On Sat, Feb 4, 2017 at 11:12 AM, Adam Wolk <adam.w...@tintagel.pl> wrote:
> Not that I know of but the diff removes a hardcoded -O2 passed to ghc.
>
> The port in general hardcodes -O2 in a couple of places:
> CMakeLists.txt
> 89:add_flag_append(CMAKE_C_FLAGS_RELEASE "-O2")
> 92:add_flag_append(CMAKE_CXX_FLAGS_RELEASE "-O2")
> 113:list(APPEND haskell_flags ${ghflags_parsed} "-O2")
>
> cmake_modules/CMakePascalInformation.cmake
> 14:set(CMAKE_Pascal_FLAGS_RELEASE_INIT "-O2 -dNDEBUG")
> 15:set(CMAKE_Pascal_FLAGS_RELWITHDEBINFO_INIT "-O2 -g -gl -gp")
>
> cmake_modules/CMakePascalInformation.cmake.orig
> 14:set(CMAKE_Pascal_FLAGS_RELEASE_INIT "-O2 -dNDEBUG")
> 15:set(CMAKE_Pascal_FLAGS_RELWITHDEBINFO_INIT "-O2 -g -gl -gp")
>
> should we patch out the -O2 optimization regardless if bulks break?
Some time ago I ran a full sweep of the hardcoded optimization flags
wrt cmake-based ports.
I might have missed these or they have been readded later... I don't
know, but I'm in favor of removing them.
Ciao!
David
