| 02:46 | <MikeSmith> | expected better than that from Bruno Haible |
| 02:48 | <MikeSmith> | side note: I’m trying to imagine how much worse that bug-tracking tool could be in terms of usability |
| 02:50 | <MikeSmith> | I gues that’s one more way the GNU Project is contributing to the greater good: Their Web-based systems are so consistently horrible and user-hostile that they make everybody else’s bad systems seem tolerable |
| 03:39 | <MikeSmith> | just when you thought there’s no way the “The is="" attribute is confusing? Maybe we should encourage only ES6 class-based extension” issue discussion could get any worse, csuwildcat shows up https://github.com/w3c/webcomponents/issues/509#issuecomment-280864156 |
| 03:40 | <MikeSmith> | somebody should lock that conversation |
| 05:57 | <annevk> | gsnedders: UTR #22 is best ignored though |
| 06:21 | <MikeSmith> | annevk: what should be referenced instead of UTR #22? |
| 06:24 | <MikeSmith> | https://encoding.spec.whatwg.org/#concept-encoding-get I guess |
| 06:26 | <annevk> | MikeSmith: yup |
| 06:27 | <annevk> | MikeSmith: treating shift-jis as shift_jis breaks sites for instance |
| 06:27 | <MikeSmith> | oh |
| 06:49 | <MikeSmith> | gsnedders: point out to Bruno that groff also uses “utf8” https://www.haible.de/bruno/packages-groff-utf8.html |
| 08:03 | <annevk> | The reason "utf8" works in Encoding is because we made it a label for "utf-8" |
| 08:03 | <annevk> | Not all browsers accepted that though until recently, Edge didn't for instance |
| 10:04 | <annevk> | mkwst_: could you comment on https://github.com/whatwg/html/pull/2365? |
| 10:27 | <mkwst_> | annevk: Commented, thanks for the ping. |
| 10:32 | <annevk> | mkwst_: thanks, I think per that answer I'm going to close your URL PR for this and also resolve the URL legacy bug stating we need to do something here |
| 10:32 | <annevk> | mkwst_: providing forward pointers to this PR |
| 10:32 | <mkwst_> | That sounds fine. I think we might want to move it in the future, but I'm happy to wait until we have a reason to do so. |
| 10:53 | <Ms2ger> | Funny how we had whining about for-const-of for a week, and it was removed within two days |
| 10:54 | <gsnedders> | annevk: but then you get into the whole discussion as to whether that is a standard, and UTRs seemed less of a risk for that |
| 10:54 | <annevk> | gsnedders: if you want to hand out wrong advice, be my guest |
| 10:55 | <annevk> | Ms2ger: not sure I'd qualify it as whining and nothing really got resolved so you'll continue to encounter this |
| 10:55 | <gsnedders> | annevk: also there's the question as to whether Encoding will ever get adoption outside of the web space |
| 10:57 | <Ms2ger> | Okay, maybe not whining... How do you feel about "being an antisocial jackass"? |
| 10:57 | <annevk> | Ms2ger: I think you're pretty close to violating the CoC |
| 10:58 | <annevk> | gsnedders: I think Go uses it, but maybe that still counts as web |
| 10:59 | <gsnedders> | annevk: success in that sense would be ICU, Win32, macOS's APIs adopting it, IMO |
| 10:59 | <gsnedders> | annevk: and I don't think there's been any indication that there's even a vague intention of? |
| 11:00 | <annevk> | gsnedders: Chrome might have discussed upstreaming to ICU, but I doubt ICU would start removing stuff, indeed |
| 11:00 | <annevk> | gsnedders: those platforms have obvious legacy too, the question is what new platforms will do |
| 11:07 | <gsnedders> | annevk: the problem, when almost nothing outside of browsers has any planned intention to support it, is that it's easy to dismiss as a web-specific thing |
| 11:07 | <gsnedders> | annevk: hell, non-web stuff may well rely on more encodings and different alias matching |
| 11:08 | <annevk> | gsnedders: and different decoding and encoding rules? Could be, but the web is bigger than all those places and more easily leaks into them than the other way around |
| 11:09 | <annevk> | gsnedders: actually, as for new platforms, they should probably stick to UTF-8 |
| 11:10 | <annevk> | gsnedders: any more thoughts on https://github.com/whatwg/url/issues/233 btw while you're here? |
| 12:25 | <annevk> | jgraham: created a fairly minimal PR for raw_headers |
| 12:49 | <jgraham> | annevk: Thanks! |
| 13:02 | <annevk> | jgraham: I want to checkout an old branch on WPT and I get "The following untracked working tree files would be overwritten by checkout" |
| 13:02 | <annevk> | jgraham: followed by a bunch of tools/ stuff |
| 13:04 | <annevk> | Ah, git checkout -f |
| 13:20 | <annevk> | jgraham: when will the new wptserve be used by WPT? |
| 13:20 | <jgraham> | annevk: When I update the submodule pointer |
| 13:20 | <jgraham> | Just looking at another review first |
| 13:33 | <jgraham> | annevk: Done |
| 15:04 | <annevk> | I wish I had not neglected WPT for so long and hope somebody else would take care of it 😟 |
| 15:04 | <annevk> | That strategy never works |
| 15:04 | <jgraham> | ? |
| 15:04 | <annevk> | jgraham: going through various labels of standards I edit and feeling bad |
| 15:04 | <annevk> | jgraham: currently cleaning up the XMLHttpRequest backlog |
| 15:04 | <jgraham> | Ah, yeah |
| 15:04 | <jgraham> | Getting people to review tests is hard |
| 15:09 | <JonathanNeal> | What was the motivation behind sectioning roots? And how were the elements defined as sectioning roots determined? https://developers.whatwg.org/sections.html#sectioning-root |
| 15:36 | <wanderview> | JakeA: annevk: so, uh... are we moving ahead on the fetch cancelation thing? what is the next step? |
| 15:36 | <JakeA> | wanderview: I guess we need to design the APIs |
| 15:36 | <annevk> | wanderview: I'm kinda hoping JakeA will take the lead |
| 15:36 | <JakeA> | It's on my TODO list to circle back on that |
| 15:37 | <JakeA> | I guess I should give it some priority |
| 15:37 | <wanderview> | I feel like its something we should ship in 2017 if we can |
| 15:37 | <annevk> | I've mostly been fixing bugs URL/Fetch/XMLHttpRequest and their tests, and also some HTML work |
| 15:37 | <JakeA> | Let me spend tomorrow on it |
| 15:37 | wanderview | notes JakeA will solve fetch cancellation tomorrow. |
| 15:37 | wanderview | goes to tweet... |
| 15:37 | <JakeA> | haha |
| 15:37 | <wanderview> | j/k |
| 15:37 | <annevk> | Let's see, two weeks from now I should be able to help out more |
| 15:38 | <annevk> | I'd like to get WPT more into shape this week and next week I've some meetings |
| 15:39 | <wanderview> | I need to write a WPT test, but can't decide where to put it |
| 15:39 | <wanderview> | need to test referrer header on fetch/xhr from worker script when the worker script was redirected |
| 15:39 | <wanderview> | does that go in fetch, workers, or some security dir? |
| 15:42 | <annevk> | wanderview: fetch/ seems fine |
| 15:42 | <annevk> | wanderview: fetch/ has lots of worker tests already |
| 15:42 | <wanderview> | well, referrer-policy/generic tests are disabled currently |
| 15:42 | <annevk> | wanderview: and some referrer tests as it has referrer features |
| 15:43 | <annevk> | wanderview: in fetch/ or in Gecko? |
| 15:43 | <wanderview> | claims because of https://github.com/w3c/web-platform-tests/issues/1874 |
| 15:43 | <wanderview> | which was fixed 1.5 years ago |
| 15:43 | <wanderview> | maybe we can turn them back on |
| 15:43 | wanderview | sighs |
| 15:44 | <wanderview> | I guess thats just a subset of tests |
| 16:28 | <annevk> | Domenic: how does IDB work? |
| 16:51 | <zcorpan> | annevk: LOL I tried to review the PR review PR for quirks, and only noticed after merging that the source file is not named quirks.bs |
| 16:51 | <zcorpan> | annevk: but maybe we can rename it to that |
| 16:54 | <zcorpan> | annevk: does that need a change to .htaccess or something? |
| 16:55 | <annevk> | zcorpan: we can change the name in pr-preview too |
| 16:56 | <annevk> | zcorpan: not sure why .htaccess needs changes |
| 16:57 | <annevk> | zcorpan: tobie can maybe help, bit busy atm |
| 16:57 | <zcorpan> | annevk: I thought there was a DirectoryIndex Overview.html |
| 16:58 | <zcorpan> | I can look at this again tomorrow. |
| 16:59 | <tobie> | zcorpan: that's fine. Just edit the config file. |
| 17:00 | <zcorpan> | tobie: thx |
| 17:02 | <annevk> | zcorpan: I cleaned that up already |
| 17:02 | <annevk> | zcorpan: Travis generates index.html resources |
| 17:05 | <gsnedders> | hsivonen: if you're about at all, I'd be curious to hear your opinion on what level HTML sanitizers should work; rubys' (IIRC) Python one in html5lib originally worked by intercepting tokens between the tokenizer and the tree constructor, but that seems fragile. Seems better to post-process the tree, IMO? |
| 17:07 | <wanderview> | oh, I guess we fixed the spelling of referer with referrer-policy |
| 17:11 | <wanderview> | the referrer-policy tests are kind of hard to understand... |
| 17:11 | <gsnedders> | that's confusing |
| 17:13 | <annevk> | gsnedders: post-process on tree, how does that work with scripts? |
| 17:13 | <gsnedders> | annevk: does any sanitizer execute scripts or do anything but strip them? |
| 17:14 | <wanderview> | annevk: if there is a referrer-policy on the top level worker script, but also a different referrer-policy on an importScripts() script... which policy should the worker have at the end? |
| 17:14 | <annevk> | gsnedders: dunno, but worth thinking about since people keep asking for browsers to solve |
| 17:15 | <annevk> | wanderview: worker has top-level |
| 17:15 | <gsnedders> | annevk: if you're sanitizing something, surely you have to treat it as totally untrusted which means you must not execute scripts |
| 17:15 | <wanderview> | annevk: meaning only the header on the top level worker script should apply? |
| 17:15 | <annevk> | wanderview: yeah |
| 17:16 | <wanderview> | annevk: I'll write a test for that... since we don't do that right |
| 17:16 | <annevk> | wanderview: you cannot update the global policy through importScripting a resource with diff policy |
| 17:16 | <annevk> | wanderview: wow |
| 17:16 | <annevk> | And thanks |
| 17:17 | <wanderview> | annevk: all of this is a mess |
| 17:17 | <wanderview> | in gecko |
| 17:17 | <wanderview> | annevk: I regressed referrer on redirected worker scripts because we don't have tests for that... fixing in https://bugzilla.mozilla.org/show_bug.cgi?id=1340652 |
| 17:17 | <wanderview> | and just noticing other broken referrer stuff in workers |
| 17:18 | <annevk> | gsnedders: maybe not nonce scripts, but I was more thinking they might have executed by the time you propose to remove them |
| 17:18 | <annevk> | wanderview: appreciate you going down the rabbit hole |
| 17:19 | <annevk> | wanderview: try to follow such a policy myself too |
| 17:21 | <wanderview> | we have a lot of disabled referrer-policy tests as well: https://dxr.mozilla.org/mozilla-central/search?q=disabled+path%3Aweb-platform%2Fmeta%2Freferrer-policy&redirect=false |
| 17:38 | <annevk> | wanderview: stumbled across https://github.com/w3c/web-platform-tests/issues/2618 |