jhuber6 wrote: > TBH, I was/am confused by that suggestion. I don't see really how it relates > to a python-language or something wanting to speed up their MLIR build?
Users have always had the flexibility to selectively build dependencies, e.g. `ninja install-clang` will not build random unrelated LLVM tools but `ninja install` will. The onus is on the end user to make use of these if they need a smaller dependency tree, components are just an LLVM approved way to 'build your own distribution'. If a python language wanted to speed up their MLIR build, then MLIR just needs to provide sufficient targets for the end user. https://github.com/llvm/llvm-project/pull/221724 _______________________________________________ cfe-commits mailing list [email protected] https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
