Ensure that BrowserMessageFilter isn't used to process a sync message on the UI thread.

Ensure that BrowserMessageFilter isn't used to process a sync message on the UI thread.
Committed: http://src.chromium.org/viewvc/chrome?view=rev&revision=79470

[brettw, 2011-03-25 05:33:26]

http://codereview.chromium.org/6713121/diff/1/content/browser/browser_message...
File content/browser/browser_message_filter.cc (right):

http://codereview.chromium.org/6713121/diff/1/content/browser/browser_message...
content/browser/browser_message_filter.cc:73: return true;
It seemed weird to me that this doesn't assert, but the "CanBe..." function
does. Either the assert should be moved here, or maybe the "CanBe" function
should be "Verify"

[jam, 2011-03-25 17:08:53]

http://codereview.chromium.org/6713121/diff/1/content/browser/browser_message...
File content/browser/browser_message_filter.cc (right):

http://codereview.chromium.org/6713121/diff/1/content/browser/browser_message...
content/browser/browser_message_filter.cc:73: return true;
On 2011/03/25 05:33:26, brettw wrote:
> It seemed weird to me that this doesn't assert, but the "CanBe..." function
> does. Either the assert should be moved here, or maybe the "CanBe" function
> should be "Verify"

I don't want the assert here since that code is shared with RVH, and I want to
keep the assert beside the big comment.

prefixing Verify to MessageCanBeDispatchedOnUI seems redundant, I'm not sure how
it helps to see that the function asserts?

[jam, 2011-03-25 21:39:14]

ping

[brettw, 2011-03-25 21:46:40]

I just think a function MessageCanBeDispatchedOnUI should return a
true or false value and I don't expect it to assert one way or
another. If you don't like VerifyMessageCanBeDispatchedOnUI or
Ensure..., can you think of a different name?

Brett

[jam, 2011-03-25 21:52:45]

I think the issue is that when I see "Verify", I don't expect it to assert
as well, hence my pushback.

On Fri, Mar 25, 2011 at 2:46 PM, Brett Wilson <brettw@chromium.org> wrote:

> I just think a function MessageCanBeDispatchedOnUI should return a
> true or false value and I don't expect it to assert one way or
> another. If you don't like VerifyMessageCanBeDispatchedOnUI or
> Ensure..., can you think of a different name?
>
> Brett
>

[brettw, 2011-03-25 23:19:45]

LGTM with CheckCanDispatchOnUI