Appcache::OnCorruptionDetected handling. The general idea is to delete everything and start over (reinitialize) if we notice corruption on disk. The is some pre-existing code to perform the reinitialization function, but that was only used in a more narrow case where the disk cache could not have been open. With the changes in this CL, reinitialization can be triggered pretty much at any time.

 
* The AppCacheDatabase classes uses an SQLConnecton ErrorHandler to detect catastrophic errors (corruption).

* The AppCacheStorageImpl's DatabaseTask class checks for having seen corruption during a task and initiates a heavy handed operation to delete everything and start over.

* The AppCacheStorageImpl::InitTask also checks for the presence of a stale/undeletable DiskCache when there is no SQL database. If found, that also triggers reinitialization.

* The AppCacheService class does exponential backoff to avoid thrashing the disk on repeated reinit attempts.

* The AppCacheDiskCache class's Disable() method has been modified such that all file handles get released.

TBR=jam
BUG=318544

Review URL: https://codereview.chromium.org/137493003

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@247967 0039d316-1c4b-4281-b951-d872f2087c98

0 files changed, 0 insertions, 0 deletions