Displaying 3 results from an estimated 3 matches for "r201072".
2014 Feb 10
2
[LLVMdev] [llvm] r201072 - [CMake] Introduce llvm_add_library().
NAKAMURA Takumi wrote:
> [CMake] Introduce llvm_add_library().
I recommend moving away from wrappers like this. They indicate that either
CMake is not providing the interfaces needed, or not propagating them, or
that they exist but are not used.
Such wrappers don't parse the arguments in the same way as the wrapped
command etc. Wrappers are not good API proxies. Additionally, you put
2014 Feb 13
2
[LLVMdev] [llvm] r201072 - [CMake] Introduce llvm_add_library().
...>
>> CMake provides CMAKE_SHARED_MODULE_SUFFIX and CMAKE_IMPORT_LIBRARY_SUFFIX.
>> Can you set those at directory scope instead of setting the target property
>> in the macro?
>
> Reasonable. They could be put into common configurator, HandleLLVMOptions.
> For now in r201072, I simply gather common logics around there.
> I'd be happy if llvm_add_library became simpler, and I will work on.
>
>> Anyway, just a note to consider moving instead in a direction of fewer
>> wrappers, rather than more. I'm not familiar enough with the llvm
>> bu...
2014 Feb 13
2
[LLVMdev] [llvm] r201072 - [CMake] Introduce llvm_add_library().
...rovides CMAKE_SHARED_MODULE_SUFFIX and CMAKE_IMPORT_LIBRARY_SUFFIX.
>> Can you set those at directory scope instead of setting the target property
>> in the macro?
>>
>>
>> Reasonable. They could be put into common configurator, HandleLLVMOptions.
>> For now in r201072, I simply gather common logics around there.
>> I'd be happy if llvm_add_library became simpler, and I will work on.
>>
>> Anyway, just a note to consider moving instead in a direction of fewer
>> wrappers, rather than more. I'm not familiar enough with the llvm
>...