I haven’t had time to look at this in detail yet, but when I originally had this idea I thought we would use lit’s discovery mechanism to find all .py files, and then invoke them using dotest.py in single process mode with a path to a specific file.
Why do we need run lines? On Wed, Apr 4, 2018 at 5:04 AM Pavel Labath via Phabricator < revi...@reviews.llvm.org> wrote: > labath added a comment. > > In https://reviews.llvm.org/D45215#1056731, @JDevlieghere wrote: > > > Alright, I'm convinced this is the way to go. > > > > - For (1) I'll see if I can get some inspiration from the visit logic in > dotest.py. I guess the functionality is similar. I agree on doing this in a > separate tool, especially if we want to remove functionality from dotest in > the long run. > > > I'm not aware of any code which would do the full visit. We have some code > in dosep.py (`find_test_files_in_dir_tree`), but this one only does file > traversal. It does not look inside the the files for classes, test methods > and such. > > The real visiting happens inside unittest2 (loader.py). The code there > seems very complicated, but I think that's because it tries to support > various things we don't use, so /I hope/ that our visiting logic can be > simpler. > > OTOH, looking at loader.py, it seems to me it may actually be possible to > let it do the enumeration for us, if we stroke it the right way. > > > https://reviews.llvm.org/D45215 > > > >
_______________________________________________ lldb-commits mailing list lldb-commits@lists.llvm.org http://lists.llvm.org/cgi-bin/mailman/listinfo/lldb-commits
