Add codepoint data and Unicode version

Signed-off-by: Mattias Andrée <m@maandree.se>

1 files changed, 139 insertions, 0 deletions

diff --git a/UnicodeData.parse.c b/UnicodeData.parse.c
new file mode 100644
index 0000000..173fb6a
--- /dev/null
+++ b/UnicodeData.parse.c
@@ -0,0 +1,139 @@
+/* See LICENSE file for copyright and license details. */
+#include "parse-common.c"
+
+
+struct segment {
+	uint32_t first, last;
+};
+
+
+static struct segment *segments = NULL;
+static size_t nsegments = 0u;
+static int ret = 0;
+
+
+static char *
+split(char **text)
+{
+	char *field = *text;
+	if (field) {
+		*text = strchr(*text, ';');
+		if (*text)
+			*(*text)++ = '\0';
+	}
+	return field;
+}
+
+
+static int
+output_character_data(char *text)
+{
+	char *field;
+
+	if (!(field = split(&text))) return 0; /* codepoint */
+
+	if (!(field = split(&text))) return 0;
+	ret |= printf(".name = \"%s\"", field);
+
+	if (!(field = split(&text))) return 0;
+	ret |= printf(", .general_category = GENERAL_CATEGORY_%s", field);
+
+	if (!(field = split(&text))) return 0;
+	ret |= printf(", .canonical_combining_classes = %su", field);
+
+	if (!(field = split(&text))) return 0;
+	ret |= printf(", .bidirectional_category = BIDIRECTIONAL_CATEGORY_%s", field);
+
+	if (!(field = split(&text))) return 0; /* TODO Character Decomposition Mapping */
+	if (!(field = split(&text))) return 0; /* TODO Decimal digit value */
+	if (!(field = split(&text))) return 0; /* TODO Digit value */
+	if (!(field = split(&text))) return 0; /* TODO Numeric value */
+
+	if (!(field = split(&text))) return 0;
+	ret |= printf(", .bidirectional_mirrored = BIDIRECTIONAL_MIRRORED_%s", field);
+
+	if (!(field = split(&text))) return 0;
+	ret |= printf(", .unicode_1_name = \"%s\"", field);
+
+	if (!(field = split(&text))) return 0; /* ISO 10646 comment field -- always empty as of Unicode 5.2.0 */
+
+	if (!(field = split(&text))) return 0; /* TODO Uppercase Mapping */
+	if (!(field = split(&text))) return 0; /* TODO Lowercase Mapping */
+	if (!(field = split(&text))) return 0; /* TODO Titlecase Mapping */
+
+	return 1;
+}
+
+
+static void
+parse_line(char *text, size_t lineno)
+{
+	unsigned long int codepoint;
+	char *end;
+
+	if (!isxdigit(*text)) {
+	malformat:
+		fprintf(stderr, "%s: line %zu in is malformatted\n", argv0, lineno);
+		exit(1);
+	}
+
+	errno = 0;
+
+	codepoint = strtoul(text, &end, 16);
+	if (errno || codepoint > ULTIMATE_CODEPOINT || *end != ';')
+		goto malformat;
+
+	if (!nsegments) {
+	new_segment:
+		ret |= printf("static const struct libcmap_codepoint_data data_from_%04lx[] = {\n", codepoint);
+		segments = realloc(segments, ++nsegments * sizeof(*segments));
+		if (!segments) {
+			fprintf(stderr, "%s: realloc %zu: %s\n", argv0, nsegments * sizeof(*segments), strerror(errno));
+			exit(1);
+		}
+		segments[nsegments - 1u].first = (uint32_t)codepoint;
+		goto new_codepoint;
+	} else if (segments[nsegments - 1u].last + 1u == (uint32_t)codepoint) {
+		ret |= printf(",\n");
+	new_codepoint:
+		segments[nsegments - 1u].last = (uint32_t)codepoint;
+		ret |= printf("\t{");
+		if (!output_character_data(text))
+			goto malformat;
+		ret |= printf("}");
+	} else {
+		if ((uint32_t)codepoint < segments[nsegments - 1u].last) {
+			fprintf(stderr, "%s: UnicodeData is not sorted\n", argv0);
+			exit(1);
+		}
+		ret |= printf("\n};\n");
+		goto new_segment;
+	}
+}
+
+
+static int
+output(void)
+{
+	size_t i;
+
+	if (!nsegments) {
+		fprintf(stderr, "%s: character list is empty\n", argv0);
+		exit(1);
+	}
+
+	ret |= printf("\n};\n");
+
+	ret |= printf("static const struct libcmap_codepoint_segment list[] = {\n");
+	for (i = 0u; i < nsegments; i++) {
+		if (i)
+			ret |= printf(",\n");
+		ret |= printf("\t{.range = {0x%04"PRIX32", 0x%04"PRIX32"}, .data = data_from_%04"PRIx32"}",
+		              segments[i].first, segments[i].last, segments[i].first);
+	}
+	ret |= printf("\n};\n");
+
+	free(segments);
+
+	return ret;
+}