| 11:42 | <Luke Warlow> | What's the correct way to preserve links to definitions when you change their contents in the HTML spec? I seem to recall there being some legacy-id attribute or something but can't seem to find it. |
| 12:06 | <annevk> | Luke Warlow: just use an id attribute on the dfn |
| 12:07 | <Luke Warlow> | But then the "new" name doesn't get linkified iirc? I guess that's probably okay? Context is HTMLOrSvgElement getting renamed |
| 12:18 | <annevk> | Luke Warlow: yes it will. It will just use the old ID for the links, which is fine. We have a lot of history captured in fragment identifiers. |
| 15:12 | <annevk> | Ms2ger: where is serialization and deserialization of WebAssembly.Memory defined? |
| 15:20 | <Ms2ger> | Uhhh |
| 15:21 | <Ms2ger> | I assumed https://webassembly.github.io/spec/web-api/#serialization but that's Module |
| 15:22 | <Ms2ger> | Is it serializable? |
| 15:25 | <annevk> | Ms2ger: WebKit has code for it, in Firefox I get an exception that suggests it is (but it might be about SharedArrayBuffer instead). I should probably create a decent test; COOP + COEP makes that a bit awkward. |
| 15:47 | <annevk> | Ms2ger: yeah, so Firefox definitely preserves the object brand through structuredClone() (WebAssembly.Memory) |
| 15:47 | <annevk> | I couldn't find test coverage on WPT, but somehow everyone is doing this? |
| 15:52 | <annevk> | https://github.com/WebAssembly/spec/issues/2155 |
| 15:55 | <Ms2ger> | Thanks |
| 15:56 | <Ms2ger> | It does ring a vague bell to me, but I can't remember the situation |
| 21:24 | <Andreu Botella> | Hey, I'm trying to understand some of the media element handling. In a media element's internal play steps, the resource selection algorithm is invoked if networkState is NETWORK_EMPTY, but as far as I can tell, in all cases where this happens, the resource selection algorithm would end up returning early and setting networkState back to NETWORK_EMPTY. If so, why invoke it? |
| 21:25 | <Andreu Botella> | It's possible that the src attribute gets set immediately after e.g. calling play(), sure, but that would also cause another invocation of the resource selection algorithm. |
| 21:26 | <Andreu Botella> | is there something I'm missing here? |