GitLab

The patch [1] might introduce the kInitialDocument state to the
RenderFrameHost. During code review, I was surprised not seeing the
transition: kInitialDocument -> kRunningUnloadHandler.

In theory, executing an unload handler from the initial empty document
can happen. This patch adds a test showing it is possible.

This turns out not to be crashing against [1]. The reason: the initial
"fake" navigation for empty document IPC is received in between the
creation and the deletion of the frame. This makes the RFH state to
transition to kActive, just in time before the deletion.

Even if this patch didn't prove its point, it it still a good addition.
This might become more useful after the removal of the "fake" initial
navigation (see [2]).

R:acolwell@chromium.org
CC:alexmos@chromium.org, rakina@chromium.org, sreeja@chromium.org

[1]: https://chromium-review.googlesource.com/c/chromium/src/+/2430976/5
[2]: https://docs.google.com/document/d/1UjdxiDJ3sgVDbp0CZpsSntdwAYNMSfc2qajX4nzXfLY

Bug: 1113357
Change-Id: Iaf146c77b767b3af37401dab968fb536efba5293
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2449474
Commit-Queue: Arthur Sonzogni <arthursonzogni@chromium.org>
Reviewed-by: Aaron Colwell <acolwell@chromium.org>
Reviewed-by: Nasko Oskov <nasko@chromium.org>
Auto-Submit: Arthur Sonzogni <arthursonzogni@chromium.org>
Cr-Commit-Position: refs/heads/master@{#814582}