Mercurial > minori

annotate dep/utf8proc/test/case.c @ 364:99c961c91809

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
core: refactor out byte stream into its own file easy dubs
author Paper <paper@paper.us.eu.org>
date Tue, 16 Jul 2024 21:15:59 -0400
parents 1faa72660932
children
Ignore whitespace changes - Everywhere: Within whitespace: At end of lines:
rev   line source
343
1faa72660932 *: transfer back to cmake from autotools
Paper <paper@paper.us.eu.org>
parents:
diff changeset
1 #include "tests.h"
1faa72660932 *: transfer back to cmake from autotools
Paper <paper@paper.us.eu.org>
parents:
diff changeset
2 #include <wctype.h>
1faa72660932 *: transfer back to cmake from autotools
Paper <paper@paper.us.eu.org>
parents:
diff changeset
3
1faa72660932 *: transfer back to cmake from autotools
Paper <paper@paper.us.eu.org>
parents:
diff changeset
4 int main(int argc, char **argv)
1faa72660932 *: transfer back to cmake from autotools
Paper <paper@paper.us.eu.org>
parents:
diff changeset
5 {
1faa72660932 *: transfer back to cmake from autotools
Paper <paper@paper.us.eu.org>
parents:
diff changeset
6 int error = 0, better = 0;
1faa72660932 *: transfer back to cmake from autotools
Paper <paper@paper.us.eu.org>
parents:
diff changeset
7 utf8proc_int32_t c;
1faa72660932 *: transfer back to cmake from autotools
Paper <paper@paper.us.eu.org>
parents:
diff changeset
8
1faa72660932 *: transfer back to cmake from autotools
Paper <paper@paper.us.eu.org>
parents:
diff changeset
9 (void) argc; /* unused */
1faa72660932 *: transfer back to cmake from autotools
Paper <paper@paper.us.eu.org>
parents:
diff changeset
10 (void) argv; /* unused */
1faa72660932 *: transfer back to cmake from autotools
Paper <paper@paper.us.eu.org>
parents:
diff changeset
11
1faa72660932 *: transfer back to cmake from autotools
Paper <paper@paper.us.eu.org>
parents:
diff changeset
12 /* some simple sanity tests of the character widths */
1faa72660932 *: transfer back to cmake from autotools
Paper <paper@paper.us.eu.org>
parents:
diff changeset
13 for (c = 0; c <= 0x110000; ++c) {
1faa72660932 *: transfer back to cmake from autotools
Paper <paper@paper.us.eu.org>
parents:
diff changeset
14 utf8proc_int32_t l = utf8proc_tolower(c);
1faa72660932 *: transfer back to cmake from autotools
Paper <paper@paper.us.eu.org>
parents:
diff changeset
15 utf8proc_int32_t u = utf8proc_toupper(c);
1faa72660932 *: transfer back to cmake from autotools
Paper <paper@paper.us.eu.org>
parents:
diff changeset
16 utf8proc_int32_t t = utf8proc_totitle(c);
1faa72660932 *: transfer back to cmake from autotools
Paper <paper@paper.us.eu.org>
parents:
diff changeset
17
1faa72660932 *: transfer back to cmake from autotools
Paper <paper@paper.us.eu.org>
parents:
diff changeset
18 check(l == c || utf8proc_codepoint_valid(l), "invalid tolower");
1faa72660932 *: transfer back to cmake from autotools
Paper <paper@paper.us.eu.org>
parents:
diff changeset
19 check(u == c || utf8proc_codepoint_valid(u), "invalid toupper");
1faa72660932 *: transfer back to cmake from autotools
Paper <paper@paper.us.eu.org>
parents:
diff changeset
20 check(t == c || utf8proc_codepoint_valid(t), "invalid totitle");
1faa72660932 *: transfer back to cmake from autotools
Paper <paper@paper.us.eu.org>
parents:
diff changeset
21
1faa72660932 *: transfer back to cmake from autotools
Paper <paper@paper.us.eu.org>
parents:
diff changeset
22 if (utf8proc_codepoint_valid(c) && (l == u) != (l == t) &&
1faa72660932 *: transfer back to cmake from autotools
Paper <paper@paper.us.eu.org>
parents:
diff changeset
23 /* Unicode 11: Georgian Mkhedruli chars have uppercase but no titlecase. */
1faa72660932 *: transfer back to cmake from autotools
Paper <paper@paper.us.eu.org>
parents:
diff changeset
24 !(((c >= 0x10d0 && c <= 0x10fa) || c >= (0x10fd && c <= 0x10ff)) && l != u)) {
1faa72660932 *: transfer back to cmake from autotools
Paper <paper@paper.us.eu.org>
parents:
diff changeset
25 fprintf(stderr, "unexpected titlecase %x for lowercase %x / uppercase %x\n", t, l, c);
1faa72660932 *: transfer back to cmake from autotools
Paper <paper@paper.us.eu.org>
parents:
diff changeset
26 ++error;
1faa72660932 *: transfer back to cmake from autotools
Paper <paper@paper.us.eu.org>
parents:
diff changeset
27 }
1faa72660932 *: transfer back to cmake from autotools
Paper <paper@paper.us.eu.org>
parents:
diff changeset
28
1faa72660932 *: transfer back to cmake from autotools
Paper <paper@paper.us.eu.org>
parents:
diff changeset
29 if (sizeof(wint_t) > 2 || (c < (1<<16) && u < (1<<16) && l < (1<<16))) {
1faa72660932 *: transfer back to cmake from autotools
Paper <paper@paper.us.eu.org>
parents:
diff changeset
30 wint_t l0 = towlower((wint_t)c), u0 = towupper((wint_t)c);
1faa72660932 *: transfer back to cmake from autotools
Paper <paper@paper.us.eu.org>
parents:
diff changeset
31
1faa72660932 *: transfer back to cmake from autotools
Paper <paper@paper.us.eu.org>
parents:
diff changeset
32 /* OS unicode tables may be out of date. But if they
1faa72660932 *: transfer back to cmake from autotools
Paper <paper@paper.us.eu.org>
parents:
diff changeset
33 do have a lower/uppercase mapping, hopefully it
1faa72660932 *: transfer back to cmake from autotools
Paper <paper@paper.us.eu.org>
parents:
diff changeset
34 is correct? */
1faa72660932 *: transfer back to cmake from autotools
Paper <paper@paper.us.eu.org>
parents:
diff changeset
35 if (l0 != (wint_t)c && l0 != (wint_t)l) {
1faa72660932 *: transfer back to cmake from autotools
Paper <paper@paper.us.eu.org>
parents:
diff changeset
36 fprintf(stderr, "MISMATCH %x != towlower(%x) == %x\n",
1faa72660932 *: transfer back to cmake from autotools
Paper <paper@paper.us.eu.org>
parents:
diff changeset
37 l, c, l0);
1faa72660932 *: transfer back to cmake from autotools
Paper <paper@paper.us.eu.org>
parents:
diff changeset
38 ++error;
1faa72660932 *: transfer back to cmake from autotools
Paper <paper@paper.us.eu.org>
parents:
diff changeset
39 }
1faa72660932 *: transfer back to cmake from autotools
Paper <paper@paper.us.eu.org>
parents:
diff changeset
40 else if (l0 != (wint_t)l) { /* often true for out-of-date OS unicode */
1faa72660932 *: transfer back to cmake from autotools
Paper <paper@paper.us.eu.org>
parents:
diff changeset
41 ++better;
1faa72660932 *: transfer back to cmake from autotools
Paper <paper@paper.us.eu.org>
parents:
diff changeset
42 /* printf("%x != towlower(%x) == %x\n", l, c, l0); */
1faa72660932 *: transfer back to cmake from autotools
Paper <paper@paper.us.eu.org>
parents:
diff changeset
43 }
1faa72660932 *: transfer back to cmake from autotools
Paper <paper@paper.us.eu.org>
parents:
diff changeset
44 if (u0 != (wint_t)c && u0 != (wint_t)u) {
1faa72660932 *: transfer back to cmake from autotools
Paper <paper@paper.us.eu.org>
parents:
diff changeset
45 fprintf(stderr, "MISMATCH %x != towupper(%x) == %x\n",
1faa72660932 *: transfer back to cmake from autotools
Paper <paper@paper.us.eu.org>
parents:
diff changeset
46 u, c, u0);
1faa72660932 *: transfer back to cmake from autotools
Paper <paper@paper.us.eu.org>
parents:
diff changeset
47 ++error;
1faa72660932 *: transfer back to cmake from autotools
Paper <paper@paper.us.eu.org>
parents:
diff changeset
48 }
1faa72660932 *: transfer back to cmake from autotools
Paper <paper@paper.us.eu.org>
parents:
diff changeset
49 else if (u0 != (wint_t)u) { /* often true for out-of-date OS unicode */
1faa72660932 *: transfer back to cmake from autotools
Paper <paper@paper.us.eu.org>
parents:
diff changeset
50 ++better;
1faa72660932 *: transfer back to cmake from autotools
Paper <paper@paper.us.eu.org>
parents:
diff changeset
51 /* printf("%x != towupper(%x) == %x\n", u, c, u0); */
1faa72660932 *: transfer back to cmake from autotools
Paper <paper@paper.us.eu.org>
parents:
diff changeset
52 }
1faa72660932 *: transfer back to cmake from autotools
Paper <paper@paper.us.eu.org>
parents:
diff changeset
53 }
1faa72660932 *: transfer back to cmake from autotools
Paper <paper@paper.us.eu.org>
parents:
diff changeset
54 }
1faa72660932 *: transfer back to cmake from autotools
Paper <paper@paper.us.eu.org>
parents:
diff changeset
55 check(!error, "utf8proc case conversion FAILED %d tests.", error);
1faa72660932 *: transfer back to cmake from autotools
Paper <paper@paper.us.eu.org>
parents:
diff changeset
56
1faa72660932 *: transfer back to cmake from autotools
Paper <paper@paper.us.eu.org>
parents:
diff changeset
57 /* issue #130 */
1faa72660932 *: transfer back to cmake from autotools
Paper <paper@paper.us.eu.org>
parents:
diff changeset
58 check(utf8proc_toupper(0x00df) == 0x1e9e &&
1faa72660932 *: transfer back to cmake from autotools
Paper <paper@paper.us.eu.org>
parents:
diff changeset
59 utf8proc_totitle(0x00df) == 0x1e9e &&
1faa72660932 *: transfer back to cmake from autotools
Paper <paper@paper.us.eu.org>
parents:
diff changeset
60 utf8proc_tolower(0x00df) == 0x00df &&
1faa72660932 *: transfer back to cmake from autotools
Paper <paper@paper.us.eu.org>
parents:
diff changeset
61 utf8proc_tolower(0x1e9e) == 0x00df &&
1faa72660932 *: transfer back to cmake from autotools
Paper <paper@paper.us.eu.org>
parents:
diff changeset
62 utf8proc_toupper(0x1e9e) == 0x1e9e,
1faa72660932 *: transfer back to cmake from autotools
Paper <paper@paper.us.eu.org>
parents:
diff changeset
63 "incorrect 0x00df/0x1e9e case conversions");
1faa72660932 *: transfer back to cmake from autotools
Paper <paper@paper.us.eu.org>
parents:
diff changeset
64 utf8proc_uint8_t str_00df[] = {0xc3, 0x9f, 0x00};
1faa72660932 *: transfer back to cmake from autotools
Paper <paper@paper.us.eu.org>
parents:
diff changeset
65 utf8proc_uint8_t str_1e9e[] = {0xe1, 0xba, 0x9e, 0x00};
1faa72660932 *: transfer back to cmake from autotools
Paper <paper@paper.us.eu.org>
parents:
diff changeset
66 utf8proc_uint8_t *s1 = utf8proc_NFKC_Casefold(str_00df);
1faa72660932 *: transfer back to cmake from autotools
Paper <paper@paper.us.eu.org>
parents:
diff changeset
67 utf8proc_uint8_t *s2 = utf8proc_NFKC_Casefold(str_1e9e);
1faa72660932 *: transfer back to cmake from autotools
Paper <paper@paper.us.eu.org>
parents:
diff changeset
68 check(!strcmp((char*)s1, "ss") &&
1faa72660932 *: transfer back to cmake from autotools
Paper <paper@paper.us.eu.org>
parents:
diff changeset
69 !strcmp((char*)s2, "ss"),
1faa72660932 *: transfer back to cmake from autotools
Paper <paper@paper.us.eu.org>
parents:
diff changeset
70 "incorrect 0x00df/0x1e9e casefold normalization");
1faa72660932 *: transfer back to cmake from autotools
Paper <paper@paper.us.eu.org>
parents:
diff changeset
71 free(s1);
1faa72660932 *: transfer back to cmake from autotools
Paper <paper@paper.us.eu.org>
parents:
diff changeset
72 free(s2);
1faa72660932 *: transfer back to cmake from autotools
Paper <paper@paper.us.eu.org>
parents:
diff changeset
73 printf("More up-to-date than OS unicode tables for %d tests.\n", better);
1faa72660932 *: transfer back to cmake from autotools
Paper <paper@paper.us.eu.org>
parents:
diff changeset
74 printf("utf8proc case conversion tests SUCCEEDED.\n");
1faa72660932 *: transfer back to cmake from autotools
Paper <paper@paper.us.eu.org>
parents:
diff changeset
75 return 0;
1faa72660932 *: transfer back to cmake from autotools
Paper <paper@paper.us.eu.org>
parents:
diff changeset
76 }