[webkit-reviews] review canceled: [Bug 40060] Add a version of JSObjectSetPrototype that yields an exception when the prototype can't be set : [Attachment 59259] Fix v1.1

bugzilla-daemon at webkit.org bugzilla-daemon at webkit.org
Tue Jun 22 03:46:10 PDT 2010


Jędrzej Nowacki <jedrzej.nowacki at nokia.com> has canceled Jędrzej Nowacki
<jedrzej.nowacki at nokia.com>'s request for review:
Bug 40060: Add a version of JSObjectSetPrototype that yields an exception when
the prototype can't be set
https://bugs.webkit.org/show_bug.cgi?id=40060

Attachment 59259: Fix v1.1
https://bugs.webkit.org/attachment.cgi?id=59259&action=review

------- Additional Comments from Jędrzej Nowacki <jedrzej.nowacki at nokia.com>
(In reply to comment #8)
> JSObjectSetPrototypeWithException would probably be better. The function
doesn't actually handle errors -- it just supports an exception out parameter.
Thanks, changed.

> > This function can stay in a private header for now, for example next
version of JSC API and then it could replace the original one.
> I don't think it's practical to replace the old function, since that would
break binary compatibility.
You mean that we won't be able to break the binary compatibility promise,
never?

> > Is API's performance a good justification to introduce the function?
> It can be. It depends on whether this is a performance-critical path. Is it?
If so, why?
As the QtScript is an API, I have a similar problem, like the JSC C API, to
really define a good use cases. Lets try a logical experiment. Could you give
me a performance-critical path for setting JSObjectSetProperty? If yes then it
is the same as mine :-) if not can I put in the implementation a sleep() call?
I try to make implementation of the QtScript clean and fast as possible. As you
pointed me, there are a few ways to get the behavior, but I believe that, they
are not so clean and fast as they could be. If you could create the
JSObjectSetPrototype function again, assuming that you would knew about the
exception, would you create it with an exception parameter or without? I think,
it should be fixed, but I don't want to push to much as I can live without it
(2x time longer as it is as much slower :-) ).


More information about the webkit-reviews mailing list