arphaman added a comment.

Thanks for working on this! Please upload the patch with the full context (git 
diff -U99999). It helps the reviewers :)

In, @alexshap wrote:

> I'm kind of interested in this fixit, but one thought which i have - probably 
> it should be more conservative (i.e. fix captures by reference, integral 
> types, etc) (since the code might rely on side-effects of 
> copy-ctors/move-ctors or extend the lifetime of an object), but fixing only 
> simple cases still seems to be useful imo. CC: @aaron.ballman , @arphaman, 
> @ahatanak

Are you talking about a more conservative warning or a more conservative fixit? 
If it doesn't make sense for us to have a fixit for a particular capture, does 
it make sense for us to have a warning for that capture in the first place?

It would be helpful to add some tests with macros to ensure that the logic for 
how the removal range is computed can handle macros. (E.g. macro that expands 
to a full/partial capture, lambda in a macro).

  rC Clang

cfe-commits mailing list

Reply via email to