[webkit-reviews] review denied: [Bug 40655] Database access in worker threads results in WebKit SPI notifications being posted from worker threads : [Attachment 64623] patch

[bugzilla-daemon at webkit.org]

Darin Adler <darin at apple.com> has denied Dumitru Daniliuc <dumi at chromium.org>'s
request for review:
Bug 40655: Database access in worker threads results in WebKit SPI
notifications being posted from worker threads
https://bugs.webkit.org/show_bug.cgi?id=40655

Attachment 64623: patch
https://bugs.webkit.org/attachment.cgi?id=64623&action=review

------- Additional Comments from Darin Adler <darin at apple.com>
What guarantees these main-thread calls don't outlive the
WebDatabaseTrackerClient and WebDatabaseManager objects? Are those immortal
objects? I see a destructor for WebDatabaseTrackerClient, so there seems to be
some code that thinks it can destroy one. It would be bad if it destroyed one
that had an outstanding DidModifyOriginData object.

> +	   DidModifyOriginData* context = new DidModifyOriginData(client,
origin->threadsafeCopy());
> +	  
callOnMainThread(&DidModifyOriginData::dispatchDidModifyOriginOnMainThread,
context);

Soon we'll be implementing a rule where you have to call adoptPtr any time you
do new. We might want to come up with a new idiom for this that avoids the raw
new/delete because we'll have to come back to all these call sites some day.

review- because of the lifetime issue