[Lldb-commits] [PATCH] D74917: [lldb/cmake] Fail configuration if a required swig dependency is not met.

Jonas Devlieghere via Phabricator via lldb-commits lldb-commits at lists.llvm.org
Thu Feb 20 16:46:28 PST 2020


JDevlieghere added a comment.

In D74917#1885615 <https://reviews.llvm.org/D74917#1885615>, @mattd wrote:

> In D74917#1885590 <https://reviews.llvm.org/D74917#1885590>, @JDevlieghere wrote:
>
> > It should never be possible for `LLDB_ENABLE_PYTHON` to be true but `SWIG_FOUND` to be false. The modules `FindPythonInterpAndLibs.cmake` and `FindLuaAndSwig.cmake` should fail early when SWIG isn't found.
>
>
> Thanks for taking a look.  Would you be okay with making the status `message()` in `FindPythonInterpAndLibs.cmake` and `FindLuaAndSwig.cmake` fatal, that way the dependency error is clearer?


That would defeat the purpose of auto-detecting these dependencies. Please take a look at D71306 <https://reviews.llvm.org/D71306> for all the details. The summary is that all optional dependencies default to `Auto`: where we enable them if we can find them. You can override this behavior by passing `LLDB_ENABLE_PYTHON=ON` to CMake, in which case not finding Python (or SWIG) will be a fatal error.

I agree that it can be confusing to figure out that Python got disabled because SWIG wasn't found. Currently we call `find_package` with `QUIET` from `FindPythonInterpAndLibs.cmake` and `FindLuaAndSwig.cmake`. I think we should remove that so that the user gets more information in CMake's configuration output. Would that address your concerns?


CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D74917/new/

https://reviews.llvm.org/D74917





More information about the lldb-commits mailing list