Bug 157491 – Crash beneath ObjCCallbackFunctionImpl::call

• Status: RESOLVED
• Resolution: FIXED
• Component: JavaScriptCore

[Michael Saboff]

It appears that some Promise builtin code is throwing an exception.  The exception is still set in the VM when we later call tryUnwrapObjcObject() from ObjCCallbackFunctionImpl::call().

[Michael Saboff]

<rdar://problem/25867374>

[Michael Saboff]

Created attachment 278469 [details]
Patch

[Saam Barati]

Comment on attachment 278469 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=278469&action=review

> Source/JavaScriptCore/ChangeLog:11
> +        Tried creating a test case, but I don't have source for the app.
> +        I can't seem to find the right combination of Promises and ObjC code.

What about just regular JS code?
Like you have two schedules promises.
The first throws
The second should still run. Would we start running the second with an exception in the VM?

[Michael Saboff]

(In reply to comment #3)
> Comment on attachment 278469 [details]
> Patch
> 
> View in context:
> https://bugs.webkit.org/attachment.cgi?id=278469&action=review
> 
> > Source/JavaScriptCore/ChangeLog:11
> > +        Tried creating a test case, but I don't have source for the app.
> > +        I can't seem to find the right combination of Promises and ObjC code.
> 
> What about just regular JS code?
> Like you have two schedules promises.
> The first throws
> The second should still run. Would we start running the second with an
> exception in the VM?

The bug here is when we try to invoke a callback that is an instance function on a wrapped ObjC object.

[WebKit Commit Bot]

Comment on attachment 278469 [details]
Patch

Clearing flags on attachment: 278469

Committed r200610: <http://trac.webkit.org/changeset/200610>

[WebKit Commit Bot]

All reviewed patches have been landed.  Closing bug.