[webkit-dev] Problems with the new NATIVE_TYPE_ERR

Joshua Bell jsbell at chromium.org
Tue Nov 13 10:44:14 PST 2012


On Tue, Nov 13, 2012 at 9:38 AM, Darin Adler <darin at apple.com> wrote:

> Four months back we added NATIVE_TYPE_ERR for IndexedDB, as a way for the
> DOM code to raise JavaScript type error exceptions rather than DOM
> exceptions
> <http://trac.webkit.org/changeset/123112>. Yesterday we started using
> NATIVE_TYPE_ERR elsewhere in the DOM <
> http://trac.webkit.org/changeset/134345>.
>
> Looking this over, I see some problems. I added comments to
> https://bugs.webkit.org/show_bug.cgi?id=91679 and
> https://bugs.webkit.org/show_bug.cgi?id=101604 describing the problems.
>
> The new error code is not handled at all in bindings other than the
> JavaScript bindings. I also think the constant for it is not well named.
>
> Using NATIVE_TYPE_ERR in its current state outside IndexedDB causes an
> immediate showstopper bug for Apple’s Mac version of WebKit and perhaps for
> others as well. We may want to roll back this recent change until we deal
> with that.
>

Rolling http://trac.webkit.org/changeset/134345 out in webkit.org/b/102106

Renaming NATIVE_TYPE_ERR when we have settled on a name (discussion in
webkit.org/b/91679

The switch from DOMException of type TYPE_MISMATCH_ERR to TypeError does
not appear to be controversial, but as arv@ mentioned in one of the bugs:

"WebIDL only defines bindings for ECMAScript. We really need someone from
Apple to take care of the ObjC bindings, someone from QT for CPP bindings
and someone for GTK to take care of the GTK bindings."

And by "take care of" we mean "define what a WebIDL TypeError even means in
binding XYZ". So - help appreciated.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.webkit.org/pipermail/webkit-dev/attachments/20121113/be97d062/attachment.html>


More information about the webkit-dev mailing list