url-mode: fix double-width characters not being handled correctly

When a regex matches a string containing double-width characters, the
CELL_SPACER values were included in the URL string. This meant the
final URL (either launched, or copied) weren't handled correctly, as
invalid UTF-8 sequences were inserted in the middle of the string.

Closes #2027

CHANGELOG.md

@@ -106,9 +106,12 @@
 * Build failure (`srgb.h` not found) when doing a parallel build.
 * Regression: reflowing (changing the window size) removing empty
   lines ([#2011][2011]).
+* `url/regex-copy` not handling double-width characters correctly
+  ([#2027][2027]).
 
 [2000]: https://codeberg.org/dnkl/foot/issues/2000
 [2011]: https://codeberg.org/dnkl/foot/issues/2011
+[2027]: https://codeberg.org/dnkl/foot/issues/2027
 
 
 ### Security

url-mode.c

@@ -347,6 +347,9 @@ regex_detected(const struct terminal *term, enum url_action action,
                 wc_count = composed->count;
             }
 
+            else if (wc[0] >= CELL_SPACER)
+                continue;
+
             /* Convert wide character to utf8 */
             for (size_t i = 0; i < wc_count; i++) {
                 char buf[16];
@@ -355,6 +358,7 @@ regex_detected(const struct terminal *term, enum url_action action,
                 if (char_len == (size_t)-1)
                     continue;
 
+
                 for (size_t j = 0; j < char_len; j++) {
                     const size_t requires_size = vline->len + char_len;
 
@@ -411,9 +415,9 @@ regex_detected(const struct terminal *term, enum url_action action,
             const size_t end = start + mlen;
 
             LOG_DBG(
-                "regex match at row %d: %.*srow/col = %dx%d",
+                "regex match at row %d: %.*s (%zu bytes), row/col = %dx%d",
                 matches[1].rm_so, (int)mlen, &search_string[matches[1].rm_so],
-                v->map[start].row, v->map[start].col);
+                mlen, v->map[start].row, v->map[start].col);
 
             tll_push_back(
                 *urls,