zturner added a comment. In http://reviews.llvm.org/D14790#292493, @krytarowski wrote:
> Not long ago we rejected dynamic dependency on `curses`(3), as a fall-back > for NetBSD (without `libpanel`(3) in the 7.0 release). Why doing the similar > thing with swig is fine? > > Why are we abstracting swig when there is no other option viable? > > Is possible to improve this approach of finding swig? We are looking for it > in CMake and later in Python? > `find_package(SWIG)`, `--find-swig`. > > Maybe just add a single switch in CMake to enable or disable it without > options. > > Please explain what does static bindings mean here. static binding = one person runs swig to generate `LLDBWrapPython.cpp` and `lldb.py`, then they check these files into the upstream. Nobody else has to run swig. Dynamic binding aka on-the-fly-binding = only the swig interface file is checked in. There is a build step that automatically runs swig locally on every developer's machine when the interface file changes and automatically generates `LLDBWrapPython.cpp` and `lldb.py` at build time. http://reviews.llvm.org/D14790 _______________________________________________ lldb-commits mailing list lldb-commits@lists.llvm.org http://lists.llvm.org/cgi-bin/mailman/listinfo/lldb-commits
