| 01:21 | <MikeSmith> | SGML5 http://lists.xml.org/archives/xml-dev/201611/msg00078.html |
| 06:19 | <annevk> | MikeSmith: lol |
| 06:21 | <MikeSmith> | annevk: gotta love the ambition there |
| 08:11 | <MikeSmith> | annevk: https://github.com/w3c/manual-of-style |
| 08:11 | <MikeSmith> | if you not see it yet |
| 08:12 | <annevk> | MikeSmith: https://www.w3.org/2001/06/manual/ in a repo? |
| 08:13 | <MikeSmith> | yeah I think so |
| 08:13 | <MikeSmith> | seems plh moved it over |
| 08:14 | <MikeSmith> | there was probably some discussion about it that I either failed to read or forgot about |
| 08:14 | <annevk> | I haven't seen anything but I'm not on many mailing lists anymore |
| 08:15 | <annevk> | I created https://wiki.whatwg.org/wiki/Style the other day |
| 08:15 | <MikeSmith> | yeah I did see that |
| 08:16 | <annevk> | tobie: I'm very sad we don't seem to make much progress on the type stuff |
| 08:16 | <annevk> | tobie: I wonder if we can export the word "record" at least and settle on an ID for it |
| 08:16 | <annevk> | tobie: that would unblock a contribution someone made to Fetch |
| 08:17 | <annevk> | Although maybe I should be more optimistic, at least we're discussing the type stuff and figuring out what it means together |
| 08:20 | <tobie> | annevk: feel free to do this. I'm out of office this week, but can merge this. |
| 09:16 | <annevk> | tobie: https://github.com/heycam/webidl/pull/243 |
| 10:23 | <tobie> | annevk: did you rebuild and see no changes or not rebuild? |
| 10:23 | <tobie> | annevk: I imagine the former, but one can't be too sure. :) |
| 10:24 | <annevk> | tobie: the former |
| 10:24 | <annevk> | tobie: I ran make, then committed |
| 10:50 | <annevk> | zcorpan: MikeSmith: is http://sgmljs.net/docs/html5.html#the-datalist-element an actual problem? |
| 11:00 | <zcorpan> | annevk: http://books.xmlschemata.org/relaxng/relax-CHP-16-SECT-1.html - afaict ambiguous patterns are allowed in RELAXNG and i would assume the existing schemas in v.nu have them, and that's not a problem |
| 11:00 | <annevk> | zcorpan: cool |
| 11:00 | <zcorpan> | but i could check more carefully |
| 11:01 | <annevk> | zcorpan: I fixed the other nit, haven't found anything else in this document that is pointing out a potential mistake |
| 11:02 | <zcorpan> | i suppose the bigger issue we have with content models is that the language itself can be ambiguous. it's not always clear how the prose maps to a schema, and what edge cases are allowed |
| 11:02 | <zcorpan> | for example, is this allowed? <dl><script></script></dl> |
| 11:04 | <zcorpan> | but making it crystal clear for conformance checker implementors would probably come at the expense of less clarity for web developers |
| 11:08 | <zcorpan> | given the priority of constituencies, the spec should optimize for web developers over MikeSmith and hsivonen |
| 11:20 | <annevk> | Well, but it should be unambiguous first |
| 11:26 | <annevk> | MikeSmith: xml-dev wow btw |
| 11:26 | <annevk> | MikeSmith: there are some threads there… |
| 11:31 | <MikeSmith> | annevk: yeah I didn’t even know xml-dev was still going |
| 11:31 | <MikeSmith> | it’s like stepping back in time to visit the mid-1990s |
| 11:31 | <MikeSmith> | retro tech |
| 11:33 | <MikeSmith> | annevk: about http://sgmljs.net/docs/html5.html#the-datalist-element, what zcorpan said |
| 11:33 | <MikeSmith> | “This means that upon encountering a template element in a datalist parent element, the parser cannot decide which of the two branches declared in the choice submodels of datalists grammar rule is to be selected for subsequent parsing.” is not true |
| 11:34 | <MikeSmith> | RelaxNG was designed from the beginning to allow that |
| 11:34 | <zcorpan> | i suppose he was talking about SP |
| 11:35 | <MikeSmith> | yeah I guess so |
| 11:35 | <MikeSmith> | well in another message he said he wrote an SGML parser |
| 11:36 | <MikeSmith> | so now we have two SGML parsers! |
| 11:36 | <MikeSmith> | it only took 20 years to get a second SGML parser |
| 11:37 | <MikeSmith> | that is like only seconds in SGML-time |
| 11:37 | <zcorpan> | must be more than 20 years? |
| 11:37 | <MikeSmith> | true I guess so |
| 11:38 | <tobie> | annevk: sorry had to run out and wanted to double check we weren't breaking stuff downstream before committing. (We aren't as nothing really was exported anyway.) |
| 11:38 | <tobie> | annevk: merged now. |
| 11:39 | <annevk> | tobie: ta |
| 11:44 | <MikeSmith> | btw for those of you who like to call out to Java executables from your node applications, there is now https://www.npmjs.com/package/vnu-jar just for you |
| 11:45 | <MikeSmith> | npm install vnu-jar |
| 11:46 | <zcorpan> | MikeSmith: actually now we have at least 4 sgml parsers. arcsgml (1974), sgmls, sp and this new one |
| 11:47 | <MikeSmith> | oh |
| 11:47 | <MikeSmith> | never heard of them |
| 11:48 | <MikeSmith> | except sp |
| 11:48 | <MikeSmith> | well sgmls sounds familiar sorta |
| 11:49 | <MikeSmith> | I wonder if James CLark wrote that one too |
| 11:54 | <MikeSmith> | http://sgmljs.net/blog.html “the custom element feature puts HTML into the class of markup meta languages like SGML and XML” is an interesting way of looking at custom elements |
| 11:54 | <MikeSmith> | but true I guess |
| 12:05 | <karlcow> | aaaah… manual of style |
| 12:05 | <karlcow> | MikeSmith and annevk made me jump a couple of years |
| 12:06 | <MikeSmith> | karlcow: yeah glad plh has revived it |
| 12:06 | <annevk> | Why not visit xml-dev and jump decades? |
| 12:06 | <MikeSmith> | heh |
| 12:07 | <karlcow> | I spent a big chunk of time on this with Susan Lesch |
| 12:07 | <karlcow> | annevk: <3 |
| 12:08 | <karlcow> | MikeSmith: that's cool. |
| 12:08 | <MikeSmith> | opens it up to collaboration, PRs |
| 12:09 | <karlcow> | these old things… like JapanBloggers⊙yc :D |
| 12:09 | <karlcow> | circa 2004 |
| 12:09 | <MikeSmith> | ah yaeh |
| 12:09 | <MikeSmith> | met Andreas Bovens on that list I think |
| 12:10 | <karlcow> | :) |
| 12:10 | <karlcow> | Opera Kaput |
| 12:15 | <MikeSmith> | there may be former Opera people now working at every major browser vendor |
| 12:16 | <annevk> | I think that's true |
| 12:17 | <annevk> | Not sure who Microsoft picked up other than David Storey |
| 12:18 | <MikeSmith> | yeah not sure either |
| 12:18 | <karlcow> | http://web.archive.org/web/19970605092055/http://www.w3.org/Overview.html |
| 12:20 | <MikeSmith> | Public Release of Jigsaw 1.0 Alpha 5 -- the HTTP 1.1 Compliant Server |
| 12:21 | <MikeSmith> | reminds me w3.org should deploy HTTP/2 support |
| 12:22 | <MikeSmith> | I guess there is HTTP/2 for Apache now |
| 12:22 | <nox> | https://twitter.com/FakeUnicode/status/800808929554857984 This thread is sad. |
| 12:24 | <MikeSmith> | nox: sad for one reason that somebody is quoting https://www.w3.org/TR/html-markup/syntax.html#comments there as if it is authoritative about anything |
| 12:24 | <nox> | MikeSmith: Yes, just replied. |
| 12:24 | <MikeSmith> | I really gotta fix that, make some floating thing as we have done for other do-not-use-this docs |
| 12:24 | <annevk> | nox: I like the Christmas tree |
| 12:25 | <nox> | annevk: Yeah. :) |
| 12:25 | <MikeSmith> | https://twitter.com/jleedev/status/800813497177161729 is great |
| 12:36 | <JakeA> | annevk: I'm trying to fix up service worker's use of bikeshed but struggling with the bikeshed docs. Is there a bikeshed spec that you'd consider "really good"? |
| 12:37 | <JakeA> | Trying to figure out the right way to mark up concepts |
| 12:37 | <annevk> | JakeA: Fetch is pretty good at this point |
| 12:38 | <JakeA> | ta |
| 12:38 | <annevk> | JakeA: a concept would just be "a <dfn>service worker</dfn> is a ... A <a>service worker</a> has an associated <dfn for="service worker">X flag</dfn>" etc. |
| 13:15 | <zcorpan> | MikeSmith: apparently sgmls was James Clark's fork of arcsgml |
| 13:15 | <zcorpan> | before he wrote sp |
| 13:21 | <zcorpan> | MikeSmith: about <!----->, is there an open bug for firefox? it's fixed in the v.nu HTML parser? |
| 13:23 | zcorpan | can't find a bug |
| 13:33 | <zcorpan> | filed https://bugzilla.mozilla.org/show_bug.cgi?id=1319410 |
| 13:52 | <annevk> | zcorpan: https://stackoverflow.com/questions/40735720/does-flow-content-has-to-have-palpable-content |
| 13:57 | <JakeA> | annevk: with "A <a>service worker</a> has an associated <dfn export>id</dfn>" - this is a case where I'd give the dfn an ID right, since its content is ambiguous? |
| 13:58 | <JakeA> | or would for="service worker" be enough? |
| 13:58 | <ondras> | uhm, quick css question -- https://jsfiddle.net/57p72fh0/1/ |
| 13:58 | <annevk> | JakeA: if that is the current markup you'd add for="service worker"; however, you also want to check what the current ID is and add that, to make sure you're not changing IDs around |
| 13:58 | <annevk> | JakeA: since the for attribute does affect the generated ID |
| 13:59 | <ondras> | why are the dt/dd layouted in in a different way in these two <dl>s ? |
| 13:59 | <JakeA> | annevk: gotcha. Was trying to eliminate the explicit IDs, but I guess that's too breaking |
| 14:00 | <annevk> | JakeA: yeah, you should definitely keep IDs on <dfn>s |
| 14:00 | <annevk> | JakeA: I noticed service workers regressed several times on that front, you want to start reducing that and improve reliability of linking to stuff |
| 14:01 | <zcorpan> | annevk: answered |
| 14:01 | <annevk> | JakeA: and basically add the original ID to any <dfn>s you modify by adding for attributes and such |
| 14:01 | <JakeA> | shall do, cheers |
| 14:02 | <zcorpan> | ondras: margins don't collapse across borders |
| 14:03 | <zcorpan> | ondras: maybe padding does what you want here? |
| 14:04 | <zcorpan> | ondras: also note <dl><div><dt><dd></div></dl> is allowed now, if that helps |
| 14:04 | <JakeA> | annevk: if a dfn has an ID, is there any point in "for"? |
| 14:04 | <JakeA> | as in, is the "for" redundant in <dfn export for="service worker" id="dfn-state"> |
| 14:04 | <annevk> | JakeA: yes, for will help disambiguate with terms from the same name |
| 14:05 | <JakeA> | cheers, thanks for all the help! |
| 14:05 | <annevk> | JakeA: so e.g. I can define <dfn for=XMLHttpRequest export>state</dfn> and we both don't end up linking to each others state and others can pick which state to link to |
| 14:06 | <JakeA> | gotcha |
| 14:06 | <ondras> | zcorpan: thanks, I am pretty sure I can solve this in various ways. But it is still not clear to me why the dt/dd are actually (vertically) centered in the border-less case. |
| 14:06 | <ondras> | in particular, what moves the dt downwards |
| 14:06 | <annevk> | JakeA: the ID doesn't carry any magic, it's just the identifier within your document, but it's good to keep it stable so links don't break |
| 14:07 | <annevk> | JakeA: the cross-spec and internal linking magic is all about contents / lt attribute and the for attribute, and there's also a couple of IDL-specific things |
| 14:07 | <JakeA> | annevk: the bit of the learning curve I'm hitting is where the magic starts & stops |
| 14:07 | <annevk> | JakeA: https://tabatkins.github.io/bikeshed/ has lots of information on this |
| 14:07 | <JakeA> | Hopefully by the end of this I'll *know* it rather than just copying it and hoping it's right :D |
| 14:08 | <annevk> | JakeA: I'm happy to review your PR |
| 14:08 | <ondras> | zcorpan: i.e. with just the first list present, how is it possible that <dd>'s top margin influences its previous sibling's vertical position |
| 14:09 | <JakeA> | annevk: Cheers. Will aim for it to produce html that's identical in all the important ways |
| 14:09 | <annevk> | ondras: because it collapses with the margin of the <dl> |
| 14:10 | <ondras> | annevk: the same happens when the <dl> does not have any margin |
| 14:11 | ondras | probably missing some rudimentary knowledge about margin collapsing |
| 14:11 | <zcorpan> | ondras: "A floating box's outer top may not be higher than the top of its containing block." |
| 14:11 | <annevk> | ondras: they get joined basically, larger one wins |
| 14:11 | <zcorpan> | i think is the rule that applies here |
| 14:11 | <ondras> | annevk: joining I know. of adjacent siblings. |
| 14:11 | <zcorpan> | the margin pushes down the containing block (the <dl> block box) |
| 14:12 | <annevk> | ondras: I'm talking about margins |
| 14:12 | <ondras> | annevk: yeah, margins of adjacent siblings |
| 14:12 | <ondras> | zcorpan: aha, interesting |
| 14:12 | <annevk> | ondras: also parent |
| 14:12 | <tobie> | annevk: Bikeshed doesn't namespace exported dfns with their spec name? |
| 14:12 | <JakeA> | annevk: in which cases would I *not* want to export a dfn? |
| 14:13 | <annevk> | tobie: yes, but if a term belongs to a specific thing it should use for |
| 14:13 | <annevk> | JakeA: when it's for internal usage |
| 14:13 | <ondras> | annevk: yeah, but parent's top margin only collapses with its first child's top margin |
| 14:13 | <annevk> | JakeA: generally most stuff is probably good to export, the export story is a little messy |
| 14:13 | <wanderview> | JakeA: what do you think about adding an install event handler to async-waituntil-polyfill that tests if the native waitUntil works async so it can turn the polyfill on/off? |
| 14:14 | <wanderview> | sorry for the non-sequitor during your conversation |
| 14:14 | <tobie> | annevk: I'd like to see a clear guide as when you're supposed to do what with Bikeshed. |
| 14:14 | <tobie> | annevk: the docs tells you everything you can do. It would be good to have good practice doc somewhere |
| 14:14 | <ondras> | zcorpan: I see, thanks. I supposed this is related to floats, so I shall read that part of the spec? |
| 14:15 | <JakeA> | wanderview: where would it store the result of the test? |
| 14:15 | <wanderview> | JakeA: thats a good point :-\ |
| 14:15 | <zcorpan> | ondras: floats, block boxes and margin collapsing |
| 14:15 | <wanderview> | JakeA: I guess there is no good way to feature detect it then |
| 14:16 | <JakeA> | wanderview: not that I could figure out. And if the test is async it's probably not worth it :( |
| 14:16 | <wanderview> | JakeA: well, thats why I thought doing it during install would be nice... but I forgot about needing to persist the result |
| 14:16 | <ondras> | "When the float occurs between two collapsing margins, the float is positioned as if it had an otherwise empty anonymous block parent taking part in the flow. The position of such a parent is defined by the rules in the section on margin collapsing." |
| 14:16 | <ondras> | that would be it. |
| 14:17 | <ondras> | cool, thanks for the info |
| 14:17 | <wanderview> | JakeA: I guess our impl won't see much use until chrome ships it as well |
| 14:17 | <JakeA> | Haven't caught up with the status of that yet :/ |
| 14:20 | <wanderview> | JakeA: well, this release is delayed an extra cycle... so we won't ship until April 2017... shouldn't be a problem for us to ship first since people mostly test with chrome |
| 14:20 | <zcorpan> | ondras: yeah, i think that's what actually pushes down the block. i read that paragraph but didn't understand it so went on looking at other places trying to find this :-D |
| 14:20 | <ondras> | :) |
| 14:21 | <ondras> | zcorpan: yeah, the behavior is the same in ff/chrome, so I assume it is correct. just puzzling for me. thanks a lot. |
| 14:22 | <zcorpan> | CSS |
| 14:22 | <zcorpan> | IS |
| 14:22 | <zcorpan> | AWESOME |
| 14:22 | <ondras> | :} |
| 14:22 | <css> | Thx! |
| 14:26 | <nox> | Can't spell awesome without me. |
| 14:26 | <Ms2ger> | Fortunately you can without nox |
| 14:26 | <nox> | Ms2ger: But you can't spell obnoxious without me. |
| 14:27 | <Ms2ger> | That's undeniable |
| 14:28 | <nox> | Ah ah. |
| 14:33 | <JakeA> | annevk: should https://w3c.github.io/ServiceWorker/#dfn-script-resource-map really be an infra map? |
| 14:34 | <JakeA> | Rather than pointing to ecmascript concepts |
| 14:35 | <annevk> | JakeA: I suspect so, but maybe ask Domenic since it involves importScripts() |
| 14:35 | <JakeA> | cheers |
| 14:53 | <wanderview> | annevk: can we lock this thread or something? its devolving into conspiracy theories, etc: https://github.com/whatwg/html/issues/315#issuecomment-262210092 |
| 14:56 | <nox> | This is gold. |
| 15:08 | <zcorpan> | wanderview: commented |
| 15:08 | <wanderview> | zcorpan: thanks |
| 15:35 | <annevk> | wanderview: 😟 |
| 16:29 | <Ms2ger> | Is there no IDL algorithm for creating an ArrayBuffer? |
| 16:29 | <nox> | Ms2ger: "Make it so" boom, done. |
| 16:40 | <Ms2ger> | annevk / Domenic: around? |
| 16:41 | <Domenic> | we've merged into a superentity ;_; |
| 16:41 | <Ms2ger> | I'm looking at https://html.spec.whatwg.org/multipage/workers.html#run-a-worker , your superentitiness |
| 16:42 | <Ms2ger> | Correct me if I'm wrong... |
| 16:42 | <Ms2ger> | "outside settings" is the realm of the Worker that's being constructed |
| 16:42 | <Ms2ger> | "inside settings" is the realm of the DedicatedWorkerGlobalScope |
| 16:43 | <Ms2ger> | We do "Fetch a classic worker script given url, outside settings, and destination." |
| 16:43 | <Ms2ger> | It returns a "classic script", whose "settings object" is "outside settings" |
| 16:43 | <Domenic> | hmmmm |
| 16:43 | <Domenic> | yes, that seems a little broken |
| 16:45 | <Ms2ger> | Seems like it's using a realm on the wrong thread... That's more than a little in C++ :) |
| 16:45 | <Domenic> | I forgot that the settings object passed to fetch a classic worker script was used for more than just the fetch client |
| 16:45 | <Domenic> | Will add another parameter |
| 16:45 | <Ms2ger> | Thanks |
| 18:14 | <jsbell> | wanderview: thanks for test feedback! We're now running the w-p-t copies locally so expect more patches as I try and reconcile our local copies so I can delete them :) |
| 18:15 | <wanderview> | jsbell: np, sounds great |
| 18:28 | <bterlson> | TabAtkins: is this a bug? http://codepen.io/anon/pen/XNRdwg Position: fixed doesn't work when a child of an element with will-change: transform. |
| 18:30 | <bterlson> | TabAtkins: I should say, chrome behaves like this, but edge does not :-P |
| 19:06 | <KiChjang> | are there any plans for an implementor-only HTML spec? |
| 19:06 | <KiChjang> | i know there's the developer-only spec |
| 19:17 | <Domenic> | KiChjang: no concrete plans, but yes vague ones. We see implementers get confused by authoring requirements way too often; it would be nice to do something about that. |
| 19:45 | <gsnedders> | Domenic: I wonder if it'd be worthwhile having a bar to one side saying "this applies to x"? |
| 19:46 | <Domenic> | yeah, it's really just a matter of tooling + marking up the spec consistently |
| 22:51 | <n0ah_> | Hello, got a question regarding CORS: is there a way to detect (client side) the header fields exposed by Access-Control-Expose-Headers? |
| 22:53 | <n0ah_> | here's my use case a browser supporting fetch but missing Headers.prototype.entries Id have to pick each header exposed using Headers.prototype.get but for that Id need the list of these headers. |
| 22:54 | <n0ah_> | Is it exposed anywhere? |
| 22:55 | <n0ah_> | If it's not the right place to ask could someone redirect me to the right channel? |
| 23:05 | <n0ah_> | a/ everyone is sleeping b/ too simple c/ too complicate |
| 23:05 | <n0ah_> | Which one is it? |
| 23:12 | <n0ah_> | the fetch spec. has this channel listed in its header |
| 23:12 | <n0ah_> | you should remove it |
| 23:14 | <caitp> | wow |