#include "heap.h"
#include "unicode.h"
#include "inputline.h"
+#include "wstat.h"
static int tests, errors;
static void test_casefold(void) {
uint32_t c, l;
- const char *input, *folded, *expected;
+ const char *input, *canon_folded, *compat_folded, *canon_expected, *compat_expected;
fprintf(stderr, "test_casefold\n");
+
+ /* This isn't a very exhaustive test. Unlike for normalization, there don't
+ * seem to be any public test vectors for these algorithms. */
for(c = 1; c < 256; ++c) {
input = utf32_to_utf8(&c, 1, 0);
- folded = utf8_casefold_canon(input, strlen(input), 0);
+ canon_folded = utf8_casefold_canon(input, strlen(input), 0);
+ compat_folded = utf8_casefold_compat(input, strlen(input), 0);
switch(c) {
default:
if((c >= 'A' && c <= 'Z')
l = 0x3BC; /* GREEK SMALL LETTER MU */
break;
case 0xDF: /* LATIN SMALL LETTER SHARP S */
- insist(!strcmp(folded, "ss"));
+ insist(!strcmp(canon_folded, "ss"));
+ insist(!strcmp(compat_folded, "ss"));
l = 0;
break;
}
if(l) {
/* Case-folded data is now normalized */
- expected = ucs42utf8(utf32_decompose_canon(&l, 1, 0));
- if(strcmp(folded, expected)) {
- fprintf(stderr, "%s:%d: casefolding %#lx got '%s', expected '%s'\n",
+ canon_expected = ucs42utf8(utf32_decompose_canon(&l, 1, 0));
+ if(strcmp(canon_folded, canon_expected)) {
+ fprintf(stderr, "%s:%d: canon-casefolding %#lx got '%s', expected '%s'\n",
__FILE__, __LINE__, (unsigned long)c,
- format(folded), format(expected));
+ format(canon_folded), format(canon_expected));
+ ++errors;
+ }
+ ++tests;
+ compat_expected = ucs42utf8(utf32_decompose_compat(&l, 1, 0));
+ if(strcmp(compat_folded, compat_expected)) {
+ fprintf(stderr, "%s:%d: compat-casefolding %#lx got '%s', expected '%s'\n",
+ __FILE__, __LINE__, (unsigned long)c,
+ format(compat_folded), format(compat_expected));
++errors;
}
++tests;
putchar('\n');
}
+/** @brief Open a Unicode test file */
+static FILE *open_unicode_test(const char *path) {
+ const char *base;
+ FILE *fp;
+ char buffer[1024];
+ int w;
+
+ if((base = strrchr(path, '/')))
+ ++base;
+ else
+ base = path;
+ if(!(fp = fopen(base, "r"))) {
+ snprintf(buffer, sizeof buffer,
+ "wget http://www.unicode.org/Public/5.0.0/ucd/%s", path);
+ if((w = system(buffer)))
+ fatal(0, "%s: %s", buffer, wstat(w));
+ if(chmod(base, 0444) < 0)
+ fatal(errno, "chmod %s", base);
+ if(!(fp = fopen(base, "r")))
+ fatal(errno, "%s", base);
+ }
+ return fp;
+}
+
/** @brief Tests for @ref lib/unicode.h */
static void test_unicode(void) {
FILE *fp;
uint32_t buffer[1024];
uint32_t *c[6], *NFD_c[6], *NFKD_c[6]; /* 1-indexed */
int cn, bn;
+ char break_allowed[1024];
fprintf(stderr, "test_unicode\n");
- if(!(fp = fopen("NormalizationTest.txt", "r"))) {
- system("wget http://www.unicode.org/Public/5.0.0/ucd/NormalizationTest.txt");
- chmod("NormalizationTest.txt", 0444);
- if(!(fp = fopen("NormalizationTest.txt", "r"))) {
- perror("NormalizationTest.txt");
- ++tests; /* don't know how many... */
- ++errors;
- return;
- }
- }
+ fp = open_unicode_test("NormalizationTest.txt");
while(!inputline("NormalizationTest.txt", fp, &l, '\n')) {
++lineno;
if(*l == '#' || *l == '@')
#define unt_check(T, A, B) do { \
++tests; \
if(utf32_cmp(c[A], T##_c[B])) { \
- fprintf(stderr, "L%d: c%d != "#T"(c%d)\n", lineno, A, B); \
+ fprintf(stderr, \
+ "NormalizationTest.txt:%d: c%d != "#T"(c%d)\n", \
+ lineno, A, B); \
fprintf(stderr, " c%d: %s\n", \
A, format_utf32(c[A])); \
fprintf(stderr, "%4s(c%d): %s\n", \
}
xfree(l);
}
+ fclose(fp);
+ fp = open_unicode_test("auxiliary/GraphemeBreakTest.txt");
+ lineno = 0;
+ while(!inputline("GraphemeBreakTest.txt.txt", fp, &l, '\n')) {
+ ++lineno;
+ if(l[0] == '#') continue;
+ bn = 0;
+ lp = l;
+ while(*lp) {
+ if(*lp == ' ' || *lp == '\t') {
+ ++lp;
+ continue;
+ }
+ if(*lp == '#')
+ break;
+ if((unsigned char)*lp == 0xC3 && (unsigned char)lp[1] == 0xB7) {
+ /* 00F7 DIVISION SIGN */
+ break_allowed[bn] = 1;
+ lp += 2;
+ continue;
+ }
+ if((unsigned char)*lp == 0xC3 && (unsigned char)lp[1] == 0x97) {
+ /* 00D7 MULTIPLICATION SIGN */
+ break_allowed[bn] = 0;
+ lp += 2;
+ continue;
+ }
+ if(isxdigit((unsigned char)*lp)) {
+ buffer[bn++] = strtoul(lp, &lp, 16);
+ continue;
+ }
+ fatal(0, "GraphemeBreakTest.txt:%d: evil line: %s", lineno, l);
+ }
+ for(cn = 0; cn <= bn; ++cn) {
+ if(utf32_is_gcb(buffer, bn, cn) != break_allowed[cn]) {
+ fprintf(stderr,
+ "GraphemeBreakTest.txt:%d: offset %d: utf32_is_gcb wrong\n",
+ lineno, cn);
+ ++errors;
+ }
+ ++tests;
+ }
+ xfree(l);
+ }
}
int main(void) {
Local Variables:
c-basic-offset:2
comment-column:40
+fill-column:79
+indent-tabs-mode:nil
End:
*/
-