[cfe-commits] PATCH: Better handle overloaded functions in typo correction

[Chandler Carruth]

On Thu, Jul 28, 2011 at 2:25 PM, Kaelyn Uhrain <rikka at google.com> wrote:

> Here's an updated version of my patch to add function overload resolution
> to the typo correction, eliminating one of the ways that Sema::CorrectTypo
> returns a non-keyword correction without a corresponding NamedDecl. It
> incorporates all of the changes and feedback from
> http://codereview.appspot.com/4747047. Given it's been almost two weeks
> since I sent out the first version and over a week since I made changes
> based on Chandler's last feedback (I was hoping he'd get a chance to look at
> those changes, but he's been swamped), I'd appreciate it of someone could
> give it a once-over and let me know whether it is ready for submission.
>

I've made a few additional comments, but my primary concern with this patch
is the representation of KeywordDecl. I think playing fast and loose with
magical pointer values is going to come back to bite us. If we can't use
'NULL' as the magical pointer value, we should find some more robust way to
represent a keyword in the correction results.

I'd certainly appreciate ideas from dgregor and any others on the list on
the best way to represent that.

To make it easier for others to review, can you attach the latest patch to
the email thread?
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/cfe-commits/attachments/20110801/2fb45308/attachment.html>