From c11d303f41f15a1a25dbf1ffcbebf91b86ef9556 Mon Sep 17 00:00:00 2001
From: Ingo Schwarze <schwarze@cvs.openbsd.org>
Date: Sat, 31 Aug 2019 13:44:30 +0000
Subject: New utility function mbtowc_left() because this functionality -
 decoding a UTF-8 multibyte character to the left of a given byte - is already
 needed at three places in line.c and will also be needed for cleanup work in
 cmdbuf.c in the future. OK millert@

---
 usr.bin/less/charset.c | 21 ++++++++++++++++++++
 usr.bin/less/funcs.h   |  1 +
 usr.bin/less/line.c    | 53 ++++++++------------------------------------------
 3 files changed, 30 insertions(+), 45 deletions(-)

diff --git a/usr.bin/less/charset.c b/usr.bin/less/charset.c
index 4be75d678a9..5739fe1b71a 100644
--- a/usr.bin/less/charset.c
+++ b/usr.bin/less/charset.c
@@ -145,6 +145,27 @@ init_charset(void)
 	setbinfmt("LESSUTFBINFMT", &utfbinfmt, "<U+%04lX>");
 }
 
+/*
+ * Like mbtowc(3), except that it converts the multibyte character
+ * preceding ps rather than the one starting at ps.
+ */
+int
+mbtowc_left(wchar_t *pwc, const char *ps, size_t psz)
+{
+	size_t sz = 0;
+	int len;
+
+	do {
+		if (++sz > psz)
+			return -1;
+	} while (utf_mode && IS_UTF8_TRAIL(ps[-sz]));
+	if ((len = mbtowc(pwc, ps - sz, sz)) == -1) {
+		(void)mbtowc(NULL, NULL, 0);
+		return -1;
+	}
+	return len == sz || (len == 0 && sz == 1) ? len : -1;
+}
+
 /*
  * Is a given character a "control" character?
  */
diff --git a/usr.bin/less/funcs.h b/usr.bin/less/funcs.h
index 2bfab74475e..5ef595a3bbe 100644
--- a/usr.bin/less/funcs.h
+++ b/usr.bin/less/funcs.h
@@ -55,6 +55,7 @@ void ch_set_eof(void);
 void ch_init(int, int);
 void ch_close(void);
 int ch_getflags(void);
+int mbtowc_left(wchar_t *, const char *, size_t);
 void init_charset(void);
 char *prchar(LWCHAR);
 char *prutfchar(LWCHAR);
diff --git a/usr.bin/less/line.c b/usr.bin/less/line.c
index aa1a2b834fc..123353d2121 100644
--- a/usr.bin/less/line.c
+++ b/usr.bin/less/line.c
@@ -437,44 +437,20 @@ backc(void)
 	wchar_t	 ch, prev_ch;
 	int	 i, len, width;
 
-	i = curr - 1;
-	if (utf_mode) {
-		while (i >= lmargin && IS_UTF8_TRAIL(linebuf[i]))
-			i--;
-	}
-	if (i < lmargin)
+	if ((len = mbtowc_left(&ch, linebuf + curr, curr)) <= 0)
 		return (0);
-	if (utf_mode) {
-		len = mbtowc(&ch, linebuf + i, curr - i);
-		if (len == -1 || i + len < curr) {
-			(void)mbtowc(NULL, NULL, MB_CUR_MAX);
-			return (0);
-		}
-	} else
-		ch = linebuf[i];
+	curr -= len;
 
 	/* This assumes that there is no '\b' in linebuf.  */
-	while (curr > lmargin && column > lmargin &&
-	    (!(attr[curr - 1] & (AT_ANSI|AT_BINARY)))) {
-		curr = i--;
-		if (utf_mode) {
-			while (i >= lmargin && IS_UTF8_TRAIL(linebuf[i]))
-				i--;
-		}
-		if (i < lmargin)
+	while (curr >= lmargin && column > lmargin &&
+	    !(attr[curr] & (AT_ANSI|AT_BINARY))) {
+		if ((len = mbtowc_left(&prev_ch, linebuf + curr, curr)) <= 0)
 			prev_ch = L'\0';
-		else if (utf_mode) {
-			len = mbtowc(&prev_ch, linebuf + i, curr - i);
-			if (len == -1 || i + len < curr) {
-				(void)mbtowc(NULL, NULL, MB_CUR_MAX);
-				prev_ch = L'\0';
-			}
-		} else
-			prev_ch = linebuf[i];
 		width = pwidth(ch, attr[curr], prev_ch);
 		column -= width;
 		if (width > 0)
 			return (1);
+		curr -= len;
 		if (prev_ch == L'\0')
 			return (0);
 		ch = prev_ch;
@@ -554,21 +530,8 @@ store_char(LWCHAR ch, char a, char *rep, off_t pos)
 	}
 	if (w == -1) {
 		wchar_t prev_ch;
-
-		if (utf_mode) {
-			for (i = curr - 1; i >= 0; i--)
-				if (!IS_UTF8_TRAIL(linebuf[i]))
-					break;
-			if (i >= 0) {
-				w = mbtowc(&prev_ch, linebuf + i, curr - i);
-				if (w == -1 || i + w < curr) {
-					(void)mbtowc(NULL, NULL, MB_CUR_MAX);
-					prev_ch = L' ';
-				}
-			} else
-				prev_ch = L' ';
-		} else
-			prev_ch = curr > 0 ? linebuf[curr - 1] : L' ';
+		if (mbtowc_left(&prev_ch, linebuf + curr, curr) <= 0)
+			prev_ch = L' ';
 		w = pwidth(ch, a, prev_ch);
 	}
 
-- 
cgit v1.2.3