This is a reland of e7d5981628
TestExpectations added for the portals test that is hitting crbug.com/838348
TBR=avi@chromium.org, dcheng@chromium.org
Original change's description:
> Handle race of SwapIn() and browser destroying the speculative MainFrame
>
> When WebContentsImpl destroys, it deletes the speculative main frame,
> but the renderer may have taken ownership of it already, and the notice
> of such action is in flight to the browser still. This leads to DCHECKs
> failing in the renderer. So inform the FrameMsg_Delete IPC what the
> intention of the browser is, there are 3 modes:
>
> - Deleting a non-main frame. This is the common case. These frames are
> all owned by the browser so it's all good, no races.
> - Deleting a speculative main frame at shutdown. This is the race we
> address here.
> - Deleting a speculative main frame because it's no longer needed. This
> race is not handled by this CL but we CHECK() it explicitly now instead
> of letting the renderer continue with a missing RenderFrame that it
> expects to be present until it crashes somewhere random later.
>
> TBR=avi@chromium.org, dcheng@chromium.org
>
> Bug: 957858, 838348
> Change-Id: I2110bdaf8b116df48037f69db6cb992fa3796e29
> Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/1594834
> Commit-Queue: danakj <danakj@chromium.org>
> Reviewed-by: Daniel Cheng <dcheng@chromium.org>
> Reviewed-by: Avi Drissman <avi@chromium.org>
> Cr-Commit-Position: refs/heads/master@{#660025}
Bug: 957858, 838348
Change-Id: Iba424cc7be7db053c7ca617677b3cacf972fa067
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/1614132
Commit-Queue: danakj <danakj@chromium.org>
Reviewed-by: Lucas Gadani <lfg@chromium.org>
Reviewed-by: Avi Drissman <avi@chromium.org>
Cr-Commit-Position: refs/heads/master@{#660128}
f23424355b