Closed brevzin closed 2 years ago
Thanks @brevzin. Can you look into the test regressions?
Thanks @brevzin. Can you look into the test regressions?
Yeah, looks like this was cause of some implementations not having the std::map
deduction guide (that we technically have a feature test macro for, that nobody actually defines, oh well). You worked around this one clever way with test_zip_to_map
, I just muscled through making a std::map
-equivalent. This should work 🤞.
Okay now everything passes, except the windows ones where it just looks like the CI setup is broken.
The current overload set is inconsistent:
m3
fails because the logic here tries to instantiatestd::map<std::pair<int, int>>
(specifically,Cont<range_value_t<Rng>>
) instead of actually using the existing deduction machinery (which is what them4
overload correctly uses).