[webkit-reviews] review denied: [Bug 186104] Stop using StorageTracker.db in LocalStroageDatabaseTracker : [Attachment 341597] Patch

[bugzilla-daemon at webkit.org]

Chris Dumez <cdumez at apple.com> has denied Sihui Liu <sihui_liu at apple.com>'s
request for review:
Bug 186104: Stop using StorageTracker.db in LocalStroageDatabaseTracker
https://bugs.webkit.org/show_bug.cgi?id=186104

Attachment 341597: Patch

https://bugs.webkit.org/attachment.cgi?id=341597&action=review




--- Comment #4 from Chris Dumez <cdumez at apple.com> ---
Comment on attachment 341597
  --> https://bugs.webkit.org/attachment.cgi?id=341597
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=341597&action=review

> Source/WebKit/UIProcess/WebStorage/LocalStorageDatabaseTracker.cpp:63
> +    SQLiteFileSystem::deleteDatabaseFile(databasePath("StorageTracker.db"));

Why are we still doing this on the main thread?

> Source/WebKit/UIProcess/WebStorage/LocalStorageDatabaseTracker.cpp:110
> +    importOrigins();

Why is this OK? This will reimport all origins from disk, even though they have
likely already been imported.

> Source/WebKit/UIProcess/WebStorage/LocalStorageDatabaseTracker.cpp:166
>	   if (!path.endsWith(".localstorage"))

Why is this still here? Can this be made an assertion?

> Source/WebKit/UIProcess/WebStorage/LocalStorageDatabaseTracker.cpp:172
> +	   if (origin)

Should we add a RELEASE_LOG_ERROR() for the case where it fails? It seems like
we'd want to know if we fail to import origins from the disk.