1) Make ctype-aware expect chars in range 0200-0204 treated

as controls in any case.
2) Fix bug with incorrect column position when standouts occurse
3) Fix bug when last standouted char cause clearing next line
in standout

8 files changed, 85 insertions, 53 deletions

diff --git a/usr.bin/more/ch.c b/usr.bin/more/ch.c
index 668e601..a971c76 100644
--- a/usr.bin/more/ch.c
+++ b/usr.bin/more/ch.c
@@ -96,7 +96,7 @@ static off_t last_piped_pos;
 #define	ch_get() \
 	((buf_head->block == ch_block && \
 	    ch_offset < buf_head->datasize) ? \
-	    buf_head->data[ch_offset] : fch_get())
+	    buf_head->data[ch_offset] & 0xff : fch_get())
 
 static
 fch_get()
@@ -130,7 +130,7 @@ fch_get()
 			 * find it already buffered.
 			 */
 			if (ispipe)
-				return(bp->data[ch_offset]);
+				return(bp->data[ch_offset] & 0xff);
 			goto found;
 		}
 	/*
@@ -189,15 +189,15 @@ read_more:
 
 	if (bs_mode) {
 		for (p = &bp->data[bp->datasize]; --n >= 0;) {
-			*--p &= 0177;
+			*--p;
 			if (*p == EOI)
 				*p = 0200;
 		}
 	}
 	else {
 		for (t = p; --n >= 0; ++p) {
-			ch = *p & 0177;
-			if (ch == '\r' && n && (p[1] & 0177) == '\n') {
+			ch = *p;
+			if (ch == '\r' && n && p[1] == '\n') {
 				++p;
 				*t++ = '\n';
 			}
@@ -233,7 +233,7 @@ found:
 		 */
 		goto read_more;
 
-	return(bp->data[ch_offset]);
+	return(bp->data[ch_offset] & 0xff);
 }
 
 /*
diff --git a/usr.bin/more/command.c b/usr.bin/more/command.c
index 51f5847..dcab842 100644
--- a/usr.bin/more/command.c
+++ b/usr.bin/more/command.c
@@ -76,6 +76,7 @@ static int wsearch;		/* Search for matches (1) or non-matches (0) */
 static
 cmd_erase()
 {
+	int c;
 	/*
 	 * backspace past beginning of the string: this usually means
 	 * abort the command.
@@ -84,7 +85,8 @@ cmd_erase()
 		return(1);
 
 	/* erase an extra character, for the carat. */
-	if (CONTROL_CHAR(*--cp)) {
+	c = *--cp & 0xff;
+	if (CONTROL_CHAR(c)) {
 		backspace();
 		--cmd_col;
 	}
@@ -119,9 +121,9 @@ cmd_char(c)
 	/* in this order, in case werase == erase_char */
 	if (c == werase_char) {
 		if (cp > cmdbuf) {
-			while (isspace(cp[-1]) && !cmd_erase());
-			while (!isspace(cp[-1]) && !cmd_erase());
-			while (isspace(cp[-1]) && !cmd_erase());
+			while (isspace(cp[-1] & 0xff) && !cmd_erase());
+			while (!isspace(cp[-1] & 0xff) && !cmd_erase());
+			while (isspace(cp[-1] & 0xff) && !cmd_erase());
 		}
 		return(cp == cmdbuf);
 	}
@@ -141,6 +143,7 @@ cmd_char(c)
 		if (CONTROL_CHAR(c)) {
 			putchr('^');
 			cmd_col++;
+			c &= ~0200;
 			c = CARAT_CHAR(c);
 		}
 		putchr(c);
@@ -273,11 +276,11 @@ exec_mca()
 		(void)search(0, cmdbuf, number, wsearch);
 		break;
 	case A_EXAMINE:
-		for (p = cmdbuf; isspace(*p); ++p);
+		for (p = cmdbuf; isspace(*p & 0xff); ++p);
 		(void)edit(glob(p));
 		break;
 	case A_TAGFILE:
-		for (p = cmdbuf; isspace(*p); ++p);
+		for (p = cmdbuf; isspace(*p & 0xff); ++p);
 		findtag(p);
 		if (tagfile == NULL)
 			break;
@@ -573,6 +576,7 @@ again:		if (sigs)
 				start_mca(A_PREFIX, "");
 			if (CONTROL_CHAR(c)) {
 				putchr('^');
+				c &= ~0200;
 				c = CARAT_CHAR(c);
 			}
 			putchr(c);
diff --git a/usr.bin/more/less.h b/usr.bin/more/less.h
index 70de6de..c4dd23f 100644
--- a/usr.bin/more/less.h
+++ b/usr.bin/more/less.h
@@ -47,7 +47,7 @@
 #define	BO_CHAR		'\203'		/* Enter boldface mode */
 #define	BE_CHAR		'\204'		/* Exit boldface mode */
 
-#define	CONTROL_CHAR(c)		(iscntrl(c))
+#define CONTROL_CHAR(c)         (!isprint(c))
 #define	CARAT_CHAR(c)		((c == '\177') ? '?' : (c | 0100))
 
 #define	TOP		(0)
diff --git a/usr.bin/more/line.c b/usr.bin/more/line.c
index 7634f35..bdef948 100644
--- a/usr.bin/more/line.c
+++ b/usr.bin/more/line.c
@@ -167,10 +167,9 @@ pappend(c)
 		switch (ln_state) {
 		case LN_NORMAL:
 			if (curr <= linebuf + 1
-			    || curr[-1] != (char)('H' | 0200))
+			    || curr[-1] != '\b')
 				break;
-			column -= 2;
-			if (c == curr[-2])
+			if (c == (curr[-2] & 0xff))
 				goto enter_boldface;
 			if (c == '_' || curr[-2] == '_')
 				goto enter_underline;
@@ -183,7 +182,7 @@ enter_boldface:
 			 * Switch into boldface mode.
 			 */
 			column--;
-			if (column + bo_width + be_width + 1 >= sc_width)
+			if (column + bo_width + be_width >= sc_width)
 				/*
 				 * Not enough room left on the screen to 
 				 * enter and exit boldface mode.
@@ -213,7 +212,7 @@ enter_underline:
 			 * the current char).  Switch into underline mode.
 			 */
 			column--;
-			if (column + ul_width + ue_width + 1 >= sc_width)
+			if (column + ul_width + ue_width >= sc_width)
 				/*
 				 * Not enough room left on the screen to 
 				 * enter and exit underline mode.
@@ -243,7 +242,7 @@ enter_underline:
 			/*
 			 * Termination of a sequence "_\bX" or "X\b_".
 			 */
-			if (c != '_' && curr[-2] != '_' && c == curr[-2])
+			if (c != '_' && curr[-2] != '_' && c == (curr[-2] & 0xff))
 			{
 				/*
 				 * We seem to have run on from underlining
@@ -260,7 +259,7 @@ enter_underline:
 			}
 ln_ul_xb_case:
 			if (c == '_')
-				c = curr[-2];
+				c = curr[-2] & 0xff;
 			curr -= 2;
 			ln_state = LN_UNDERLINE;
 			break;
@@ -268,7 +267,7 @@ ln_ul_xb_case:
 			/*
 			 * Termination of a sequnce "X\bX".
 			 */
-			if (c != curr[-2] && (c == '_' || curr[-2] == '_'))
+			if (c != (curr[-2] & 0xff) && (c == '_' || curr[-2] == '_'))
 			{
 				/*
 				 * We seem to have run on from
@@ -286,7 +285,7 @@ ln_bo_xb_case:
 			ln_state = LN_BOLDFACE;
 			break;
 		case LN_UNDERLINE:
-			if (column + ue_width + bo_width + 1 + be_width >= sc_width)
+			if (column + ue_width + bo_width + be_width >= sc_width)
 				/*
 				 * We have just barely enough room to 
 				 * exit underline mode and handle a possible
@@ -301,7 +300,7 @@ ln_bo_xb_case:
 				ln_state = LN_BO_XB;
 				break;
 			}
-			if (column + be_width + ul_width + 1 + ue_width >= sc_width)
+			if (column + be_width + ul_width + ue_width >= sc_width)
 				/*
 				 * We have just barely enough room to 
 				 * exit underline mode and handle a possible
@@ -378,30 +377,31 @@ ln_bo_xb_case:
 
 	if (c == '\b') {
 		if (ln_state == LN_NORMAL)
-			NEW_COLUMN(2);
+			NEW_COLUMN(0);
 		else
 			column--;
-		*curr++ = ('H' | 0200);
+		*curr++ = c;
 		return(0);
 	} 
 
-	if (CONTROL_CHAR(c)) {
-		/*
-		 * Put a "^X" into the buffer.  The 0200 bit is used to tell
-		 * put_line() to prefix the char with a ^.  We don't actually
-		 * put the ^ in the buffer because we sometimes need to move
-		 * chars around, and such movement might separate the ^ from
-		 * its following character.
-		 */
+	switch ((char)c) {
+	case UL_CHAR:
+	case UE_CHAR:
+	case BO_CHAR:
+	case BE_CHAR:
+		c &= ~0200;
+		/* fall through */
+	case '\200':
 		NEW_COLUMN(2);
-		*curr++ = (CARAT_CHAR(c) | 0200);
-		return(0);
+		break;
+	default:
+		if (CONTROL_CHAR(c))
+			NEW_COLUMN(2);
+		else
+			NEW_COLUMN(1);
+		break;
 	}
 
-	/*
-	 * Ordinary character.  Just put it in the buffer.
-	 */
-	NEW_COLUMN(1);
 	*curr++ = c;
 	return (0);
 }
diff --git a/usr.bin/more/output.c b/usr.bin/more/output.c
index 2205f1f..4524bd5 100644
--- a/usr.bin/more/output.c
+++ b/usr.bin/more/output.c
@@ -54,6 +54,8 @@ extern int tabstop;
 extern int screen_trashed;
 extern int any_display;
 extern char *line;
+extern int mode_flags;
+static int last_pos_highlighted = 0;
 
 /* display the line which is in the line buffer. */
 put_line()
@@ -75,14 +77,19 @@ put_line()
 	if (line == NULL)
 		line = "";
 
+	if (last_pos_highlighted)
+	{
+		clear_eol();
+		last_pos_highlighted = 0;
+	}
 	column = 0;
 	for (p = line;  *p != '\0';  p++)
 	{
-		switch (c = *p)
+		switch ((char)(c = *p & 0xff))
 		{
 		case UL_CHAR:
 			ul_enter();
-			column += ul_width +1;
+			column += ul_width;
 			break;
 		case UE_CHAR:
 			ul_exit();
@@ -90,7 +97,7 @@ put_line()
 			break;
 		case BO_CHAR:
 			bo_enter();
-			column += bo_width +1;
+			column += bo_width;
 			break;
 		case BE_CHAR:
 			bo_exit();
@@ -108,15 +115,11 @@ put_line()
 			column--;
 			break;
 		default:
-			if (c & 0200)
+			if (c == 0200 || CONTROL_CHAR(c))
 			{
-				/*
-				 * Control characters arrive here as the
-				 * normal character [CARAT_CHAR(c)] with
-				 * the 0200 bit set.  See pappend().
-				 */
+				c &= ~0200;
 				putchr('^');
-				putchr(c & 0177);
+				putchr(CARAT_CHAR(c));
 				column += 2;
 			} else
 			{
@@ -124,6 +127,8 @@ put_line()
 				column++;
 			}
 		}
+		if (column == sc_width && mode_flags)
+			last_pos_highlighted = 1;
 	}
 	if (column < sc_width || !auto_wrap || ignaw)
 		putchr('\n');
diff --git a/usr.bin/more/prim.c b/usr.bin/more/prim.c
index d5af8f3..99361cc 100644
--- a/usr.bin/more/prim.c
+++ b/usr.bin/more/prim.c
@@ -611,8 +611,8 @@ search(search_forward, pattern, n, wantmatch)
 	 */
 	if (caseless && pattern != NULL)
 		for (p = pattern;  *p;  p++)
-			if (isupper(*p))
-				*p = tolower(*p);
+			if (isupper(*p & 0xff))
+				*p = tolower(*p & 0xff);
 #ifdef RECOMP
 
 	/*
@@ -764,7 +764,7 @@ search(search_forward, pattern, n, wantmatch)
 		 */
 		if (caseless)
 			for (p = q = line;  *p;  p++, q++)
-				*q = isupper(*p) ? tolower(*p) : *p;
+				*q = isupper(*p & 0xff) ? tolower(*p & 0xff) : *p;
 
 		/*
 		 * Remove any backspaces along with the preceeding char.
diff --git a/usr.bin/more/screen.c b/usr.bin/more/screen.c
index edfae1e..2dc99ec 100644
--- a/usr.bin/more/screen.c
+++ b/usr.bin/more/screen.c
@@ -106,6 +106,11 @@ int bo_width, be_width;		/* Printing width of boldface sequences */
 int ul_width, ue_width;		/* Printing width of underline sequences */
 int so_width, se_width;		/* Printing width of standout sequences */
 
+int mode_flags = 0;
+#define M_SO 1
+#define M_UL 2
+#define M_BO 4
+
 /*
  * These two variables are sometimes defined in,
  * and needed by, the termcap library.
@@ -503,6 +508,12 @@ bell()
  */
 clear()
 {
+	if (mode_flags & M_SO)
+		so_exit();
+	if (mode_flags & M_UL)
+		ul_exit();
+	if (mode_flags & M_BO)
+		bo_exit();
 	tputs(sc_clear, sc_height, putchr);
 }
 
@@ -512,6 +523,12 @@ clear()
  */
 clear_eol()
 {
+	if (mode_flags & M_SO)
+		so_exit();
+	if (mode_flags & M_UL)
+		ul_exit();
+	if (mode_flags & M_BO)
+		bo_exit();
 	tputs(sc_eol_clear, 1, putchr);
 }
 
@@ -521,6 +538,7 @@ clear_eol()
 so_enter()
 {
 	tputs(sc_s_in, 1, putchr);
+	mode_flags |= M_SO;
 }
 
 /*
@@ -529,6 +547,7 @@ so_enter()
 so_exit()
 {
 	tputs(sc_s_out, 1, putchr);
+	mode_flags &= ~M_SO;
 }
 
 /*
@@ -538,6 +557,7 @@ so_exit()
 ul_enter()
 {
 	tputs(sc_u_in, 1, putchr);
+	mode_flags |= M_UL;
 }
 
 /*
@@ -546,6 +566,7 @@ ul_enter()
 ul_exit()
 {
 	tputs(sc_u_out, 1, putchr);
+	mode_flags &= ~M_UL;
 }
 
 /*
@@ -554,6 +575,7 @@ ul_exit()
 bo_enter()
 {
 	tputs(sc_b_in, 1, putchr);
+	mode_flags |= M_BO;
 }
 
 /*
@@ -562,6 +584,7 @@ bo_enter()
 bo_exit()
 {
 	tputs(sc_b_out, 1, putchr);
+	mode_flags &= ~M_BO;
 }
 
 /*
diff --git a/usr.bin/more/ttyin.c b/usr.bin/more/ttyin.c
index 52ff92e..2d3567e 100644
--- a/usr.bin/more/ttyin.c
+++ b/usr.bin/more/ttyin.c
@@ -75,5 +75,5 @@ getchr()
 			quit();
 		}
 	} while (result != 1);
-	return (c & 0177);
+	return (c & 0xff);
 }