diff options
author | Richard Smith <richard-llvm@metafoo.co.uk> | 2018-07-27 17:13:18 +0000 |
---|---|---|
committer | Richard Smith <richard-llvm@metafoo.co.uk> | 2018-07-27 17:13:18 +0000 |
commit | 7ed5fb2d22942648408376457dcc11912066ea6f (patch) | |
tree | a2cdf5599fab6ae29b5027dc16711b28e6cde415 /clang/lib/Sema/SemaInit.cpp | |
parent | 04ccfda0750b17773808d0e7c237dc618468f147 (diff) | |
download | bcm5719-llvm-7ed5fb2d22942648408376457dcc11912066ea6f.tar.gz bcm5719-llvm-7ed5fb2d22942648408376457dcc11912066ea6f.zip |
Add missing temporary materialization conversion on left-hand side of .
in some member function calls.
Specifically, when calling a conversion function, we would fail to
create the AST node representing materialization of the class object.
llvm-svn: 338135
Diffstat (limited to 'clang/lib/Sema/SemaInit.cpp')
-rw-r--r-- | clang/lib/Sema/SemaInit.cpp | 10 |
1 files changed, 0 insertions, 10 deletions
diff --git a/clang/lib/Sema/SemaInit.cpp b/clang/lib/Sema/SemaInit.cpp index 0dddbdc36fd..3ee5ec4a492 100644 --- a/clang/lib/Sema/SemaInit.cpp +++ b/clang/lib/Sema/SemaInit.cpp @@ -7406,16 +7406,6 @@ InitializationSequence::Perform(Sema &S, if (S.DiagnoseUseOfDecl(FoundFn, Kind.getLocation())) return ExprError(); - // FIXME: Should we move this initialization into a separate - // derived-to-base conversion? I believe the answer is "no", because - // we don't want to turn off access control here for c-style casts. - CurInit = S.PerformObjectArgumentInitialization(CurInit.get(), - /*Qualifier=*/nullptr, - FoundFn, Conversion); - if (CurInit.isInvalid()) - return ExprError(); - - // Build the actual call to the conversion function. CurInit = S.BuildCXXMemberCallExpr(CurInit.get(), FoundFn, Conversion, HadMultipleCandidates); if (CurInit.isInvalid()) |