Scryer-prolog: atomic/1 sometimes incorrectly fails

[triska]

For instance:

?- atomic(6).
   false.

Per 8.3.5.4, this should succeed.

Quoting from the standard:

8.3.5 atomic/1

8.3.5.1 Description

atomic(X) is true if X is a member of the set A or I or
F (7.1.4, 7.1.2, 7.1.3) and is false if X is a member of
the set V or CT (7.1.1, 7.1.5).

[srenatus]

_💭Is there an ISO test suite of some sort?_

[triska]

See for example this publication:

Szabó P., Szeredi P. (2006) Improving the ISO Prolog Standard by Analyzing Compliance Test Results. In: Etalle S., Truszczyński M. (eds) Logic Programming. ICLP 2006. Lecture Notes in Computer Science, vol 4079. Springer, Berlin, Heidelberg

[pmoura]

https://github.com/LogtalkDotOrg/logtalk3/tree/master/tests/prolog

Unfortunately, Scryer Prolog is still missing standard functionality (e.g. stream I/O) that would allow Logtalk and thus its tests suites (~6000 tests) to run automatically using it (e.g. as part of the of a CI pipeline).

[srenatus]

@triska @pmoura thanks for the links and insights. Indeed,

run automatically using it (e.g. as part of the of a CI pipeline)

sounds like what we might want to strive for! I wasn't aware of the gaps. What's missing beyond stream IO, would you know, @pmoura?

[pmoura]

@srenatus Logtalk backend Prolog compiler requirements are described at:

https://logtalk.org/backend_requirements.html

Related to CI, when using GitHub, there's also:

https://github.com/logtalk-actions

Note that a setup action for Scryer Prolog is also available (may need to be updated and may also be possible to improve it for the current status of Scryer Prolog development):

https://github.com/logtalk-actions/setup-scryer-prolog

[srenatus]

Sorry for derailing; I think the original issue (atomic(6)) has been resolved. ✔️

[triska]

Yes, thank you a lot!