MSE may input partial media segment, which could cause the WebMDemuxer and libnestegg to error upon encountering an incomplete block which can't be recovered from.
this will allow to limit read to known complete blocks.
A typical non-fragmented mp4 would have the ftyp atom located at the beginning of the mp4 and the moov at the end. We would to calculate the location of the metadata by spanning the byte range of the two atoms.
As such, we would typically allocate an amount of memory equivalent to the size of the mp4.
Instead we now reconstruct the metadata to only have the ftyp and moov atoms contiguously.
The spec for getRegistration() and getRegistrations() states:
"Resolve promise with a ServiceWorkerRegistration object, setting its
service worker client to service worker client, which represents
registration."
indicating a new ServiceWorkerRegistration is created per call. The existing
Blink test depends on Blink's implementation detail which caches the
registration.
The second test, which checks for NetworkError is incorrect.
[[Update]] step 12 states:
"If response is a network error or response's status is not in the range 200 to 299, then:
1. Reject p with a TypeError."
which specifically converts the NetworkError (due to 404) to a TypeError.
Also fixes controller-on-reload.https.html
Update web-platform-tests expected data
A typical non-fragmented mp4 would have the ftyp atom located at the beginning of the mp4 and the moov at the end. We would to calculate the location of the metadata by spanning the byte range of the two atoms.
As such, we would typically allocate an amount of memory equivalent to the size of the mp4.
Instead we now reconstruct the metadata to only have the ftyp and moov atoms contiguously.
The only logic change is that we now call UpdateCaret() before
dispatching CaretStateChangedEvent.
This resolves a bug that the text selection dialog flashes when long
tapping on an empty content.
There's a bug that when a frame is focusable but not selectable, we
won't focus on it because we call IsSelectable() before ChangeFocus().
By moving the check into SelectWord(), we'll have a chance to focus on
it.
This resolves a issue that when long press to select a word on a new
opened app, the selection highlight is gray instead of blue.
We want the focus changing behavior by long tap as close as to the one
by single tap.
The only functional change is that we always clear old focus and
re-focus the window if a focusable frame cannot be found. This behavior
is the same as the single tap implemented in
EventStateManager::PostHandleEvent().
Besides, ChangeFocus now returns the new focusable frame instead of bool
which provides more information.
