refactor: simplify the Lifecycle Watcher #12295
Conversation
Lightning00Blade
force-pushed
the
simplify-the-watcher
branch
4 times, most recently
from
ab924b9
to
d040223
Compare
Lightning00Blade
force-pushed
the
simplify-the-watcher
branch
from
d040223
to
9d15d3f
Compare
| // CDP provided the correct order for Loading Events | ||
| // And NetworkIdle is a global state | ||
| // Consider removing | ||
| for (const child of frame.childFrames()) { |
There was a problem hiding this comment.
so was it here only for the empty frames? isn't this code also checks for networkidle events which could come in any order of the frames/subframes? Is it definitely not possible that the load or domcontent loader events arrive after the main frame load/domcontentloaded?
There was a problem hiding this comment.
So for the Lifecycle events they now (after the linked CL) come in order, that's only true for parsing so anything coming from the HTML or Scripts (evaluateOnNewDocument included) will be correct.
As for the NetworkIdle I have to double check but I think the state for every frame is track on the same NetworkIdle detector so it's global for the page (comment also suggest that).
There was a problem hiding this comment.
I am not sure it can be tracked globally with OOPIFs
There was a problem hiding this comment.
Yeah you may be correct. I don't think we have such test. I will write one here.
We can expect the correct order of event for
DOMContentLoadedandloadevents,that the main frame's will come last upstream CL.