chars: create a dedicated function for getting the length of a character

Instead of calling in twenty places parse_mbchar(pointer, NULL, NULL),
use a simpler and faster char_length(pointer).  This saves pushing two
unneeded parameters onto the stack, avoids two needless ifs, and elides
an intermediate variable.

Its main purpose will follow in a later commit: to speed up searching.

src/chars.c

@@ -280,6 +280,18 @@ char *make_mbchar(long chr, int *chr_mb_len)
 	return chr_mb;
 }
 
+/* Return the length (in bytes) of the character located at *pointer. */
+int char_length(const char *pointer)
+{
+	/* If possibly a multibyte character, get its length; otherwise, it's 1. */
+	if ((signed char)*pointer < 0) {
+		int length = mblen(pointer, MAXCHARLEN);
+
+		return (length > 0 ? length : 1);
+	} else
+		return 1;
+}
+
 /* Parse a multibyte character from buf.  Return the number of bytes
  * used.  If chr isn't NULL, store the multibyte character in it.  If
  * col isn't NULL, add the character's width (in columns) to it. */
@@ -355,7 +367,7 @@ size_t move_mbleft(const char *buf, size_t pos)
 		/* Move forward again until we reach the original character,
 		 * so we know the length of its preceding character. */
 		while (before < pos) {
-			charlen = parse_mbchar(buf + before, NULL, NULL);
+			charlen = char_length(buf + before);
 			before += charlen;
 		}
 
@@ -369,7 +381,7 @@ size_t move_mbleft(const char *buf, size_t pos)
  * after the one at pos. */
 size_t move_mbright(const char *buf, size_t pos)
 {
-	return pos + parse_mbchar(buf + pos, NULL, NULL);
+	return pos + char_length(buf + pos);
 }
 
 /* This function is equivalent to strcasecmp() for multibyte strings. */

src/cut.c

@@ -35,8 +35,8 @@ void do_deletion(undo_type action)
 
 	/* When in the middle of a line, delete the current character. */
 	if (openfile->current->data[openfile->current_x] != '\0') {
-		int charlen = parse_mbchar(openfile->current->data +
+		int charlen = char_length(openfile->current->data +
-										openfile->current_x, NULL, NULL);
+										openfile->current_x);
 		size_t line_len = strlen(openfile->current->data +
 										openfile->current_x);
 #ifndef NANO_TINY

src/prompt.c

@@ -253,7 +253,7 @@ void do_statusbar_right(void)
 void do_statusbar_delete(void)
 {
 	if (answer[typing_x] != '\0') {
-		int charlen = parse_mbchar(answer + typing_x, NULL, NULL);
+		int charlen = char_length(answer + typing_x);
 
 		charmove(answer + typing_x, answer + typing_x + charlen,
 						strlen(answer) - typing_x - charlen + 1);

src/proto.h

@@ -213,6 +213,7 @@ char control_mbrep(const char *c, bool isdata);
 int length_of_char(const char *c, int *width);
 int mbwidth(const char *c);
 char *make_mbchar(long chr, int *chr_mb_len);
+int char_length(const char *pointer);
 int parse_mbchar(const char *buf, char *chr, size_t *col);
 size_t move_mbleft(const char *buf, size_t pos);
 size_t move_mbright(const char *buf, size_t pos);

src/rcfile.c

@@ -1226,8 +1226,8 @@ void parse_rcfile(FILE *rcstream, bool syntax_only, bool headers_only)
 				free(option);
 			} else {
 				whitespace = option;
-				whitelen[0] = parse_mbchar(whitespace, NULL, NULL);
+				whitelen[0] = char_length(whitespace);
-				whitelen[1] = parse_mbchar(whitespace + whitelen[0], NULL, NULL);
+				whitelen[1] = char_length(whitespace + whitelen[0]);
 			}
 		} else
 #endif

src/search.c

@@ -931,7 +931,7 @@ void do_find_bracket(void)
 
 	/* Find the halfway point in matchbrackets, where the closing ones start. */
 	for (size_t i = 0; i < charcount; i++)
-		halfway += parse_mbchar(matchbrackets + halfway, NULL, NULL);
+		halfway += char_length(matchbrackets + halfway);
 
 	/* When on a closing bracket, we have to search backwards for a matching
 	 * opening bracket; otherwise, forward for a matching closing bracket. */
@@ -948,8 +948,8 @@ void do_find_bracket(void)
 			wanted_ch += move_mbright(wanted_ch, 0);
 	}
 
-	ch_len = parse_mbchar(ch, NULL, NULL);
+	ch_len = char_length(ch);
-	wanted_ch_len = parse_mbchar(wanted_ch, NULL, NULL);
+	wanted_ch_len = char_length(wanted_ch);
 
 	/* Copy the two complementary brackets into a single string. */
 	strncpy(bracket_pair, ch, ch_len);

src/text.c

@@ -1548,7 +1548,7 @@ ssize_t break_line(const char *line, ssize_t goal, bool snap_at_nl)
 			else if (lastblank > 0)
 				return lastblank;
 
-			charlen = parse_mbchar(line, NULL, NULL);
+			charlen = char_length(line);
 			line += charlen;
 			index += charlen;
 		}
@@ -1558,13 +1558,13 @@ ssize_t break_line(const char *line, ssize_t goal, bool snap_at_nl)
 
 	/* Move the pointer back to the last blank, and then step beyond it. */
 	line = line - index + lastblank;
-	charlen = parse_mbchar(line, NULL, NULL);
+	charlen = char_length(line);
 	line += charlen;
 
 	/* Skip any consecutive blanks after the last blank. */
 	while (*line != '\0' && is_blank_mbchar(line)) {
 		lastblank += charlen;
-		charlen = parse_mbchar(line, NULL, NULL);
+		charlen = char_length(line);
 		line += charlen;
 	}
 
@@ -1599,7 +1599,7 @@ size_t indent_length(const char *line)
 /* Copy a character from one place to another. */
 void copy_character(char **from, char **to)
 {
-	int charlen = parse_mbchar(*from, NULL, NULL);
+	int charlen = char_length(*from);
 
 	if (*from == *to) {
 		*from += charlen;
@@ -1624,11 +1624,11 @@ void squeeze(linestruct *line, size_t skip)
 	 * pass over all blanks after these; 3) leave anything else unchanged. */
 	while (*from != '\0') {
 		if (is_blank_mbchar(from)) {
-			from += parse_mbchar(from, NULL, NULL);
+			from += char_length(from);
 			*(to++) = ' ';
 
 			while (*from != '\0' && is_blank_mbchar(from))
-				from += parse_mbchar(from, NULL, NULL);
+				from += char_length(from);
 		} else if (mbstrchr(punct, from) != NULL) {
 			copy_character(&from, &to);
 
@@ -1636,16 +1636,16 @@ void squeeze(linestruct *line, size_t skip)
 				copy_character(&from, &to);
 
 			if (*from != '\0' && is_blank_mbchar(from)) {
-				from += parse_mbchar(from, NULL, NULL);
+				from += char_length(from);
 				*(to++) = ' ';
 			}
 			if (*from != '\0' && is_blank_mbchar(from)) {
-				from += parse_mbchar(from, NULL, NULL);
+				from += char_length(from);
 				*(to++) = ' ';
 			}
 
 			while (*from != '\0' && is_blank_mbchar(from))
-				from += parse_mbchar(from, NULL, NULL);
+				from += char_length(from);
 		} else
 			copy_character(&from, &to);
 	}

src/winio.c

@@ -1903,7 +1903,7 @@ char *display_string(const char *buf, size_t column, size_t span,
 			if (start_col < column) {
 				converted[index++] = control_mbrep(buf, isdata);
 				column++;
-				buf += parse_mbchar(buf, NULL, NULL);
+				buf += char_length(buf);
 			}
 		}
 #ifdef ENABLE_UTF8
@@ -1916,7 +1916,7 @@ char *display_string(const char *buf, size_t column, size_t span,
 			/* Display the right half of a two-column character as ']'. */
 			converted[index++] = ']';
 			column++;
-			buf += parse_mbchar(buf, NULL, NULL);
+			buf += char_length(buf);
 		}
 #endif
 	}