Likely simple bugs to fix in the format/2-3 predicates

[pmoura]

Bug in the implementation of ~c (code): the following calls are expected to print an A:

| ?- format("~c", [65]).

yes
| ?- format('~c', [65]).

yes

Bug in the implementation of ~i (ignore): the following calls are expected to omit the middle argument and thus print ac:

| ?- format("~a~i~a", [a,b,c]).
ab

yes
| ?- format('~a~i~a', [a,b,c]).
ab

yes

[spabreu]

Hi Paulo,

Bug in the implementation of ~c (code): the following calls are expected to print an A: | ?- format("~c", [65]).

yes | ?- format('~c', [65]).

yes

In this case, I don’t think there's a bug:

| ?- format("~c", [65]), flush_output. A

yes

cheers, -spa

[pmoura]

Hi Salvador.

Interesting. But other Prolog systems don't require a call to flush_output/0 (or flush_output/1 for format/3 calls).

[pmoura]

Additionally, the test definition implicitly closes the output stream, which is expected to flush any pending output.

[pmoura]

Fixes confirmed. Thanks!