diff options
Diffstat (limited to 'text/test-text.c')
| -rw-r--r-- | text/test-text.c | 139 |
1 files changed, 139 insertions, 0 deletions
diff --git a/text/test-text.c b/text/test-text.c new file mode 100644 index 0000000..e8077b5 --- /dev/null +++ b/text/test-text.c @@ -0,0 +1,139 @@ +/* Vega's text against Unicode's conformance tests (docs/design/004-text.md): + * test-text UCD-DIRECTORY + * BidiTest.txt (Bidi_Class sequences, each paragraph direction its bitset + * names) and BidiCharacterTest.txt (code points, paired brackets included): + * every line's levels and visual order. Prints one JSON line of counts; + * exits 1 on the first failures (up to ten shown). */ +#define _POSIX_C_SOURCE 200809L +#include "text.h" +#include <stdio.h> +#include <stdlib.h> +#include <string.h> + +enum { MAX = 512 }; +static unsigned failures; + +static void fail(const char *file, size_t line, const char *what) { + if (failures++ < 10) fprintf(stderr, "%s:%zu: %s\n", file, line, what); +} + +static const char *const names[VEGA_BIDI_CLASSES] = { + "L", "R", "AL", "EN", "ES", "ET", "AN", "CS", "NSM", "BN", "B", "S", "WS", "ON", + "LRE", "LRO", "RLE", "RLO", "PDF", "LRI", "RLI", "FSI", "PDI"}; + +static int class_named(const char *name) { + for (int c = 0; c < VEGA_BIDI_CLASSES; ++c) + if (!strcmp(names[c], name)) return c; + return -1; +} + +/* Levels as the tests write them ("x" for removed), into levels; how many. */ +static size_t parse_levels(char *text, uint8_t *levels) { + size_t n = 0; + for (char *token = strtok(text, " \t\r\n"); token && n < MAX; token = strtok(NULL, " \t\r\n")) + levels[n++] = token[0] == 'x' ? VEGA_BIDI_REMOVED : (uint8_t)atoi(token); + return n; +} + +static size_t parse_order(char *text, uint32_t *order) { + size_t n = 0; + for (char *token = strtok(text, " \t\r\n"); token && n < MAX; token = strtok(NULL, " \t\r\n")) + order[n++] = (uint32_t)strtoul(token, NULL, 10); + return n; +} + +static bool same_result(const uint8_t *levels, const uint8_t *want, size_t count, const uint32_t *order, + size_t order_count) { + if (memcmp(levels, want, count)) return false; + uint32_t got[MAX]; + return vega_bidi_order(levels, count, got) == order_count && !memcmp(got, order, order_count * sizeof(*got)); +} + +static size_t bidi_test(const char *path) { + FILE *in = fopen(path, "r"); + if (!in) { perror(path); exit(2); } + char line[4096]; + uint8_t want[MAX], levels[MAX]; + uint32_t order[MAX]; + size_t want_count = 0, order_count = 0, cases = 0, number = 0; + while (fgets(line, sizeof(line), in)) { + ++number; + if (line[0] == '#' || line[0] == '\n' || line[0] == '\r') continue; + if (!strncmp(line, "@Levels:", 8)) { want_count = parse_levels(line + 8, want); continue; } + if (!strncmp(line, "@Reorder:", 9)) { order_count = parse_order(line + 9, order); continue; } + if (line[0] == '@') continue; + char *semicolon = strchr(line, ';'); + if (!semicolon) continue; + *semicolon = 0; + unsigned bitset = (unsigned)strtoul(semicolon + 1, NULL, 16); + VegaBidiClass classes[MAX]; + size_t count = 0; + for (char *token = strtok(line, " \t"); token && count < MAX; token = strtok(NULL, " \t")) { + int c = class_named(token); + if (c < 0) { fail(path, number, "unknown class"); break; } + classes[count++] = (VegaBidiClass)c; + } + if (count != want_count) { fail(path, number, "levels of another length"); continue; } + static const int directions[3] = {VEGA_BIDI_AUTO, VEGA_BIDI_LTR, VEGA_BIDI_RTL}; + for (unsigned d = 0; d < 3; ++d) { + if (!(bitset >> d & 1)) continue; + ++cases; + if (vega_bidi_levels(classes, NULL, count, directions[d], levels) < 0) { fail(path, number, "no memory"); continue; } + if (!same_result(levels, want, count, order, order_count)) { + char what[64]; + snprintf(what, sizeof(what), "direction %u differs", d); + fail(path, number, what); + } + } + } + fclose(in); + return cases; +} + +static size_t bidi_character_test(const char *path) { + FILE *in = fopen(path, "r"); + if (!in) { perror(path); exit(2); } + char line[8192]; + size_t cases = 0, number = 0; + while (fgets(line, sizeof(line), in)) { + ++number; + if (line[0] == '#' || line[0] == '\n' || line[0] == '\r') continue; + char *field[5]; + field[0] = line; + for (unsigned f = 1; f < 5; ++f) { + field[f] = strchr(field[f - 1], ';'); + if (!field[f]) break; + *field[f]++ = 0; + } + uint32_t codes[MAX], order[MAX]; + VegaBidiClass classes[MAX]; + uint8_t want[MAX], levels[MAX]; + size_t count = 0; + for (char *token = strtok(field[0], " \t"); token && count < MAX; token = strtok(NULL, " \t")) { + codes[count] = (uint32_t)strtoul(token, NULL, 16); + classes[count] = vega_bidi_class(codes[count]); + ++count; + } + int direction = atoi(field[1]), paragraph = atoi(field[2]); + size_t want_count = parse_levels(field[3], want), order_count = parse_order(field[4], order); + ++cases; + int got = vega_bidi_levels(classes, codes, count, direction == 0 ? VEGA_BIDI_LTR : direction == 1 ? VEGA_BIDI_RTL : VEGA_BIDI_AUTO, + levels); + if (want_count != count) fail(path, number, "levels of another length"); + else if (got != paragraph) fail(path, number, "paragraph level differs"); + else if (!same_result(levels, want, count, order, order_count)) fail(path, number, "levels or order differ"); + } + fclose(in); + return cases; +} + +int main(int argc, char **argv) { + const char *ucd = argc > 1 ? argv[1] : "/usr/share/unicode"; + char path[1024]; + snprintf(path, sizeof(path), "%s/BidiTest.txt", ucd); + size_t bidi = bidi_test(path); + snprintf(path, sizeof(path), "%s/BidiCharacterTest.txt", ucd); + size_t characters = bidi_character_test(path); + printf("{\"bidi_test_cases\":%zu,\"bidi_character_test_cases\":%zu,\"failures\":%u}\n", bidi, characters, failures); + return failures ? 1 : 0; +} |