| 09:43 | <nicolo-ribaudo> | Why does the HTML spec link to https://drafts.fxtf.org rather than to the specs on w3.org? |
| 09:44 | <nicolo-ribaudo> | The <filter-value-list> definition we link to is giving me an error 500, but probably the canonical one should be https://www.w3.org/TR/filter-effects-1/#typedef-filter-value-list rather than https://drafts.fxtf.org/filter-effects/#typedef-filter-value-list |
| 09:44 | <sideshowbarker> | We can’t trust the W3C /TR versions to be up to date |
| 09:45 | <sideshowbarker> | but hang on |
| 09:46 | <sideshowbarker> | the canonical versions (latest versions) should rightly just be getting served from https://w3c.github.io/fxtf-drafts |
| 09:47 | <sideshowbarker> | I think Andreu Botella may have a back-burner thing going to make that happen |
| 09:48 | <sideshowbarker> | the problem is basically that some W3C groups thought it would be a fun good idea to mint their own vanity domains |
| 09:48 | <nicolo-ribaudo> | Oh actually for this one specifically it's probably https://drafts.csswg.org/filter-effects-1/ |
| 09:48 | <sideshowbarker> | ah yeah maybe problem solved then, in this particular case |
| 09:50 | <sideshowbarker> | but in the general case, they’ve kind of made of made it confusing as hell for any implementors to clearly figure out which spec/version/level they’re supposed to be implementing from |
| 09:50 | <sideshowbarker> | so I hope we unwind this all eventually |
| 09:51 | <nicolo-ribaudo> | Ok it seems like for all fxtf drafts the canonical URL is on the csswg.org website |
| 09:52 | <nicolo-ribaudo> | At list according to the link in the metadata of their repo |
| 09:52 | <sideshowbarker> | well we really need to just EOL that entire domain |
| 09:52 | <sideshowbarker> | well maybe it’s clear enough now for you at least to move on with what you were trying to do |
| 09:53 | <sideshowbarker> | but we still have a chronic problem that newer implementors have to go through this idiot dance every single time, N times |
| 11:19 | <nektro> | is https://github.com/web-platform-tests/wpt/blob/master/url/resources/IdnaTestV2.json#L168-L172 being used per https://github.com/web-platform-tests/wpt/blob/master/url/IdnaTestV2.window.js#L8-L10 ? |
| 12:53 | <annevk> | nektro: not for that particular test runner. But you could imagine hooking this JSON resource directly to your C++ implementation in some manner instead of going through JavaScript. |
| 13:22 | <Andreu Botella> | they were moved to drafts.csswg.org last month or so |
| 13:23 | <Andreu Botella> | after drafts.fxtf.org being broken for a while |
| 13:23 | <Andreu Botella> | I need to get around to that at some point, yeah... |
| 17:19 | <smaug> | Stephen Chenney: You might recall something about https://issues.chromium.org/issues/404183708. Chrome seems to behave differently to others when it comes to ::selection in light DOM and selection happening in shadow DOM. Is there perhaps some spec issue about this? |
| 17:25 | <smaug> | Looks like there is at least this https://github.com/w3c/csswg-drafts/issues/12497 |
| 17:34 | <smaug> | (ok, I was told elsewhere that something might have changed in some spec at some point but I can't quite find a proper spec for this) |
| 18:11 | <Noam Rosenthal> | https://drafts.csswg.org/css-pseudo-4/#highlight-pseudo-element:~:text=The%20highlight%20pseudo%2Delements%20do%20not%20necessarily%20fit%20into%20the%20element%20tree%2C%20and%20can%20arbitrarily%20cross%20element%20boundaries%20without%20honoring%20its%20nesting%20structure. seems relevant |
| 18:11 | <Noam Rosenthal> | ... though a bit handwavy |
| 21:21 | <smaug> | "a bit" 😛 |
| 21:32 | <smaug> | ChatGPT nor Claude seem to find the answer either. It is quite well hidden. |