[css-contain-2] WIP: Stop monkey-patching the HTML spec. #12105
Conversation
This prevents bikeshed from generating the spec locally.
| 2. Set <var>element</var>'s [=currently relevant to the user=] flag to true. | ||
| 1. If <var>isFirstIteration</var> is false, then [=continue=]. | ||
| 1. If <var>element</var> is [=relevant to the user=], then: | ||
| 1. If <var>element</var>'s [=currently relevant to the user=] flag is false, set <var>changed</var> to true. |
There was a problem hiding this comment.
This condition seems to be the same as the second condition above
If element is not [=currently relevant to the user=] and element is [=relevant to the user=], then
in fact I don't think it would hit since we update currently relevant to the user on line 2009. Or am I missing a subtlety here?
There was a problem hiding this comment.
Yeah I think this branch is dead effectively, but just because of the mistake you pointed out above.
| 1. If <var>element</var> is not [=currently relevant to the user=] and <var>element</var> is [=relevant to the user=], then: | ||
| 2. Set <var>changed</var> to true. | ||
| 2. Set <var>element</var>'s [=currently relevant to the user=] flag to true. | ||
| 1. If <var>isFirstIteration</var> is false, then [=continue=]. |
There was a problem hiding this comment.
Because of the comment below, it's unclear to me the role isFirstIteration is playing here. changed is set in either case
There was a problem hiding this comment.
The important bit is that if isFirstIteration is false, then non-initial determination is not supposed to happen (this is what implements the semantics you were talking about in your other comments). I think I just botched the condition in line 2007 (it's supposed to deal only with the initial proximity determination).
I'll fix, thanks for the review.
| @@ -2308,14 +2339,7 @@ Restrictions and Clarifications {#cv-notes} | |||
| </wpt> | |||
|
|
|||
| 3. If an element starts or stops [=skipped contents|skipping its contents=], | |||
| this change happens | |||
| after the requestAnimationFrame callbacks | |||
| of the frame that renders the effects of the change | |||
There was a problem hiding this comment.
So there's a bit of a subtlety here. What this is saying is that if we have frames 1 and 2 (separate update the rendering steps), and we determine that an element is relevant in step 1, then the effect of that relevance (ie it allows rendering, doesn't have strict containment etc) happen on frame 2. The difference is that if you try and observe the size of the element between rAFs in a timeout, then between frame 1 and 2 it should act as if it's still not relevant wrt size containment. See the note in this spec below which tries to explain that.
I don't mind if we change this behavior, but we need a larger discussion for that.
There was a problem hiding this comment.
That's what the intention of the currently relevant to the user flag is (maybe I botched the language).
"currently relevant to the user" is meant to be the thing that has effect on the rendering. But you're right I might have messed up my prose ;)
This exports an "update content relevancy for a document" algorithm that HTML can call at the right time.