Re: less-374 patch (was Re: less 358)

[Jungshik Shin]

On Mon, 20 May 2002, Jungshik Shin wrote:
> On Mon, 20 May 2002, Zvi Har'El wrote:

> > I am using less on a UTF-8 Redhat Linux 7.3 machine. I am having troubles
> > with using man, because of the overstiking is not handled properly.
> > I read the

>   Anyway, attached is a *simplistic*(not perfect)  patch against
> less 374(the newest at less home page)

  Attached is a newer patch. This one works better than the previous
one although it still has the problem mentioned in my previous
message(and quoted below). With this, two hyphens in a row(U+2010) overstruck
with themselves are rendered correctly.  I don't have to be bothered
by hollow boxes(for U+2010) in man pages any more.

> 'pattern'). It's not perfect because it only checks the first
> octet of a two or three octet-long UTF-8 char to see if it's
> identical with the char. preceding backspace.

   Jungshik Shin

--- line.c.orig Mon May 20 11:56:34 2002
+++ line.c      Mon May 20 14:38:02 2002
@@ -555,6 +555,13 @@
        return (r);
 }
 
+#define IS_UTF8_4BYTE(c) ( ((c) & 0xf8) == 0xf0 )
+#define IS_UTF8_3BYTE(c) ( ((c) & 0xf0) == 0xe0 )
+#define IS_UTF8_2BYTE(c) ( ((c) & 0xe0) == 0xc0 )
+#define IS_UTF8_TRAIL(c) ( ((c) & 0xc0) == 0x80 )
+
+
+
        static int
 do_append(c, pos)
        int c;
@@ -592,12 +599,20 @@
                 * or just deletion of the character in the buffer.
                 */
                overstrike--;
-               if (utf_mode && curr > 1 && (char)c == linebuf[curr-2])
+               if (utf_mode && IS_UTF8_4BYTE(c) && curr > 2 && (char)c == 
+linebuf[curr-3])
+               {
+                       backc();
+                       backc();
+                       backc();
+                       STORE_CHAR(linebuf[curr], AT_BOLD, pos);
+                       overstrike = 3;
+               } else if (utf_mode && (IS_UTF8_3BYTE(c) || overstrike==2 && 
+IS_UTF8_TRAIL(c)) && curr > 1 && (char)c == linebuf[curr-2])
                {
                        backc();
                        backc();
+                       STORE_CHAR(linebuf[curr], AT_BOLD, pos);
                        overstrike = 2;
-               } else if (utf_mode && curr > 0 && (char)c == linebuf[curr-1])
+               } else if (utf_mode && curr > 0 && (IS_UTF8_2BYTE(c) || overstrike==1 
+&& IS_UTF8_TRAIL(c)) && (char)c == linebuf[curr-1])
                {
                        backc();
                        STORE_CHAR(linebuf[curr], AT_BOLD, pos);