Initialize the simple cache backend at runtime.

We will need to choose the backend at runtime for A/B testing. This change is 
the first step towards backend comparisons: it moves the logic of choosing among 
cache backends to a single point (cache_creator.cc) and makes it logically 
separate from BackendImpl. Next step would be to inject performance-analyzer 
backend into this cache backend choice logic. The analyzer is going to be under 
a compile-time flag. 

With this change the Simple Cache Backend operation gets enabled when the 
"SimpleCacheTrial" experiment is set.

Cache Initialization. 

The Simple Cache Backend now shares the common initialization/cleanup code. So 
it can delete inconsistent cache, cache of a different version etc 
asynchronously. It is just a positive side-effect of moving the backend choice 
to a single point. 

The CacheCreator with almost no change. 

To see the difference:
shell> out/Debug/chrome --enable-logging=stderr --v=1 \ 
  --user-data-dir=/tmp/unique --use-simple-cache-backend=on \ 
  http://url 2>&1 | grep "Simple Cache"

BUG=173390, 173384
TBR=maruel

Review URL: https://chromiumcodereview.appspot.com/12794003

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@191830 0039d316-1c4b-4281-b951-d872f2087c98

1 files changed, 1 insertions, 0 deletions

diff --git a/PRESUBMIT.py b/PRESUBMIT.py
index b304285..3c62eb2 100644
--- a/PRESUBMIT.py
+++ b/PRESUBMIT.py
@@ -159,6 +159,7 @@ _BANNED_CPP_FUNCTIONS = (
       True,
       (
         r"^content[\\\/]shell[\\\/]shell_browser_main\.cc$",
+        r"^net[\\\/]disk_cache[\\\/]cache_util\.cc$",
       ),
     ),
 )