| 00:44 | <falken> | wanderview: good question. I don't think there is polished documentation. https://bugs.chromium.org/p/chromium/issues/detail?id=501569 was the main bug which links to a design doc. |
| 01:57 | <wanderview> | falken: says the design doc is internal? |
| 01:58 | <wanderview> | oh, I see it now |
| 02:00 | <wanderview> | falken: does it let an extension intercept a page's network requests like a script blocker addon would want to do? |
| 02:01 | <falken> | wanderview: i'm actually still wading through the tests so I'm learning as I go, but I don't think it does. It's like if there were a page at blah://example/page with a service worker from blah://example/sw.js and the SW intercepts requests from the page. |
| 02:02 | <falken> | wanderview: except in this case "blah://" is "chrome-extension://" |
| 02:02 | <wanderview> | falken: oh... I think our addon team is very confused then... they keep asking for this as a way for an addon to provide alternate responses to web content |
| 02:03 | <wanderview> | falken: fwiw, I've been opposed to trying to do that because I don't want any kind of cross-origin capability in the service worker code... its too dangerous IMO |
| 02:04 | <falken> | wanderview: yea... i don't think they can do that. and agreed about cross-origin concern. |
| 02:04 | <wanderview> | falken: I guess I can go suggest doing it through WebRequest extension API again |
| 02:08 | <wanderview> | falken: thankjs |
| 04:52 | <domfarolino> | Domenic: Hmm what is the original IRCCloud view? Is there a new/optional UI update? |
| 08:23 | <nox> | annevk: https://github.com/servo/servo/pull/20693/files#diff-43ef31cdf6bd590d4201c07f57a9cd9b :( |
| 08:23 | <nox> | You using await/async in there. |
| 08:25 | <annevk> | nox: it already used that, I just made it apply to more globals |
| 08:26 | <annevk> | nox: but also, async/await is great; Servo should really update |
| 08:26 | <nox> | Really? It used to not fail to parse. |
| 08:26 | <annevk> | nox: https://github.com/w3c/web-platform-tests/commit/5adab93dc7ec658c1f3ced37c9934d3fe90f6251 is all I changed |
| 08:26 | <nox> | Disregard me then. |
| 08:27 | <annevk> | nox: seems you're concerned with something that happened a month ago: https://github.com/w3c/web-platform-tests/commit/b202bbb5aa0d235b22bac11fe902eab1094ef9d2 |
| 10:55 | <annevk> | mkwst: https://github.com/whatwg/fetch/issues/717 |
| 10:55 | <annevk> | JakeA: https://github.com/whatwg/fetch/pull/702 |
| 10:57 | <JakeA> | annevk: ta. Don't suppose you could nudge the script so the diff is updated with linked terms? |
| 10:58 | <annevk> | JakeA: can do |
| 10:59 | <annevk> | JakeA: should be done within a minute or so |
| 11:01 | <JakeA> | Ta! It's updated already |
| 12:51 | <annevk> | JakeA: one more look at https://github.com/whatwg/fetch/pull/702? |
| 12:52 | <annevk> | JakeA: I think I got the indenting right, but still |
| 13:20 | <JakeA> | Done |
| 14:37 | <annevk> | wanderview: last sentence of https://github.com/whatwg/fetch/issues/717#issuecomment-384653109 ; do you mean "would not"? |
| 14:37 | <wanderview> | annevk: yea, I'll fix... thanks |
| 14:39 | <annevk> | wanderview: so the main use case is to avoid having to rewrite your CSP? |
| 14:39 | <wanderview> | annevk: no... |
| 14:39 | <wanderview> | annevk: the use case is restricting what the service worker can do |
| 14:40 | <annevk> | But it doesn't really restrict it with this, right? |
| 14:40 | <wanderview> | annevk: sure it does... connect-src: * is *a lot* more permissive in that you can stick any URL in there with zero effort... requiring someone to XSS the main page through an image element is harder |
| 14:41 | <annevk> | wanderview: well, the service worker is often in charge of the main page, so... |
| 14:42 | <annevk> | (and also in charge of its CSP policy, iirc) |
| 14:42 | <wanderview> | annevk: so CSP is pretty pointless on service worker sites? |
| 14:43 | <wanderview> | if you want to do anything like the common use case you have to open it up pretty wide |
| 14:44 | <annevk> | Yeah, it is somewhat, but not entirely |
| 14:45 | <annevk> | The CSP folks didn't really care much about this when I raised it |
| 14:45 | <annevk> | Maybe once it becomes an actual problem they'll get back to me... |
| 14:45 | <annevk> | Mike West is hard to find these days |
| 14:45 | <wanderview> | annevk: do the content type sniffing things make this any better? to block sending non-image data back for an image destination request? |
| 14:47 | <annevk> | wanderview: I'm not entirely sure what you're talking about, but I suspect not |
| 14:47 | <wanderview> | annevk: I think for CORB and things we were doing more validation that responses were actually the correct type |
| 14:47 | <wanderview> | or there was a desire to |
| 14:48 | <wanderview> | ^I think^I thought |
| 15:02 | <annevk> | wanderview: CORB could not be read anyway, unless you count Spectre, then maybe |
| 15:05 | <annevk> | Kinda dislike the name therefore |
| 15:19 | <wanderview> | annevk: it always makes me think of CORBA |
| 17:02 | <annevk> | "The Common Object Request Broker Architecture (CORBA) is a standard defined by the Object Management Group (OMG)..." |
| 17:02 | <annevk> | wow |
| 17:11 | <wanderview> | annevk: yes, its quite exciting |
| 17:11 | <wanderview> | annevk: I used it a lot in my past life in the early 2000's |
| 18:04 | <wanderview> | falken: then chrome-extension:// service worker thing sounds a bit like foreign fetch... is that correct? |
| 18:04 | <wanderview> | it seems the addon team here also wants this for moz-extension:// |
| 18:43 | <hober> | TabAtkins: is there a tool for bikeshed where I can feed it a spec#anchor url and have it spit out how you'd reference that in bikeshed source? |
| 18:43 | <hober> | TabAtkins: ideally with what i need to put in link-defaults, if anything |
| 18:50 | <TabAtkins> | hober: Not yet, but I could probably produce something... |
| 18:52 | <TabAtkins> | If I actually tell you what's needed for a link, you wouldn't ever need link-defaults; that just lets you write shorter things. |
| 19:05 | <wanderview> | the thing I always want when linking to a spec is being able to reference a specific step |
| 19:53 | <md1> | yo |
| 19:53 | <md1> | is this where websockets was born |
| 19:58 | <TabAtkins> | wanderview: Yeah, making that work automatically is on my roadmap. |
| 19:59 | <wanderview> | TabAtkins: that would be awesome |
| 20:02 | <hober> | basically i want to paste/input something like html#attr-meta-http-equiv-refresh (or the full url) and have it spit out [=something=] or <{something}> or whatever |
| 20:07 | <md1> | hi |
| 20:17 | <md1> | yo tantek |
| 20:17 | <md1> | wat u doin |
| 20:19 | <tantek> | ? |
| 20:23 | <TabAtkins> | Sigh, randos. |
| 20:30 | <hober> | TabAtkins: do section links work for rfcs? |
| 20:30 | <hober> | cross-spec ones |
| 20:31 | <TabAtkins> | Section links "work" for everything, but if they're not a spec that Shepherd's watching, I can't verify the anchor is right or sub in the actual section heading. Instead you get the hash displayed. |
| 20:32 | <TabAtkins> | I should do something about that, but it depends on me finally having per-spec persistent data. |
| 20:33 | <hober> | right. is there a <pre class='foo'> block i can put in my file with section to pretty section name mappings? |
| 20:33 | <TabAtkins> | Not yet, but that's particularly easy to add. I can put that together this afternoon for you, if you want. |
| 20:34 | <TabAtkins> | Also, your previous request is lucky bug #1234 https://github.com/tabatkins/bikeshed/issues/1234 |
| 20:34 | <hober> | oh, it's super not urgent. but it'd be nice. |
| 20:42 | <TabAtkins> | k, well https://github.com/tabatkins/bikeshed/issues/1235 for ya |
| 20:47 | <hober> | thanks :) |