Simplify the SafeBrowsingService logic some:

* |database_loaded_| was entirely unnecessary.  The only ways in which checking it differed from checking |database_| were that it took longer to be set, and if true it meant that ProtocolManager::Initialize() had been called.  Erik and I determined that ProtocolManager::Initialize() was safe to call irrespective of whether the database had finished loading, and that this didn't add additional network burden in the case of a corrupt/unavailable database (especially since in those cases |database_loaded_| would still have been set true).

* |resetting_| can be set to false on the db thread, saving a little bit of code (due to a reorg) and allowing callers to proceed immediately when they happen to call the IO thread before the db thread calls back to it after finishing a reset.

I also made CheckUrl() queue the request when |database_| is NULL instead of just dropping it.  This seems more correct given that |database_| is guaranteed to eventually become non-NULL, so this means we won't skip checking requests that happen right during startup.  I am a bit concerned that this means we may slow startup by delaying all the network requests until the safe browsing DB spins up.  Feedback desired.

BUG=none
TEST=none
Review URL: http://codereview.chromium.org/384110

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@31950 0039d316-1c4b-4281-b951-d872f2087c98

0 files changed, 0 insertions, 0 deletions