[cfe-dev] RFC: Easier AST Matching by Default

[Stephen Kelly via cfe-dev]

On 27/05/2020 11:42, Stephen Kelly wrote:
> I have a proof of concept patch which adds a virtual method to 
> ClangTidyCheck which returns an overriding optional<TraversalKind> to 
> use for the check. If you reimplement that method in your downstream 
> checks to return AsIs then you don't have to change individual 
> matchers yet.
>
> What do you think?
>
> I have the patch on a computer I will have access to in about 2 hours.


I have uploaded that patch here: https://reviews.llvm.org/D80623

It is WIP as it does not include a test. Let me know if you think that 
is an interesting direction and we can finish the patch.


> Sorry for the disruption. At least the issue is getting visibility 
> here and feedback already indicates that the release notes could be 
> more informative.


I have added a detailed porting guide to the release notes. Hopefully by 
the time anyone reads this email they will be updated here:

  https://clang.llvm.org/docs/ReleaseNotes.html#ast-matchers

I will be away now for a few hours. I'll look more into the bug 
discovered via the transform library when I return.

Thanks,

Stephen.


-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/cfe-dev/attachments/20200527/14037326/attachment-0001.html>