[PATCH] Fix WinEHPrepare bug with multiple catch handlers

[Reid Kleckner]

I understand why we need to split the block after llvm.eh.endcatch, but I don't really see why we should be treating icmps specially. Maybe we should pre-populate the value map with all dominating EH value stores instead? I'm looking at the broken test case and looking at how to fix it.

Feel free to commit the basic block splitting separately.


REPOSITORY
  rL LLVM

http://reviews.llvm.org/D8682

EMAIL PREFERENCES
  http://reviews.llvm.org/settings/panel/emailpreferences/