Rtools42 is an MinGW distribution for package development in R. Compilation of swipl-devel is successful, and so is ?- check_installation.
, but I get an error in ctest for pcre, related to missing re_config(jittarget(X))
if there is no PCRE JIT compilation, that is, re_config(jit(false))
.
Unsure what to do. Currently, re_config(jittarget(X))
fails, although it might also raise an error. In test_pcre.pl, it seems that there may have been an error message in a previous version. I could also think of a scenario in which jittarget(X) returns something like X = “JIT is not enabled”, or just “NA”.
Maybe someting like this (test_pcre.pl, line 350).
re_test(config_all3, DocSorted = ConfigsSorted) :-
Doc = [bsr2(_),
compiled_widths(_),
depthlimit(_),
heaplimit(_),
jit(_),
jittarget(_),
linksize(_),
matchlimit(_),
never_backslash_c(_),
newline2(_),
parenslimit(_),
stackrecurse(_),
unicode(_),
unicode_version(_),
version(_)
],
sort(Doc, DocSorted),
bagof(C, re_config_for_doc(C), Configs), %%%%%%%%%%%%% here
assertion(ground(Configs)),
msort(Configs, ConfigsSorted), % note msort/2, in case of dups
% For more easily finding the first mismatch:
maplist(assertion_eq, DocSorted, ConfigsSorted).
re_config_for_doc(jittarget(none)) :-
re_config(jit(false)).
re_config_for_doc(X) :-
re_config(X).
And a similar thing below (added the condition):
re_test(config_jittarget, [condition(re_config(jit(true)))]) :-
re_config(jittarget(V)),
must_be(atom, V).
unicode and unicode_version may have the same issue.