Clean up dependencies

[h-vetinari]

No description provided.

[conda-forge-webservices]

Hi! This is the friendly automated conda-forge-linting service.

I just wanted to let you know that I linted all conda-recipes in your PR (recipe) and found it was in an excellent condition.

[isuruf]

Can we please keep make? ninja's verbose mode gives very little info and recent llvmdev-feedstock issue would have been easier to debug if we used make.

[h-vetinari]

Can we please keep make? ninja's verbose mode gives very little info and recent llvmdev-feedstock issue would have been easier to debug if we used make.

OK, if you have that preference, I don't mind.

(by default, I try to harmonize the build system across platforms with ninja where possible; one less divergence to handle).

[h-vetinari]

Hm, seems the NATIVE_FLAGS here aren't being picked up correctly, or rather, the lack of ZSTD_ROOT / LIBXML2_ROOT in https://github.com/llvm/llvm-project/blob/llvmorg-16.0.0-rc4/llvm/cmake/modules/LLVMConfig.cmake.in#L70-L84 leads to the one from the host prefix being found:

-- Detecting CXX compile features - done
-- Found ZLIB: $BUILD_PREFIX/lib/libz.so (found version "1.2.13")  
-- Found zstd: $PREFIX/lib/libzstd.so  
-- Found LibXml2: $PREFIX/lib/libxml2.so (found version "2.10.3") 

[h-vetinari]

Thanks for the help with 367a295 Isuru, much better than what I was trying to do in conda-forge/llvmdev-feedstock#204, obviously.

I've looked a bit more closely how the native config is being built in upstream cross-compilation, and it's basically a normal CMake invocation; this also explains why CMAKE_PREFIX_PATH works out-of-the-box.

In any case, this should now be more consistent, both internally and with our llvmdev builds.

There shouldn't be any surprises here, but let me know if you want some time to review at your leisure.

[h-vetinari]

Included in #206