Use an explicit boolean has_new_first_party_for_cookies instead...

Use an explicit boolean has_new_first_party_for_cookies instead
of an empty, invalid URL to indicate whether the first party for
cookies URL needs changing when following a redirect.

R=eroman
BUG=25133
TEST=none

Committed: http://src.chromium.org/viewvc/chrome?view=rev&revision=32260

[eroman, 2009-11-18 01:58:37]

lgtm

http://codereview.chromium.org/405011/diff/1/11
File chrome/browser/renderer_host/resource_dispatcher_host.h (right):

http://codereview.chromium.org/405011/diff/1/11#newcode149
Line 149: // has_new_first_party_for_cookies.
For the last sentence, can you add:

"and new_first_party_for_cookies will not be used."

http://codereview.chromium.org/405011/diff/1/15
File chrome/common/resource_dispatcher_unittest.cc (right):

http://codereview.chromium.org/405011/diff/1/15#newcode40
Line 40: // TODO(wtc): should we return a new first party for cookies URL?
Why is this a TODO? Doesn't seem like we need to do anything here.

http://codereview.chromium.org/405011/diff/1/15#newcode41
Line 41: *has_new_first_party_for_cookies = false;
Since you default it to false in the callers, I guess this is not necessary (I
don't mind either way).

http://codereview.chromium.org/405011/diff/1/15#newcode253
Line 253: // TODO(wtc): should we return a new first party for cookies URL?
Same comment as above.

[wtc, 2009-11-18 02:25:50]

http://codereview.chromium.org/405011/diff/1/15
File chrome/common/resource_dispatcher_unittest.cc (right):

http://codereview.chromium.org/405011/diff/1/15#newcode40
Line 40: // TODO(wtc): should we return a new first party for cookies URL?
On 2009/11/18 01:58:37, eroman wrote:
> Why is this a TODO? Doesn't seem like we need to do anything here.

I removed the two TODO comments in this file.

http://codereview.chromium.org/405011/diff/1/15#newcode41
Line 41: *has_new_first_party_for_cookies = false;
On 2009/11/18 01:58:37, eroman wrote:
> Since you default it to false in the callers, I guess this is not necessary (I
> don't mind either way).

In general I think a function should set out parameters,
rather than relying on the caller initializing out
parameters to some values.

[darin (slow to review), 2009-11-18 04:42:56]

The CL description doesn't say why you are making this change.  I couldn't grok
it from the bug either.  Can you amend the CL description with the reason for
this change?  Does it fix a bug or does it just make the code more clear? 
Thanks!