From: Bruno Cardoso Lopes <bruno.cardoso@gmail.com>
Date: Wed, 22 Mar 2017 00:11:21 +0000 (+0000)
Subject: [Modules] Rebuild modules on umbrella header mismatch
X-Git-Url: https://granicus.if.org/sourcecode?a=commitdiff_plain;h=625d61d6c5d39ebcde60a01d8955c8701734dade;p=clang

[Modules] Rebuild modules on umbrella header mismatch

This restores behavior pre-r230064 since after PCMCache work (r298278)
we don't reload PCMs from disk within the same compiler invocation.

Testcases from r230064 are still left around since they still guarantee
the correct behavior we're expecting.

rdar://problem/19889777

git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@298464 91177308-0d34-0410-b5e6-96231b3b80d8
---

diff --git a/lib/Serialization/ASTReader.cpp b/lib/Serialization/ASTReader.cpp
index d9033c830a..c6b21aadbd 100644
--- a/lib/Serialization/ASTReader.cpp
+++ b/lib/Serialization/ASTReader.cpp
@@ -4890,13 +4890,9 @@ ASTReader::ReadSubmoduleBlock(ModuleFile &F, unsigned ClientLoadCapabilities) {
         if (!CurrentModule->getUmbrellaHeader())
           ModMap.setUmbrellaHeader(CurrentModule, Umbrella, Blob);
         else if (CurrentModule->getUmbrellaHeader().Entry != Umbrella) {
-          // This can be a spurious difference caused by changing the VFS to
-          // point to a different copy of the file, and it is too late to
-          // to rebuild safely.
-          // FIXME: If we wrote the virtual paths instead of the 'real' paths,
-          // after input file validation only real problems would remain and we
-          // could just error. For now, assume it's okay.
-          break;
+          if ((ClientLoadCapabilities & ARR_OutOfDate) == 0)
+            Error("mismatched umbrella headers in submodule");
+          return OutOfDate;
         }
       }
       break;