GitLab

This reverts commit 8c5fb8cb.

Reason for revert: DCHECK(!installed_native_timer_); in dev builds
when moving and resizing windows. Seems interactive_ui_tests might
not cover those cases..?

Original change's description:
> [base] Migrate the Win MessagePump to DoSomeWork
>
> This CL migrates both MessagePumpForUI and ForIO as they use the same
> base class and it's easier to do both at once.
>
> Tweaks to MessagePumpBase and/or both pumps:
>  * Use DoSomeWork() instead of DoWork()/DoDelatedWork()
>  * Replace |delayed_work_time_| by a timeout value computed on the stack
>    from the result of DoSomeWork(). Conversely making
>    ScheduleDelayedWork() a no-op as desired.
>  * Refactor GetCurrentDelay() into GetSleepTimeoutMs() and use modern
>    safe_conversions.h to perform the math.
>
> MessagePumpForUI specific tweaks:
>  * Fixed crbug.com/929263: ::SetTimer() used 0 as an nIDEvent but
>    ::KillTimer() used |this| so they weren't even interacting...
>  * Fixed bug: even with the previous bug fixed. HandleTimerMessage()
>    could be invoked while in a non-native DoRunLoop() (e.g. since
>    KillTimer doesn't remove WM_TIMER messages already enqueued) and as
>    such result in redundant sources of scheduling since it would then
>    continue to RescheduleTimer() and confusing state for DoRunLoop()).
>  * Ignore WM_TIMER messages which weren't scheduled by our ::SetTimer().
>  * Simplified usage of native ::SetTimer(). We now *only* rely on it
>    when in a native nested loop and kill it immediately after to avoid
>    double-scheduling.
>
> Note: MessageLoopTest.PostDelayedTask_SharedTimer_SubPump tests the
> necessity for a native timer when entering a native nested loop and
> confirms that the new behavior is WAI.
>
> Also adding regression tests for a tricky use case (and finding that
> it's already broken on ToT..!). Filed crbug.com/930940 to follow-up.
>
> Bug: 885371, 929263, 930940
> Change-Id: I7f2d18a1a5a494c6a4f31fcc0fad9b4de25d70a7
> Reviewed-on: https://chromium-review.googlesource.com/c/1455266
> Reviewed-by: Alex Clarke <alexclarke@chromium.org>
> Reviewed-by: François Doray <fdoray@chromium.org>
> Commit-Queue: Gabriel Charette <gab@chromium.org>
> Auto-Submit: Gabriel Charette <gab@chromium.org>
> Cr-Commit-Position: refs/heads/master@{#631231}

TBR=gab@chromium.org,fdoray@chromium.org,alexclarke@chromium.org

# Not skipping CQ checks because original CL landed > 1 day ago.

Bug: 885371, 929263, 930940
Change-Id: Ic179d0493d01d4f44330927d9ccdb0f1454d2e78
Reviewed-on: https://chromium-review.googlesource.com/c/1471254Reviewed-by: Gabriel Charette <gab@chromium.org>
Commit-Queue: Gabriel Charette <gab@chromium.org>
Cr-Commit-Position: refs/heads/master@{#631885}