On Thu, Jan 31, 2013 at 9:46 AM, David Blaikie <[email protected]> wrote:
> +Chandler, CMake owner > > Not sure if we need to do the flag support check (ala > -Wcovered-switch-default) when adding -Wmissing-field-initializers (in > case we're dealing with non-Clang non-GCC compilers (that are still > GCC compatible?)?) > Yea, i think at least the else branch needs the test. Personally, I would sink both sides under the test just for clarity. > > On Thu, Jan 31, 2013 at 6:51 AM, Edwin Vane <[email protected]> wrote: > > Hi gribozavr, dblaikie, > > > > gcc produces false positives for empty braces so turning the warning > > off. Instead, turning the warning on for clang so proper warnings aren't > > missed. > > > > http://llvm-reviews.chandlerc.com/D358 > > > > Files: > > cmake/modules/HandleLLVMOptions.cmake > > > > Index: cmake/modules/HandleLLVMOptions.cmake > > =================================================================== > > --- cmake/modules/HandleLLVMOptions.cmake > > +++ cmake/modules/HandleLLVMOptions.cmake > > @@ -178,6 +178,14 @@ > > elseif( LLVM_COMPILER_IS_GCC_COMPATIBLE ) > > if (LLVM_ENABLE_WARNINGS) > > add_llvm_definitions( -Wall -W -Wno-unused-parameter > -Wwrite-strings ) > > + # Turn off missing field initializer warnings for gcc to avoid > noise from > > + # false positives with empty {}. Turn them on otherwise (they're > off by > > + # default for clang). > > + if (CMAKE_COMPILER_IS_GNUCXX) > > + add_llvm_definitions( -Wno-missing-field-initializers ) > > + else() > > + add_llvm_definitions( -Wmissing-field-initializers ) > > + endif() > > if (LLVM_ENABLE_PEDANTIC) > > add_llvm_definitions( -pedantic -Wno-long-long ) > > endif (LLVM_ENABLE_PEDANTIC) >
_______________________________________________ cfe-commits mailing list [email protected] http://lists.cs.uiuc.edu/mailman/listinfo/cfe-commits
