text: move regex related functions to separate file

Eventually this should probably be rewritten to use an iternal
regex engine, currently it has unacceptable memory usage, it
copies the whole text.

1 files changed, 71 insertions, 0 deletions

diff --git a/text-regex.c b/text-regex.c
new file mode 100644
index 0000000..cff8587
--- /dev/null
+++ b/text-regex.c
@@ -0,0 +1,71 @@
+#include <stdlib.h>
+#include <regex.h>
+
+#include "text-regex.h"
+
+struct Regex {
+	const char *string;
+	regex_t regex;
+};
+
+Regex *text_regex_new(void) {
+	Regex *r = calloc(1, sizeof(Regex));
+	if (!r)
+		return NULL;
+	regcomp(&r->regex, "\0\0", 0); /* this should not match anything */
+	return r;
+}
+
+int text_regex_compile(Regex *regex, const char *string, int cflags) {
+	regex->string = string;
+	int r = regcomp(&regex->regex, string, cflags);
+	if (r)
+		regcomp(&regex->regex, "\0\0", 0);
+	return r;
+}
+
+void text_regex_free(Regex *r) {
+	if (!r)
+		return;
+	regfree(&r->regex);
+	free(r);
+}
+
+int text_search_range_forward(Text *txt, size_t pos, size_t len, Regex *r, size_t nmatch, RegexMatch pmatch[], int eflags) {
+	char *buf = malloc(len + 1);
+	if (!buf)
+		return REG_NOMATCH;
+	len = text_bytes_get(txt, pos, len, buf);
+	buf[len] = '\0';
+	regmatch_t match[nmatch];
+	int ret = regexec(&r->regex, buf, nmatch, match, eflags);
+	if (!ret) {
+		for (size_t i = 0; i < nmatch; i++) {
+			pmatch[i].start = match[i].rm_so == -1 ? EPOS : pos + match[i].rm_so;
+			pmatch[i].end = match[i].rm_eo == -1 ? EPOS : pos + match[i].rm_eo;
+		}
+	}
+	free(buf);
+	return ret;
+}
+
+int text_search_range_backward(Text *txt, size_t pos, size_t len, Regex *r, size_t nmatch, RegexMatch pmatch[], int eflags) {
+	char *buf = malloc(len + 1);
+	if (!buf)
+		return REG_NOMATCH;
+	len = text_bytes_get(txt, pos, len, buf);
+	buf[len] = '\0';
+	regmatch_t match[nmatch];
+	char *cur = buf;
+	int ret = REG_NOMATCH;
+	while (!regexec(&r->regex, cur, nmatch, match, eflags)) {
+		ret = 0;
+		for (size_t i = 0; i < nmatch; i++) {
+			pmatch[i].start = match[i].rm_so == -1 ? EPOS : pos + (size_t)(cur - buf) + match[i].rm_so;
+			pmatch[i].end = match[i].rm_eo == -1 ? EPOS : pos + (size_t)(cur - buf) + match[i].rm_eo;
+		}
+		cur += match[0].rm_eo;
+	}
+	free(buf);
+	return ret;
+}