cons forms satisfy pair?

[vito]

previously, neither pair? nor list? would return true for this form:

[a & b]

so this changes pair? to also return true for Cons values (square brackets).

admittedly, the naming with list/cons/pair is awfully inconsistent between Go and Bass right now.

in Bass, a list? value must be a linked list - a List whose second value is another List. both Cons and Pair can satisfy this.

in Go, List is the interface implemented by Cons and Pair, but it doesn't ensure that the second value is a list, so it's not always a list?.

in Bass, a pair? value used to be ONLY the Pair type. now it will be true of Cons as well.

here's the old truth table:

         []  ()  [1 2] (1 2) [1 & 2] (1 & 2)
list?    1   1     1     1      0       0
pair?    0   0     0     1      0       1
empty?   1   1     0     0      0       0

here's the new truth table:

         []  ()  [1 2] (1 2) [1 & 2] (1 & 2)
list?    1   1     1     1      0       0
pair?    0   0     1     1      1       1
empty?   1   1     0     0      0       0

there is no cons? - just as there is no keyword?, but there is symbol?. there may be a good reason to add both in the future, but I'm waiting for it to arrive.

an alternative approach here might be to have Cons support decoding into Pair, but that may be the path to darkness. Value.Decode explicitly calls out that it shouldn't be used for converting between Value types, though I don't remember if that's coming from a footgun or if it's speculative.