diff options
author | Bruno Cardoso Lopes <bruno.cardoso@gmail.com> | 2017-02-07 21:54:57 +0000 |
---|---|---|
committer | Bruno Cardoso Lopes <bruno.cardoso@gmail.com> | 2017-02-07 21:54:57 +0000 |
commit | 17da34d2bd42e763ad9995a01a5d8784137a476c (patch) | |
tree | 0a305b1566f5a4478097027dc5b6bf0ecca09c01 /llvm/lib/Transforms/Utils/PredicateInfo.cpp | |
parent | c0d5590a3b908ce583b9e24099b5028026a93456 (diff) | |
download | llvm-17da34d2bd42e763ad9995a01a5d8784137a476c.zip llvm-17da34d2bd42e763ad9995a01a5d8784137a476c.tar.gz llvm-17da34d2bd42e763ad9995a01a5d8784137a476c.tar.bz2 |
[PCH] Fix a regression when PCH is used with -fmodules
Following up on r291465 after a regression in r276159. When we use
-fmodule-name=X while building a PCH, modular headers in X will be
textually included and the compiler knows that we are not building
module X, so don't serialize such headers in the PCH as being part of a
module, because at this point they are not.
This was causing subtle bugs and malformed AST crashes, for instance,
when using the PCH in subsequent compiler invocation with -fmodules, the
HFI for a modular header would map to the PCH, which would force a
module load of and unexistent module ID.
rdar://problem/30171164
llvm-svn: 294361
Diffstat (limited to 'llvm/lib/Transforms/Utils/PredicateInfo.cpp')
0 files changed, 0 insertions, 0 deletions