[PATCH] D8822: Proposed fix for PR23076 (conditional branch debug line info)

[Adrian Prantl via cfe-commits]

aprantl added a comment.

The linetable produced by GCC appears to be the most accurate one (stopping at the && for the short-circuit evaluation and then returning there for the actual & and the branch). I would not have a problem with clang implementing it.
The downside of it is that it breaks the flow of stepping when the cursor goes back up one line even if that is what is actually happening.

I don't think that getting rid of the short-circuit breakpoint would be a very big loss, and not returning to the operator after returning from the RHS also shouldn't break any useful workflow.

What happens if && is an overloaded C++ operator — would still be a breakpoint before the call or would it also disappear?


http://reviews.llvm.org/D8822