Eliminate paint timing monkey patching #3923
Conversation
|
Hmmm, |
|
If you look at the Travis logs, you'll see that the CI does more-expensive HTML conformance checking which isn't done at build time. In particular: |
|
Thanks, fixed the issue. |
There was a problem hiding this comment.
@tdresser, this is quite nice; I'm really happy with what we came up with here.
I pushed a few separate fixup commits, separated for easier review. Note the source formatting one for future reference, and check me that "present" -> "presentable" was correct.
There's a separate issue with paint timing where you're passing "document" to the "mark paint timing" algorithm, but that spec is not very clear about how things are scoped (document? browsing context? global?). I'll file that issue over there, although I'm pretty sure it won't impact this. So I'm OK merging this as soon as you confirm my changes.
source
Outdated
| @@ -89002,7 +89002,7 @@ dictionary <dfn>PromiseRejectionEventInit</dfn> : <span>EventInit</span> { | |||
| <p><i>Rendering opportunites</i>: If there are <span data-x="browsing context">browsing | |||
| contexts</span> <var>B</var> that do not have a <span>rendering opportunity</span>, then | |||
| remove from <var>docs</var> all <code>Document</code> objects whose <span | |||
| data-x="concept-document-bc">browsing context</span> is in <var>B</var>.</p> | |||
| data-x="concept-document-bc">browsing context</span> is <var>B</var>.</p> | |||
There was a problem hiding this comment.
In this context, B is a list of browsing contexts, so I think this was correct previously?
There was a problem hiding this comment.
Oh, great catch. I'll rename to "browsingContexts" for clarity. (There is also a "B" up in step 7.2, which technically doesn't collide.)
|
Yeah, we'll need to clear up paint timing once this has landed. Thanks for fixing the indenting, I'll be more intentional about making sure it's sane next time. |
This PR removes the "Modifications to other specifications" section because whatwg/html#3923 has already landed, so the HTML spec change does not need to be described here.
This eliminates paint timing monkey patching, and fixes some minor issues with when we mark paint timing.
This is necessary for us to correctly report event duration for the event timing spec. Issue here.
@domenic - you took a look at this here.
/acknowledgements.html ( diff )
/infrastructure.html ( diff )
/webappapis.html ( diff )