#include "heap.h"
#include "unicode.h"
#include "inputline.h"
+#include "wstat.h"
static int tests, errors;
+static int fail_first;
+
+static void count_error() {
+ ++errors;
+ if(fail_first)
+ abort();
+}
/** @brief Checks that @p expr is nonzero */
#define insist(expr) do { \
if(!(expr)) { \
- ++errors; \
+ count_error(); \
fprintf(stderr, "%s:%d: error checking %s\n", \
__FILE__, __LINE__, #expr); \
} \
if(w == 0) { \
fprintf(stderr, "%s:%d: %s returned 0\n", \
__FILE__, __LINE__, #GOT); \
- ++errors; \
+ count_error(); \
} else if(strcmp(w, g)) { \
fprintf(stderr, "%s:%d: %s returned:\n%s\nexpected:\n%s\n", \
__FILE__, __LINE__, #GOT, format(g), format(w)); \
- ++errors; \
+ count_error(); \
} \
++tests; \
} while(0)
char *u8; \
\
insist(validutf8(CHARS)); \
- ucs = utf82ucs4(CHARS); \
+ ucs = utf8_to_utf32(CHARS, strlen(CHARS), 0); \
insist(ucs != 0); \
- insist(!ucs4cmp(w, ucs)); \
- u8 = ucs42utf8(ucs); \
+ insist(!utf32_cmp(w, ucs)); \
+ u8 = utf32_to_utf8(ucs, utf32_len(ucs), 0); \
insist(u8 != 0); \
insist(!strcmp(u8, CHARS)); \
} while(0)
fprintf(stderr, "test_utf8\n");
+#define validutf8(S) utf8_valid((S), strlen(S))
/* empty string */
break;
}
if(l) {
+ uint32_t *d;
/* Case-folded data is now normalized */
- canon_expected = ucs42utf8(utf32_decompose_canon(&l, 1, 0));
+ d = utf32_decompose_canon(&l, 1, 0);
+ canon_expected = utf32_to_utf8(d, utf32_len(d), 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(canon_folded), format(canon_expected));
- ++errors;
+ count_error();
}
++tests;
- compat_expected = ucs42utf8(utf32_decompose_compat(&l, 1, 0));
+ d = utf32_decompose_compat(&l, 1, 0);
+ compat_expected = utf32_to_utf8(d, utf32_len(d), 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;
+ count_error();
}
++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 Run breaking tests for utf32_grapheme_boundary() etc */
+static void breaktest(const char *path,
+ int (*breakfn)(const uint32_t *, size_t, size_t)) {
+ FILE *fp = open_unicode_test(path);
+ int lineno = 0;
+ char *l, *lp;
+ size_t bn, n;
+ char break_allowed[1024];
+ uint32_t buffer[1024];
+
+ while(!inputline(path, 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, "%s:%d: evil line: %s", path, lineno, l);
+ }
+ for(n = 0; n <= bn; ++n) {
+ if(breakfn(buffer, bn, n) != break_allowed[n]) {
+ fprintf(stderr,
+ "%s:%d: offset %zu: mismatch\n",
+ path, lineno, n);
+ count_error();
+ }
+ ++tests;
+ }
+ xfree(l);
+ }
+ fclose(fp);
+}
+
/** @brief Tests for @ref lib/unicode.h */
static void test_unicode(void) {
FILE *fp;
int cn, bn;
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, " c%d: %s\n", \
+ 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", \
#T, B, format_utf32(T##_c[B])); \
- ++errors; \
+ count_error(); \
} \
} while(0)
unt_check(NFD, 3, 1);
}
xfree(l);
}
+ fclose(fp);
+ breaktest("auxiliary/GraphemeBreakTest.txt", utf32_is_grapheme_boundary);
+ breaktest("auxiliary/WordBreakTest.txt", utf32_is_word_boundary);
}
int main(void) {
+ fail_first = !!getenv("FAIL_FIRST");
insist('\n' == 0x0A);
insist('\r' == 0x0D);
insist(' ' == 0x20);