Bug 165192 – Make ApplicationCacheHost::maybeLoadSynchronously more robust

• Status: RESOLVED
• Resolution: FIXED
• Component: WebCore Misc.

[youenn fablet]

In case out parameter "data" is not null and shouldLoadResourceFromApplicationCache returns false and resource is null, we might dereference a null pointer.
"data" out parameter cannot be null right now but we should not expect callers to always do that

[youenn fablet]

Created attachment 295712 [details]
Patch

[Alex Christensen]

Comment on attachment 295712 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=295712&action=review

We should also stop using ResourceHandle directly with ApplicationCacheGroup

> Source/WebCore/ChangeLog:10
> +        We refactor the code to ensure that this case cannot happen anymore.

If it can happen now, we should add a test that reproduces it before this change.

[youenn fablet]

(In reply to comment #2)
> Comment on attachment 295712 [details]
> Patch
> 
> View in context:
> https://bugs.webkit.org/attachment.cgi?id=295712&action=review
> 
> We should also stop using ResourceHandle directly with ApplicationCacheGroup
> 
> > Source/WebCore/ChangeLog:10
> > +        We refactor the code to ensure that this case cannot happen anymore.
> 
> If it can happen now, we should add a test that reproduces it before this
> change.

I don't think this can happen.
This change is coming from discussions with Keith based on some static analysis he did.

[WebKit Commit Bot]

Comment on attachment 295712 [details]
Patch

Clearing flags on attachment: 295712

Committed r210682: <http://trac.webkit.org/changeset/210682>

[WebKit Commit Bot]

All reviewed patches have been landed.  Closing bug.