Avoid a NOT_REACHED() hit when we get an empty 404 page and

Avoid a NOT_REACHED() hit when we get an empty 404 page and
link doctor is down or disabled.

We weren't fully specifying WebURLError, which caused us to
be confused when looking up the strings for the error code.

We still don't have a nice error message for this, but at least
we don't DCHECK in debug and this will now at least cause the
local error page to say there was an http 404 error.

BUG=76457
Committed: http://src.chromium.org/viewvc/chrome?view=rev&revision=88543

[mmenke, 2011-06-08 22:06:07]

LGTM

[tony, 2011-06-08 22:08:26]

+more content owners

[jam, 2011-06-09 00:12:20]

lgtm

[darin (slow to review), 2011-06-13 21:35:40]

In this case, maybe we should just try to leave the original 404 page, as served
to us by the web server, visible to the user instead of trying to show our own
lame error page instead?

[tony, 2011-06-13 21:39:18]

On 2011/06/13 21:35:40, darin wrote:
> In this case, maybe we should just try to leave the original 404 page, as
served
> to us by the web server, visible to the user instead of trying to show our own
> lame error page instead?

We only show our lame error page if the web server returns no content (a 404 in
the header, but no content).

We removed the 512 byte check a few months ago since IE7+ doesn't do the 512
byte check either.

[darin (slow to review), 2011-06-13 21:48:05]

On Mon, Jun 13, 2011 at 2:39 PM, <tony@chromium.org> wrote:

> On 2011/06/13 21:35:40, darin wrote:
>
>> In this case, maybe we should just try to leave the original 404 page, as
>>
> served
>
>> to us by the web server, visible to the user instead of trying to show our
>> own
>> lame error page instead?
>>
>
> We only show our lame error page if the web server returns no content (a
> 404 in
> the header, but no content).
>
> We removed the 512 byte check a few months ago since IE7+ doesn't do the
> 512
> byte check either.
>
>
I see.  If didReceiveDocumentData is called, then we will show the original
404 page.

SGTM



>
>
http://codereview.chromium.**org/6995091/<http://codereview.chromium.org/6995...
>