Add a second line of defense for receiving a bad message in the renderer.

Add a second line of defense for receiving a bad message in the renderer.

BUG=88949
TEST=none

Committed: http://src.chromium.org/viewvc/chrome?view=rev&revision=104010

[Charlie Reis, 2011-10-04 20:45:01]

I used the same IPC_BEGIN_MESSAGE_EX approach as RenderViewHost to handle the
failed IPC deserialization, with a CHECK(false) to kill the renderer if it
happens.

I also moved Karl's kMaxURLChars check in TabContents above the
render_manager_.Navigate line so that we wouldn't do unnecessary work in
RenderViewHostManager (like firing up a pending RVH) for a URL we won't visit.

Justin, can you review the CHECK?  Brett, can you take a look for owner's
approval?  Thanks!

[jschuh, 2011-10-04 21:08:10]

One documentation nit. Other than that lgtm.

http://codereview.chromium.org/8142009/diff/1/content/renderer/render_view.cc
File content/renderer/render_view.cc (right):

http://codereview.chromium.org/8142009/diff/1/content/renderer/render_view.cc...
content/renderer/render_view.cc:695: // Kill the renderer.
Nit: "Kill the renderer" seems too terse. I worry someone might be inclined to
remove it. How about something like "kill the renderer to avoid potential
spoofing attacks."

[brettw, 2011-10-04 21:56:30]

lgtm

[Charlie Reis, 2011-10-04 21:58:38]

Thanks!

http://codereview.chromium.org/8142009/diff/1/content/renderer/render_view.cc
File content/renderer/render_view.cc (right):

http://codereview.chromium.org/8142009/diff/1/content/renderer/render_view.cc...
content/renderer/render_view.cc:695: // Kill the renderer.
On 2011/10/04 21:08:10, Justin Schuh wrote:
> Nit: "Kill the renderer" seems too terse. I worry someone might be inclined to
> remove it. How about something like "kill the renderer to avoid potential
> spoofing attacks."

Done.