DescriptionRevert of Move beforeunload hang timer duties to its own timer. (patchset #6 id:100001 of https://codereview.chromium.org/2738943002/ )
Reason for revert:
Speculative, might have caused https://crbug.com/705306
Original issue's description:
> Move beforeunload hang timer duties to its own timer.
>
> 1. Just because a beforeunload dialog was closed, it doesn't mean that the frame
> necessarily was waiting for a beforeunload ack. (If navigation happened in the
> renderer entirely, it could decide to run beforeunload without the browser
> knowing; the whole "waiting for beforeunload ack" deal is in the situation where
> the browser forces the page to run a beforeunload handler if it has one.) Thus
> blindly restarting the beforeunload ack timer just because it was a beforeunload
> dialog was wrong. Thus this reverts r453323.
>
> 2. r453323 was trying to address a real problem: what if the parent frame was
> waiting for a beforeunload ack but the child frame had the beforeunload handler
> and showed the dialog? But the problem is more general, as _any_ parent frame
> could have been asked to run beforeunload handlers and could have an ack
> pending. So that's what's with the walking up the frame tree to both suspend the
> timer when the dialog is shown, and to resume it when the dialog is closed.
>
> BUG=418266
> TEST=no user-visible change
> CQ_INCLUDE_TRYBOTS=master.tryserver.chromium.linux:linux_site_isolation
>
> Review-Url: https://codereview.chromium.org/2738943002
> Cr-Commit-Position: refs/heads/master@{#457223}
> Committed: https://chromium.googlesource.com/chromium/src/+/2d7d2c3632187e2cfaaa5c6ec26d3adb3ae784b0
TBR=creis@chromium.org,dtapuska@chromium.org,avi@chromium.org
# Not skipping CQ checks because original CL landed more than 1 days ago.
BUG=418266
Patch Set 1 #
Messages
Total messages: 7 (2 generated)
|