rebis-dev: Unexpected round brackets for characters '2' and '3'

[triska]

With the rebis-dev branch, I get:

?- C = '1'.
   C = '1'.   % expected
?- C = '2'.
   C = ('2'). % unexpected (expected: '2')
?- C = '3'.
   C = ('3'). % unexpected (expected: '3')
?- C = '4'.
   C = '4'.   % expected
?- C = '5'.
   C = '5'.   % expected

[UWN]

And I get:

?- member(C,"0123456789").
   C = '0'
;  C = '1'
;  C = '2'
;  C = '3'
;  C = '4'
;  C = '5'
;  C = '6'
;  C = '7'
;  C = ('8')
;  C = '9'
;  false.
?- end_of_file.
ulrich@p0:/opt/gupu/scryer-prolog$ git describe
v0.8.123-805-gbcacb49

[josd]

It looks non-deterministic:

$ scryer-prolog
?- use_module(library(lists)).
   true.
?- member(C,"0123456789").
   C = '0'
;  C = ('1')
;  C = '2'
;  C = '3'
;  C = ('4')
;  C = ('5')
;  C = '6'
;  C = '7'
;  C = '8'
;  C = '9'
;  false.
?-

[triska]

This works as expected now, thank you a lot!

I hope that internally, characters will simply be treated as atoms (of length 1) in the future of rebis-dev?

[mthom]

I hope that internally, characters will simply be treated as atoms (of length 1) in the future of rebis-dev?

Yes, that's one case covered by the planned short atom optimization.

[UWN]

... And that "abc" = [A|As] does not involve any lookup in an atom table?

[mthom]

Right.