Internal: slightly inaccurate `as_join_by()` implementation for character about dplyr HOT 8 OPEN

We do allow this though:

unclass(dplyr::join_by("a" == "a"))
#> $exprs
#> $exprs[[1]]
#> "a" == "a"
#> 
#> 
#> $condition
#> [1] "=="
#> 
#> $filter
#> [1] "none"
#> 
#> $x
#> [1] "a"
#> 
#> $y
#> [1] "a"

But maybe that should be the "special" case, and coming from character should try to generate the "typical" case of a == a

from dplyr.

What are the semantics of dplyr::join_by("a" == "a") ? Shouldn't the result be identical to dplyr::join_by(a == a) ?

from dplyr.