On Tue, Jul 7, 2020 at 8:37 AM Rhodri James <rhodri@kynesim.co.uk> wrote:
I'm not quite convinced about making "_" non-binding, mostly because of
the knock-on effects in the PEP for other types of patterns.  It seems
to breed more special cases, and I can't help but feel that's a bad sign.

After spending some time with it in my head and in the actual testbed, it seems like a reasonable compromise. case Point(x,x): being illegal while x,x=(1,2) is perfectly legal is a bit of contention, but grammar also doesn't need to be perfectly consistent in order to learn to speak a language, it just needs to minimize the context-switching. It'd be so simple if we could use case *:, but alas.