[DevTools] Preserve DevTools for crashed page.

[DevTools] Preserve DevTools for crashed page.

BUG=461774
Committed: https://crrev.com/f551ed7462e214be796648de39c44e2788b71356
Cr-Commit-Position: refs/heads/master@{#334604}

[dgozman, 2015-06-11 13:03:26]

dgozman@chromium.org changed reviewers:
+ nasko@chromium.org, pfeldman@chromium.org

[dgozman, 2015-06-11 13:03:26]

Take a look please.

[nasko, 2015-06-11 16:56:12]

https://codereview.chromium.org/1181673002/diff/20001/content/browser/devtool...
File content/browser/devtools/render_frame_devtools_agent_host.cc (right):

https://codereview.chromium.org/1181673002/diff/20001/content/browser/devtool...
content/browser/devtools/render_frame_devtools_agent_host.cc:367: frame_crashed_
= false;
Does this boolean track the pending or the current one? If the main frame has
crashed and a pending cross-site navigation has started but is now cancelled,
the main frame can still be in crashed state, no?

[dgozman, 2015-06-15 14:59:06]

PTAL

https://codereview.chromium.org/1181673002/diff/20001/content/browser/devtool...
File content/browser/devtools/render_frame_devtools_agent_host.cc (right):

https://codereview.chromium.org/1181673002/diff/20001/content/browser/devtool...
content/browser/devtools/render_frame_devtools_agent_host.cc:367: frame_crashed_
= false;
On 2015/06/11 16:56:12, nasko (slow) wrote:
> Does this boolean track the pending or the current one? If the main frame has
> crashed and a pending cross-site navigation has started but is now cancelled,
> the main frame can still be in crashed state, no?

It's intended to be the current one. RenderProcessGone does not give any
information about which one has crashed though.
Renamed to |current_frame_crashed_|, removed this line.

[nasko, 2015-06-15 22:25:50]

LGTM with one question.

https://codereview.chromium.org/1181673002/diff/40001/content/browser/devtool...
File content/browser/devtools/render_frame_devtools_agent_host.cc (right):

https://codereview.chromium.org/1181673002/diff/40001/content/browser/devtool...
content/browser/devtools/render_frame_devtools_agent_host.cc:473: if (old_host
== new_host && !current_frame_crashed_)
Will it really make a difference whether the current frame has crashed or not?

[dgozman, 2015-06-16 13:30:36]

https://codereview.chromium.org/1181673002/diff/40001/content/browser/devtool...
File content/browser/devtools/render_frame_devtools_agent_host.cc (right):

https://codereview.chromium.org/1181673002/diff/40001/content/browser/devtool...
content/browser/devtools/render_frame_devtools_agent_host.cc:473: if (old_host
== new_host && !current_frame_crashed_)
On 2015/06/15 22:25:49, nasko (slow) wrote:
> Will it really make a difference whether the current frame has crashed or not?

Yes, because if the current frame has crashed, renderer will be recreated on the
navigation and we should reattach there.

[dgozman, 2015-06-16 13:30:41]

The CQ bit was checked by dgozman@chromium.org

[commit-bot: I haz the power, 2015-06-16 13:31:09]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1181673002/40001

[commit-bot: I haz the power, 2015-06-16 15:06:57]

Committed patchset #3 (id:40001)

[commit-bot: I haz the power, 2015-06-16 15:07:56]

Patchset 3 (id:??) landed as
https://crrev.com/f551ed7462e214be796648de39c44e2788b71356
Cr-Commit-Position: refs/heads/master@{#334604}