mirror of
https://github.com/Mercury-Language/mercury.git
synced 2026-04-17 10:23:46 +00:00
ee001d38b4
library/string.m:
Add predicate to test if a string is in UTF-8 or UTF-16,
depending on the target language.
NEWS:
Announce the addition.
tests/hard_coded/string_well_formed.exp:
tests/hard_coded/string_well_formed.m:
Add basic test case.
tests/hard_coded/string_well_formed_utf8.exp:
tests/hard_coded/string_well_formed_utf8.exp2:
tests/hard_coded/string_well_formed_utf8.exp3:
tests/hard_coded/string_well_formed_utf8.inp:
tests/hard_coded/string_well_formed_utf8.m:
Add more thorough test for UTF-8. The input file is from
https://www.cl.cam.ac.uk/~mgk25/ucs/examples/UTF-8-test.txt
tests/hard_coded/Mmakefile:
Enable the tests.
44 lines
1.3 KiB
Mathematica
44 lines
1.3 KiB
Mathematica
%---------------------------------------------------------------------------%
|
|
% vim: ts=4 sw=4 et ft=mercury
|
|
%---------------------------------------------------------------------------%
|
|
|
|
:- module string_well_formed.
|
|
:- interface.
|
|
|
|
:- import_module io.
|
|
|
|
:- pred main(io::di, io::uo) is det.
|
|
|
|
%---------------------------------------------------------------------------%
|
|
%---------------------------------------------------------------------------%
|
|
|
|
:- implementation.
|
|
|
|
:- import_module int.
|
|
:- import_module string.
|
|
|
|
%---------------------------------------------------------------------------%
|
|
|
|
main(!IO) :-
|
|
Empty = "",
|
|
Good = "\U0001F600", % 4 UTF-8 code units, 2 UTF-16 code units
|
|
Bad1 = string.between(Good, 0, 1),
|
|
Bad2 = string.between(Good, 0, length(Good) - 1),
|
|
Bad3 = Bad1 ++ Good,
|
|
Bad4 = Good ++ Bad2,
|
|
test_well_formed("Empty", Empty, !IO),
|
|
test_well_formed("Good", Good, !IO),
|
|
test_well_formed("Bad1", Bad1, !IO),
|
|
test_well_formed("Bad2", Bad2, !IO),
|
|
test_well_formed("Bad3", Bad3, !IO),
|
|
test_well_formed("Bad4", Bad4, !IO).
|
|
|
|
:- pred test_well_formed(string::in, string::in, io::di, io::uo) is det.
|
|
|
|
test_well_formed(Label, S, !IO) :-
|
|
( if string.is_well_formed(S) then
|
|
io.write_string(Label ++ " is well-formed\n", !IO)
|
|
else
|
|
io.write_string(Label ++ " is not well-formed\n", !IO)
|
|
).
|