Add an exact search method to the Keychain adapter, and modify unit tests to ...

Add an exact search method to the Keychain adapter, and modify unit tests to us that instead of the raw keychain version.
Overhaul the single-item-search test to be shorter, clearer, and more complete.
Convert more namespaced methods that now have no public callers to private helpers of the Keychain adapter, and refactor the helpers.

BUG=none
TEST=none; no behavioral changes

Committed: http://src.chromium.org/viewvc/chrome?view=rev&revision=19964

[pink (ping after 24hrs), 2009-07-06 14:34:59]

LGTM

http://codereview.chromium.org/149160/diff/1/2
File chrome/browser/password_manager/password_store_mac_internal.h (right):

http://codereview.chromium.org/149160/diff/1/2#newcode49
Line 49: // them. The caller is responsible for Free-ing the returned items.
does this mean they should call free() on them, or some other method of cleanup?
It's a little confusing given your wording. I guess Free() is a method on
MacKeychain?

http://codereview.chromium.org/149160/diff/1/2#newcode70
Line 70: // If the return value is false, the state of the out params is
undefined.
Are you allowed to pass NULL for any of the out params if you don't care about
the value?

[stuartmorgan, 2009-07-06 17:26:34]

http://codereview.chromium.org/149160/diff/1/2
File chrome/browser/password_manager/password_store_mac_internal.h (right):

http://codereview.chromium.org/149160/diff/1/2#newcode49
Line 49: // them. The caller is responsible for Free-ing the returned items.
On 2009/07/06 14:34:59, pink wrote:
> does this mean they should call free() on them, or some other method of
cleanup?
> It's a little confusing given your wording. I guess Free() is a method on
> MacKeychain?

Yes, it's MacKeychain method. Wording clarified.

http://codereview.chromium.org/149160/diff/1/2#newcode70
Line 70: // If the return value is false, the state of the out params is
undefined.
On 2009/07/06 14:34:59, pink wrote:
> Are you allowed to pass NULL for any of the out params if you don't care about
> the value?

Yes, the "all out params that are non-NULL will be set" was intended to indicate
that. Suggestions for clearer wording are welcome :)

[pink (ping after 24hrs), 2009-07-06 20:15:46]

still LGTM

http://codereview.chromium.org/149160/diff/1/2
File chrome/browser/password_manager/password_store_mac_internal.h (right):

http://codereview.chromium.org/149160/diff/1/2#newcode70
Line 70: // If the return value is false, the state of the out params is
undefined.
how did i not notice that?! oy.