be that and use an override encoding and URL doesn't [00:59:00.0000] zcorpan: and that URLSearchParams always uses utf-8 [00:59:27.0000] ok thanks [01:06:47.0000] mkwst: so for https://github.com/whatwg/fetch/pull/273 you didn't update xref at all? [01:06:52.0000] mkwst: or it wasn't needed? [01:07:19.0000] I updated it locally, and apparently didn't send a PR. 2 seconds. [01:08:23.0000] https://github.com/whatwg/xref/pull/11 [01:20:12.0000] mkwst: wasn't sure whether https://github.com/whatwg/fetch/issues/269 should be closed, will leave that to you [01:20:39.0000] Probably not. It needs an HTML patch I'm writing now. [01:20:48.0000] And a CSP patch, but that's less relevant to Fetch. [01:23:07.0000] GitHub is super slow? [01:31:05.0000] annevk: yes (see: https://status.github.com/messages) [01:36:09.0000] ta [01:36:19.0000] TabAtkins: so I got conflicts now for "origin" [01:36:28.0000] TabAtkins: then I use "spec" as suggested to disambiguate [01:36:46.0000] TabAtkins: but by default "spec" just picks the first in HTML, even it says noexport! [01:37:00.0000] TabAtkins: that's somewhat absurd, no? [01:51:16.0000] zcorpan: Where's the best place to submit PRs for CSSOM? :) [01:51:22.0000] Just file something at https://www.w3.org/Bugs/Public/buglist.cgi?component=CSSOM&list_id=62622&product=CSS&resolution=--- ? [02:00:50.0000] mkwst: https://github.com/w3c/csswg-drafts/ - but create a fork, branches on this repo don't work because of the hg sync [02:01:10.0000] zcorpan_: Thanks! [02:02:01.0000] /me waits an hour for GitHub to catch up. :/ [02:08:56.0000] /me lights his peace pipe and enjoys an early-evening break [02:14:38.0000] zcorpan_: If you have a moment, can you help me understand how "fetch a CSS style sheet" relates to "create a CSS style sheet"? [02:14:52.0000] The latter has data I want to use in the former (namely the "owner node" of the sheet). [02:15:35.0000] "create" is called from HTML. [02:15:42.0000] "fetch" doesn't seem to be called at all. [02:16:51.0000] mkwst: https://github.com/whatwg/html/issues/968 [02:17:26.0000] mkwst: also i think several things in this area are currently broken, i haven't spent time yet to fix it [02:19:04.0000] Got it. Ok. Then I'll defer this PR a bit and file a bug for myself to come back to it. :) [02:19:12.0000] mkwst: happy to allocate time for this soon though and help with whatever you want to fix [02:19:54.0000] I know jochen__ wants clarity here for Referrer Policy, and it would be helpful for CSP and SRI, which both need to stuff things into Fetch for various checks. [02:20:30.0000] *shrug* Not the most important thing ever, but it would be nice to have a clear path from `` in HTML through to Fetch for stylesheets. [02:20:31.0000] Yeah, he's always reminding me how CSS doesn't have its act together [02:20:50.0000] But note that fixing that would also requiring some deeper fixes, such as defining what `background-image` and such do [02:20:59.0000] Yup. Baby steps. [02:21:54.0000] Yeah, it's only a two decade old feature [02:22:05.0000] Would be a bit soon for it to be properly defined [02:22:29.0000] Fetch isn't two decades old. :) [02:22:37.0000] mkwst: background-image is [02:22:57.0000] I see. [02:23:38.0000] 1) Here's a URL. 2) ??? 3) It works!! [02:23:43.0000] Also, I think I could argue that parts of Fetch maybe are that old [02:23:57.0000] zcorpan_, hey, you know if attr() case-sensitivity is specced? [02:24:09.0000] Ms2ger: what aspect of it? [02:24:16.0000] E.g., the redirect behavior has existed since, well, browsers [02:25:01.0000] zcorpan_, something that says "it's case-insensitive in HTML documents, case-sensitive elsewhere" [02:29:22.0000] zcorpan_, will file on HTML, I guess [02:29:24.0000] Ms2ger: i don't find anything about that [02:32:48.0000] html http://software.hixie.ch/utilities/js/live-dom-viewer/saved/4034 xhtml http://software.hixie.ch/utilities/js/live-dom-viewer/saved/4035 [02:33:37.0000] Also http://test.csswg.org/suites/css21_dev/nightly-unstable/html4/content-attr-case-001.htm :) [02:35:47.0000] "The case-sensitivity of attribute names depends on the document language." says CSS 2.1 [02:36:09.0000] https://github.com/whatwg/html/issues/991 [02:36:12.0000] but no such text in https://drafts.csswg.org/css-values-3/#funcdef-attr [02:40:21.0000] I'd file an issue there, but they use Tracker [02:40:38.0000] And that's really not worth my time [02:43:18.0000] TabAtkins, ^ [02:55:49.0000] squash and merge is very satisfying [03:12:45.0000] mkwst: should the HTML parser hide nonce attributes? [03:15:04.0000] annevk: Why? [03:15:34.0000] mkwst: probably not worth the complexity, but it reduces the risk of nonce being reused by an attacker [03:16:06.0000] An attacker could only gain access to the nonce if they had the ability to run script on the page though, right? [03:16:16.0000] At that point, it's not clear what we'd be defending against. [03:16:37.0000] mkwst: sure, it's defense-in-depth, similar to your credential scheme [03:17:07.0000] mkwst: e.g., it might prevent exfiltration [03:19:54.0000] annevk: File a bug? It's certainly worth considering, though I'm not sure the complexity is worth it. [03:20:06.0000] annevk: I've been thinking about something else that might be worth it, though: [03:20:28.0000] An issue that keeps coming up is injection of partial tags to capture attributes on subsequent tagfs. [03:21:09.0000] Like, an attacker injects ` can do that too [05:35:39.0000] so not document.write after load event [05:35:46.0000] but perhaps the current setup is fine [05:41:11.0000] cancelBubble has been in UIEvent in Gecko at least 15 years [05:43:58.0000] hmm, it used to have some different behavior [05:44:20.0000] it has always set the stopPropagation flag [05:44:52.0000] but if I read this code from 1999 right, one was able to set it only during bubble phase or something [05:45:14.0000] /me tries to recall the meaning of various flags in old gecko [05:45:54.0000] ok, cancelBubble was added in a commit "Adding old 4.x capture/release compatibility code" [05:46:04.0000] (unfortunately no bug number) [09:35:29.0000] smaug____: I sure hope toggling stop propagation is a sufficient implementation [09:49:32.0000] yeah [09:49:35.0000] I wonder what Edge does [09:50:04.0000] annevk: one option is to try to remove the attribute [09:50:14.0000] I have no idea how often it is being used [09:51:12.0000] smaug____: hopefully philipj can figure that out, but with everyone having something… [09:51:34.0000] oh, "EventCancelBubble 3.2310%" [09:51:42.0000] if I interpret chrome status right [09:53:21.0000] Geez [10:04:00.0000] oh, right. IE didn't use to have .stopPropagation() but only cancelBubble [10:04:03.0000] no wonder 2016-04-10 [04:14:51.0000] what's the use case for https://encoding.spec.whatwg.org/#utf-8-decode ? (UTF-8 decode with BOM removal but not honoring UTF-16 BOMs) [04:15:02.0000] i.e. which specs use that hook? [04:36:32.0000] hsivonen: workers, JS modules, EventSource, XHR JSON, WebVTT, etc [06:37:00.0000] annevk: thanks [07:30:22.0000] annulen et al.: There was little to no discussion on any of my XXX comments in mimesniff. [07:33:03.0000] GPHemsley: annulen? [07:33:35.0000] Hey, I'm just responding to whoever asked me a question. [07:34:38.0000] GPHemsley: in https://mimesniff.spec.whatwg.org/#rules-for-identifying-an-unknown-mime-type 7.1.1 ends with "What about feeds?" question. I wonder, was there any discussion on this topic? [07:34:39.0000] also, what about SVG? AFAIU, according to this section it should end up as text/xml [07:40:48.0000] Ah [14:44:00.0000] Are there no web platform tests for the encoding standard as it applies to parsing? Only for TextEncoder/TextDecoder? /cc jgraham Ms2ger [14:53:19.0000] Domenic: not really [14:57:44.0000] Well, I'm going to try to make a jsdom volunteer write some, then... [15:02:53.0000] And SVG tests are in the CSSWG test suite, I assume :-/ [15:03:15.0000] or... don't exist? [15:04:00.0000] SVG tests are in their own repo [15:04:18.0000] I'm not entirely sure about the status wrt SVG2 [15:07:46.0000] Domenic: or, uh, maybe they aren't in any repo? [15:07:55.0000] I found https://www.w3.org/Graphics/SVG/WG/wiki/Test_Suite_Overview ... [15:08:01.0000] yeah [15:08:09.0000] maybe there's nothing being updated for SVG2 yet :\ [15:08:20.0000] /me wonders who to ping who'll know [15:08:21.0000] looks like nothing on github at least [15:08:31.0000] heycam|away or TabAtkins have been my recent go-tos for SVG questions [15:08:46.0000] I'll just tell this contributor to put them in web-platform-tests/to-upstream/svg/ [15:08:50.0000] and we'll figure it out later [15:21:06.0000] https://www.w3.org/Graphics/SVG/WG/wiki/SVG2/Authoring_Guide has nothing ontestsuites [15:21:43.0000] https://www.w3.org/Graphics/SVG/WG/wiki/Main_Page#Testing says they should go into wpt [15:26:17.0000] Domenic: so yeah, the svg directory in wpt is the current SVG testsuite 2016-04-11 [09:55:09.0000] TabAtkins: Could you add the now-bikeshedded https://w3c.github.io/permissions/ to Shepherd? [09:56:17.0000] jyasskin: that is great news [09:56:38.0000] TabAtkins: also add https://storage.spec.whatwg.org/ if it isn't already [09:57:01.0000] jyasskin: I should have some later this week to get persistent-storage in [09:58:17.0000] annevk: Cool. I'm going to go through the comments folks made on #66 and get them fixed up, in spare time between Bluetooth work. I should be able to accept PRs fairly quickly. [10:10:58.0000] So, what's the input to the navigation algorithm these days? A URL object or a URL string? [10:11:23.0000] The text looks like it expects an object, but https://html.spec.whatwg.org/#following-hyperlinks:navigate seems to call it with a string [10:27:59.0000] jgraham: there's no consistency on that yet, but it should be an object in the end [10:28:14.0000] jgraham: I'm hoping to get around to refactoring the navigate algorithm in due course [10:28:38.0000] jgraham: probably starting with making it clearer what the actual arguments to the algorithm are [10:36:46.0000] annevk: So one should expect the final outcome to be that you have to pass in a url object? [10:37:16.0000] jgraham: I think so, otherwise navigating to blob URLs would not work correctly [10:38:55.0000] OK [10:39:04.0000] Thanks [10:56:20.0000] smaug____: any chance you could update https://bugzilla.mozilla.org/show_bug.cgi?id=1211894? [11:11:44.0000] /me tries to page in that change soon [11:20:31.0000] ta [12:16:24.0000] annevk: so do you know if Edge has the same behavior as Blink? [12:37:56.0000] smaug____: I don't know apart from what Travis wrote in the corresponding issue [12:38:17.0000] smaug____: I still need to get Edge on a VM [12:44:44.0000] Someone provide me something to test, I can run it in Edge. [12:45:07.0000] I feel like Microsoft should send free windows laptops to all spec editors or something [12:46:13.0000] Hmm, free hardware from all the vendors for spec editors! [12:47:58.0000] yeah, i guess apple should provide us with macs too [12:48:11.0000] or we could be employed by companies that let us expense such things, that works too [12:49:32.0000] /me has a macbook just to boot it twice a year to test something on OSX [14:31:01.0000] of the two links for logs for this channel, only the second one works, the first link shows empty logs for today and the past several days: https://wiki.whatwg.org/wiki/IRC [15:35:24.0000] Domenic: guess they didn't like ur edits https://github.com/w3c/webcomponents/commit/03b9f3ec9e456c86d088b8f211c84d15818bbbdc [15:54:21.0000] miketaylr: they took all the text [15:54:56.0000] Domenic: oh, so they did like your edits [15:55:10.0000] unclear it's legit [15:55:18.0000]

CEReactions]

[10:21:48.0000] 
annevk: Gecko's relatedTarget retargeting may not be the good one. it happens during event handling

[10:22:43.0000] 
smaug____: I'm just talking about target retargeting for now

[10:22:59.0000] 
Domenic: It does not, no, because Bikeshed reparses the IDL and throws out any markup in it. ^_^

[10:23:05.0000] 
annevk: but let me read that bug again once I've dealt with these housing cooperative papers

[10:23:07.0000] 
uh oh

[10:23:09.0000] 
(It'll work for HTML where y'all still mark up all your IDL manually.)

[10:23:11.0000] 
in 10 mins or so

[10:23:26.0000] 
smaug____: enjoy and thanks

[10:23:36.0000] 
So I guess I'll file a bug on Bikeshed to figure this out, and in the meantime DOM will just not have clickable annotations

[10:24:02.0000] 
It might be fine to clash with Constructor linking - the definition of a Constructor is just the constructor method, which is super-obvious.

[10:24:09.0000] 
All the rest can link to their concept definition.

[11:01:15.0000] 
JakeA: ping?

[12:22:01.0000] 
annevk: is ShadowRoot's innerHTML defined anywhere?

[12:23:08.0000] 
Domenic: that's why I'm considering taking over parts of P&S…

[12:23:16.0000] 
annevk: ah OK...

[12:26:53.0000] 
So many ReSpec specs.....

[12:43:22.0000] 
font-family: monospace, Droid Sans Fallback, Helvetica Neue, sans-serif; https://resources.whatwg.org/standard.css

[12:43:28.0000] 
that... makes little sense

[12:46:30.0000] 
contradiction in the spec: " In all cases, the input event comes before the corresponding change event (if any)." ... "the user agent must queue a task to fire a simple event that bubbles named input at the input element. The corresponding change event, if any, will be fired when the control loses focus."

[12:47:53.0000] 
/me files a bug

[13:11:51.0000] 
mounir: so what would a web page do with memorypressure?

[13:12:23.0000] 
I mean, it could sure try to not keep certain stuff alive, but it would be still up to the GC/CC/whatever to actually release the stuff

[13:12:34.0000] 
release blobs or canvases maybe

[13:13:09.0000] 
mounir: oh, I see, UA is expected to GC afterwards

[13:15:25.0000] 
maybe this could work

[13:18:21.0000] 
mounir: though, do we really need any levels?

[13:36:25.0000] 
I’m rewriting flex in JS, and I’m trying to group (row with row-reverse) and (column with column-reverse). I’m thinking the former means the container has a mainAxis of inline and later means the container has a mainAxis of block. Or should I just call them horizontal and vertical?

[14:14:21.0000] 
Correct - row means inline, column means block

[14:14:38.0000] 
JonathanNeal: ^^^

[14:15:07.0000] 
Awesome. I thought I had caught that. And the label for that is `axis`?

[14:15:30.0000] 
What do you mean by "label"?

[14:17:47.0000] 
/me hopes to explain shortly.

[14:33:47.0000] 
TabAtkins: something like this https://gist.github.com/jonathantneal/500a650ce01ddea12d555e127cbb49ba#flex-container- I’m trying to define what happens as I calculate flexbox.

[14:39:48.0000] 
I’m describing an `axis` property on the flex container. It is either `inline` or `block` based on `flex-direction`.

[15:38:01.0000] 
JonathanNeal: Then yes.


2016-04-21
[23:05:47.0000] 
Anyone know a simple way to draw tree structures with SVG?

[23:15:34.0000] 
annevk: google says http://fperucic.github.io/treant-js/

[23:17:14.0000] 
or http://jimblackler.net/treefun/index.html

[23:20:31.0000] 
I like that one from Jim, I wonder if the copyleft license poses an issue though

[23:21:08.0000] 
well if you are just using it to create diagrams the license doesn’t matter, right?

[23:21:23.0000] 
or you are wanting to embed something?

[23:21:48.0000] 
Yeah maybe that's right

[23:22:19.0000] 
No need for embedding, though I suppose there is a need for that project staying alive so we can create more diagrams as needed...

[23:24:26.0000] 
sure

[23:26:07.0000] 
What I want is to create similar graphics to those in Shadow DOM but with SVG I can edit and make accessible

[23:26:33.0000] 
And perhaps some extra graphics too to explain event dispatch and how shadow trees mess with that, etc.

[23:28:05.0000] 
I wonder if TabAtkins has diagrams in CSS specs he used something to create

[23:28:22.0000] 
I mean other than his railroad diagrams

[23:29:02.0000] 
in general he seems to have a knack for diagrams

[23:29:41.0000] 
anyway, that interactive treefun tool seems pretty serviceable

[01:29:13.0000] 
annevk, r? https://github.com/w3c/web-platform-tests/pull/2875

[01:33:49.0000] 
Ms2ger: merged

[03:14:12.0000] 
hello, why does noscript element contain markup as raw text node when scripting is enabled? What's the purpose of this?

[03:15:18.0000] 
Why can't it just parse contents, create DOM nodes, but browser would hide this subtree

[03:16:08.0000] 
Neu: I suspect it's a browser-introduced hack from back in the day to prevent  and such to load

[03:16:31.0000] 
hmm. yea, this makes sense

[03:23:48.0000] 
I'm reading https://www.w3.org/TR/2012/WD-html5-20121025/the-noscript-element.html : "The noscript element must contain only link, style, and meta elements." Does this mean I must drop all other tags? E.g. firefox for this case (simplified)  migrates  into body.

[03:25:02.0000] 
Neu: I recommend reading https://html.spec.whatwg.org/multipage/scripting.html#the-noscript-element

[03:25:14.0000] 
Neu: what is allowed as contents depends on the context

[03:25:29.0000] 
Neu: if you place