This is a highly unstable theoretical fork of SWI-Prolog

(>95% complete)

• undo/1 Hook
• delimited continuations (working/passing tests)
• attvars may override any predicate
• sicstus:attribute/1
• sicstus:put_atts/2
• sicstus:get_atts/2
• xsb:verify_attributes/2
• xsb:attv_unify/2
• listing/1 shows variables

(>40% complete)

• eclipse:meta_attributes/1 (copy_term/2, ==/2, =@=/2, etc)
• sicstus:verify_attributes/3 (90%)
• dra-tabling interpretor (works .. needs C storage) https://github.com/logicmoo/swipl-devel/blob/ATT_LOGICMOO/library/tabling/dra_interp.pl#L5692-L6101

[11979 steps, 1001 new answers tabled (1001 in all)]
Fib of 1000 is 70330367711422815821835254877183549770181269836358732742604905087154537118196933579742249494562611733487750449241765991088186363265450223647106012053374121273867339111198139373125598767690091902245245323403501
% 6,729,826 inferences, 4.405 CPU in 4.410 seconds (100% CPU, 1527685 Lips)

Tries based ssytme in C that is efficant as copy_term?

if i want to store p(q(r,A,B,t),A) = foo. i create a few levels deep of HTs .. p/2 -> q/4 -> r -> V1 -> V2 -> t -> V1 = foo

• delimited continuations tabling (60%)

work in progress

• This library is described in the paper "Tabling as Library with Delimited Control"
• by Benoit Desouter, Marko van Dooren and Tom Schrijvers. (Email: Benoit dot Desouter at UGent dot be)
• http://www-ps.informatik.uni-kiel.de/kdpd2013/talks/schrijvers.pdf
• http://arxiv.org/pdf/1507.08087v1.pdf

(0% complete)

• $schedule_wakeup/1 to not create a choicepoint block
• call/1, apply/2 will not create a choicepoint block (settable)

?- call(member(X,[1,2])),apply(call,[!]).
1
No

• catch/3 will not create a choicepoint block (settable)
• All predicates treated as transparent (settable)
• callers context is not changed
• new search order: (settable)
• definition source module context
• caller context

Expected incompatbilities to main SWI-Prolog

• meta_predicate/1 sets module_transparent/1 in all cases
• More issues: https://github.com/logicmoo/fluentvars/issues