[cfe-commits] r152609 - in /cfe/trunk/lib: AST/DeclBase.cpp Sema/SemaLookup.cpp Serialization/ASTWriter.cpp

[Chandler Carruth]

On Mon, Mar 12, 2012 at 9:58 PM, Nick Lewycky <nlewycky at google.com> wrote:

> Like what? There is absolutely no user-visible behaviour change (or at
> least, there shouldn't be!)


You changed some places that used to do a lookup when the shouldn't, to not
do such a lookup. While that change may not have been observable, you also
added an assert to the path which should not have been taken. Had you not
cleaned up these other places, the assert would have started firing. I'm
looking to ensure we have test coverage for those places which before you
cleanup would have hit the assert you added. That way we can be sure these
places never regress and start using the bad code path again.

It's a small thing, yes, but I worry about regressing on these things. We
spent a lot of time tracking down bad things, and adding an assert is only
part of the process.

That said, it's entirely possible we *already* have test coverage for these
paths. In which case, I'm happy. Just wanted to make sure.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/cfe-commits/attachments/20120314/5dd4b81b/attachment.html>