[PATCH] D32566: Revert rL301328 and add tests for the regressions introduced.

[Richard Smith via Phabricator via cfe-commits]

rsmith added a comment.

I'm OK with this from a mechanical perspective. But there's also a libclang design question here: what should the libclang methods to query template arguments for a type cursor representing an alias template specialization actually do? Should there be some way for a libclang user to choose what result they get?

One way we could make the behavior fully consistent (and more backwards-compatible with pre-clang-4.0) would be to revert both this and https://reviews.llvm.org/D26663, and provide the information that https://reviews.llvm.org/D26663 wanted to expose by another set of interface functions. What do other people with an interest in the libclang interface think?


Repository:
  rL LLVM

https://reviews.llvm.org/D32566