| 15:42 | <emilio> | jarhar: what's the rationale in https://github.com/whatwg/html/pull/12263 for making moveBefore() more async than appendChild? |
| 15:43 | <emilio> | I was reviewing the Firefox patch implementing this and it feels super weird to have code like: |
| 16:01 | <jjaschke> | annevk: ^ |
| 16:33 | <Noam Rosenthal> | Is it really more async @emilio:mozilla.org? Seems like if you are in isMove then postConnectionSteps won't be called at all |
| 16:34 | <Noam Rosenthal> | I guess it would be more obvious if it was in something like updateInPostMoveSteps but I wasn't part of the details here |
| 16:39 | <emilio> | Noam Rosenthal: right, point being that, on insertion, move steps queue a microtask, but post connection steps do it sync |
| 16:39 | <emilio> | So yes, it's more async, afaict? |
| 16:39 | <emilio> | On removal, both queue a microtask, so that's fine |
| 16:40 | <Noam Rosenthal> | Yea, I thought you meant something else than you did on first read |
| 16:41 | <emilio> | Yeah our code uses the same code with a flag to implement both post-connection and move steps, so the snippet of code doesn't quite read as the spec, sorry :) |
| 16:41 | <emilio> | But it made this mismatch much more obvious :) |
| 16:50 | <jjaschke> | Spec algo is here |
| 16:55 | <jarhar> | yeah the rationale is that we have post-insertion steps, but no post-removal and no post-move steps |
| 16:56 | <jarhar> | if we had post-removal or post-move then we would want to use them, but this edge case for doing this stuff with option elements does not justify creating post-removal and post-move steps imo |
| 16:57 | <jarhar> | we could also avoid using post-insertion steps and make that async too, but it ended up making a normal case during parsing become inconsistent about when the options get cloned: https://github.com/whatwg/html/pull/12263#issuecomment-4743852899 |
| 17:40 | <anaskim> | Hi all, as I said in the last WHATNOT, I updated the explainer for platform-provided behaviors for CE https://github.com/MicrosoftEdge/MSEdgeExplainers/blob/main/PlatformProvidedBehaviors/explainer.md and I'd like to get initial feedback on this direction. Let me know what you think 🙂 |
| 20:44 | <Noam Rosenthal> | Might be good to have this as a note in the spec in the place where we queue a microtask |