First draft of Unicode support in pterm. It's pretty complete: it

does UTF-8 copy and paste (falling back to normal strings if
necessary), it understands X font encodings and translates things
accordingly so that if you have a Unicode font you can ask for
virtually any single-byte encoding and get it (Mac-Roman pterm,
anyone?), and so on. There's work left to be done (wide fonts for
CJK spring to mind), but I reckon this is a pretty good start.

[originally from svn r2395]

unix/uxucs.c

@@ -1,17 +1,18 @@
 #include <stdio.h>
 #include <stdlib.h>
 #include <ctype.h>
+#include <locale.h>
+#include <limits.h>
+#include <wchar.h>
 
 #include <time.h>
+
 #include "putty.h"
 #include "terminal.h"
 #include "misc.h"
 
 /*
  * Unix Unicode-handling routines.
- * 
- * FIXME: currently trivial stub versions assuming all codepages
- * are ISO8859-1.
  */
 
 int is_dbcs_leadbyte(int codepage, char byte)
@@ -22,48 +23,151 @@ int is_dbcs_leadbyte(int codepage, char byte)
 int mb_to_wc(int codepage, int flags, char *mbstr, int mblen,
 	     wchar_t *wcstr, int wclen)
 {
-    int ret = 0;
-    while (mblen > 0 && wclen > 0) {
-	*wcstr++ = (unsigned char) *mbstr++;
-	mblen--, wclen--, ret++;
-    }
-    return ret;			       /* FIXME: check error codes! */
+    if (codepage == DEFAULT_CODEPAGE) {
+	int n = 0;
+	mbstate_t state = { 0 };
+
+	setlocale(LC_CTYPE, "");
+
+	while (mblen > 0) {
+	    size_t i = mbrtowc(wcstr+n, mbstr, (size_t)mblen, &state);
+	    if (i == (size_t)-1 || i == (size_t)-2)
+		break;
+	    n++;
+	    mbstr += i;
+	    mblen -= i;
+	}
+
+	setlocale(LC_CTYPE, "C");
+
+	return n;
+    } else
+	return charset_to_unicode(&mbstr, &mblen, wcstr, wclen, codepage,
+				  NULL, NULL, 0);
 }
 
 int wc_to_mb(int codepage, int flags, wchar_t *wcstr, int wclen,
 	     char *mbstr, int mblen, char *defchr, int *defused)
 {
-    int ret = 0;
+    /* FIXME: we should remove the defused param completely... */
     if (defused)
 	*defused = 0;
-    while (mblen > 0 && wclen > 0) {
-	if (*wcstr >= 0x100) {
-	    if (defchr)
-		*mbstr++ = *defchr;
-	    else
-		*mbstr++ = '.';
-	    if (defused)
-		*defused = 1;
-	} else
-	    *mbstr++ = (unsigned char) *wcstr;
-	wcstr++;
-	mblen--, wclen--, ret++;
-    }
-    return ret;			       /* FIXME: check error codes! */
+
+    if (codepage == DEFAULT_CODEPAGE) {
+	char output[MB_LEN_MAX];
+	mbstate_t state = { 0 };
+	int n = 0;
+
+	setlocale(LC_CTYPE, "");
+
+	while (wclen > 0) {
+	    int i = wcrtomb(output, wcstr[0], &state);
+	    if (i == (size_t)-1 || i > n - mblen)
+		break;
+	    memcpy(mbstr+n, output, i);
+	    n += i;
+	    wcstr++;
+	    wclen--;
+	}
+
+	setlocale(LC_CTYPE, "C");
+
+	return n;
+    } else
+	return charset_from_unicode(&wcstr, &wclen, mbstr, mblen, codepage,
+				    NULL, NULL, 0);
 }
 
 void init_ucs(void)
 {
     int i;
-    /* Find the line control characters. FIXME: this is not right. */
+
+    /*
+     * In the platform-independent parts of the code, font_codepage
+     * is used only for system DBCS support - which we don't
+     * support at all. So we set this to something which will never
+     * be used.
+     */
+    font_codepage = -1;
+
+    /*
+     * line_codepage should be decoded from the specification in
+     * cfg.
+     */
+    line_codepage = charset_from_mimeenc(cfg.line_codepage);
+    if (line_codepage == CS_NONE)
+	line_codepage = charset_from_xenc(cfg.line_codepage);
+    /* If it's still CS_NONE, we should assume direct-to-font. */
+
+    /* FIXME: this is a hack. Currently fonts with incomprehensible
+     * encodings are dealt with by pretending they're 8859-1. It's
+     * ugly, but it's good enough to stop things crashing. Should do
+     * something better here. */
+    if (line_codepage == CS_NONE)
+	line_codepage = CS_ISO8859_1;
+
+    /*
+     * Set up unitab_line, by translating each individual character
+     * in the line codepage into Unicode.
+     */
+    for (i = 0; i < 256; i++) {
+	char c[1], *p;
+	wchar_t wc[1];
+	int len;
+	c[0] = i;
+	p = c;
+	len = 1;
+	if (1 == charset_to_unicode(&p,&len,wc,1,line_codepage,NULL,L"",0))
+	    unitab_line[i] = wc[0];
+	else
+	    unitab_line[i] = 0xFFFD;
+    }
+
+    /*
+     * Set up unitab_xterm. This is the same as unitab_line except
+     * in the line-drawing regions, where it follows the Unicode
+     * encoding.
+     * 
+     * (Note that the strange X encoding of line-drawing characters
+     * in the bottom 32 glyphs of ISO8859-1 fonts is taken care of
+     * by the font encoding, which will spot such a font and act as
+     * if it were in a variant encoding of ISO8859-1.)
+     */
+    for (i = 0; i < 256; i++) {
+	static const wchar_t unitab_xterm_std[32] = {
+	    0x2666, 0x2592, 0x2409, 0x240c, 0x240d, 0x240a, 0x00b0, 0x00b1,
+	    0x2424, 0x240b, 0x2518, 0x2510, 0x250c, 0x2514, 0x253c, 0x23ba,
+	    0x23bb, 0x2500, 0x23bc, 0x23bd, 0x251c, 0x2524, 0x2534, 0x252c,
+	    0x2502, 0x2264, 0x2265, 0x03c0, 0x2260, 0x00a3, 0x00b7, 0x0020
+	};
+	if (i >= 0x5F && i < 0x7F)
+	    unitab_xterm[i] = unitab_xterm_std[i & 0x1F];
+	else
+	    unitab_xterm[i] = unitab_line[i];
+    }
+
+    /*
+     * Set up unitab_scoacs. The SCO Alternate Character Set is
+     * simply CP437.
+     */
+    for (i = 0; i < 256; i++) {
+	char c[1], *p;
+	wchar_t wc[1];
+	int len;
+	c[0] = i;
+	p = c;
+	len = 1;
+	if (1 == charset_to_unicode(&p,&len,wc,1,CS_CP437,NULL,L"",0))
+	    unitab_scoacs[i] = wc[0];
+	else
+	    unitab_scoacs[i] = 0xFFFD;
+    }
+
+    /* Find the line control characters. */
     for (i = 0; i < 256; i++)
-	if (i < ' ' || (i >= 0x7F && i < 0xA0))
+	if (unitab_line[i] < ' '
+	    || (unitab_line[i] >= 0x7F && unitab_line[i] < 0xA0))
 	    unitab_ctrl[i] = i;
 	else
 	    unitab_ctrl[i] = 0xFF;
-
-    for (i = 0; i < 256; i++) {
-	unitab_line[i] = unitab_scoacs[i] = i;
-	unitab_xterm[i] = (i >= 0x5F && i < 0x7F) ? ((i+1) & 0x1F) : i;
-    }
 }