`na_matches = "never"` doesn't work if matching columns contain only NAs

[moodymudskipper]

dplyr::inner_join(data.frame(a=NA), data.frame(a=NA), by = "a", na_matches = "never")
#> Error in `dplyr::inner_join()`:
#> ! Can't convert `out[merge]$a` <logical> to match type of `a` <vctrs_unspecified>.

^{Created on 2023-03-22 with reprex v2.0.2}

[zajichek]

Any conversation on making na_matches = "never" the default argument? I wouldn't have expected that NA values would be joined by default, but maybe that's just me. Anyway, just wondering.