[PATCH] D31444: LTO: call getRealLinkageName on IRNames before feeding to getGUID

[Bob Haarman via Phabricator via llvm-commits]

inglorion added a comment.

Since refactoring getGUID is going to touch a fair few more places than this diff does, is it OK to do that in a follow-up? I'm still struggling to reduce the test case for this change from "ninja FileCheck" to something smaller, but I really don't want to hold it up any longer than necessary. I have a few more changes that will make Clang build Clang with ThinLTO on Windows, and I'm excited to get them all committed.


https://reviews.llvm.org/D31444