xd.  add rune formatting, and also add an optional argument
to -s, specifying the swab length.  the default behavior is not changed.

Reference: /n/atom/patch/applied2013/xdrs
Date: Thu Aug  1 18:06:42 CES 2013
Signed-off-by: quanstro@quanstro.net

--- /sys/src/cmd/xd.c	Thu Aug  1 18:06:16 2013
+++ /sys/src/cmd/xd.c	Thu Aug  1 18:06:16 2013
@@ -2,10 +2,11 @@
 #include <libc.h>
 #include <bio.h>
 
-unsigned char	odata[16];
-unsigned char	data[16];
+uchar		odata[16];
+uchar		data[32];
 int		ndata;
-unsigned long	addr;
+int		nread;
+ulong		addr;
 int		repeats;
 int		swizzle;
 int		flush;
@@ -14,13 +15,17 @@
 void		xprint(char *, ...);
 void		initarg(void), swizz(void);
 enum{
-	Narg=10
+	Narg=10,
+
+	TNone=0,
+	TAscii,
+	TRune,
 };
 typedef struct Arg Arg;
 typedef void fmtfn(char *);
 struct Arg
 {
-	int	ascii;		/* 0==none, 1==ascii */
+	int	chartype;	/* TNone, TAscii, TRunes */
 	int	loglen;		/* 0==1, 1==2, 2==4, 3==8 */
 	int	base;		/* 0==8, 1==10, 2==16 */
 	fmtfn	*fn;		/* function to call with data */
@@ -29,7 +34,7 @@
 }arg[Narg];
 int	narg;
 
-fmtfn	fmt0, fmt1, fmt2, fmt3, fmtc;
+fmtfn	fmt0, fmt1, fmt2, fmt3, fmtc, fmtr;
 fmtfn *fmt[4] = {
 	fmt0,
 	fmt1,
@@ -50,6 +55,10 @@
 	" %.3uo",	" %.3ud",	" %.2ux",
 };
 
+char *rfmt[1][1] = {
+	" %2.2C",
+};
+
 char *afmt[2][3] = {
 	"%.7luo ",	"%.7lud ",	"%.7lux ",
 	"%7luo ",	"%7lud ",	"%7lux ",
@@ -77,15 +86,11 @@
 				goto Usage;
 			continue;
 		}
-		if(argv[0][0] == 'S'){
-			swizzle = 2;
-			if(argv[0][1])
-				goto Usage;
-			continue;
-		}
 		if(argv[0][0] == 's'){
 			swizzle = 4;
 			if(argv[0][1])
+				swizzle = atoi(argv[0]+1);
+			if(swizzle <= 0 || swizzle > 8)
 				goto Usage;
 			continue;
 		}
@@ -119,7 +124,13 @@
 		while(argv[0][0]){
 			switch(argv[0][0]){
 			case 'c':
-				ap->ascii = 1;
+				ap->chartype = TAscii;
+				ap->loglen = 0;
+				if(argv[0][1] || argv[0][-1]!='-')
+					goto Usage;
+				break;
+			case 'R':
+				ap->chartype = TRune;
 				ap->loglen = 0;
 				if(argv[0][1] || argv[0][-1]!='-')
 					goto Usage;
@@ -151,12 +162,14 @@
 				break;
 			default:
 			Usage:
-   fprint(2, "usage: xd [-u] [-r] [-s] [-a{odx}] [-c|{b1w2l4v8}{odx}] ... file ...\n");
+   fprint(2, "usage: xd [-u] [-r] [-s[n]] [-a{odx}] [-c|-R|{b1w2l4v8}{odx}] ... file ...\n");
 				exits("usage");
 			}
 			argv[0]++;
 		}
-		if(ap->ascii)
+		if(ap->chartype == TRune)
+			ap->fn = fmtr;
+		else if(ap->chartype == TAscii)
 			ap->fn = fmtc;
 		else
 			ap->fn = fmt[ap->loglen];
@@ -184,7 +197,7 @@
 		fprint(2, "xd: too many formats (max %d)\n", Narg);
 		exits("usage");
 	}
-	ap->ascii = 0;
+	ap->chartype = TNone;
 	ap->loglen = 2;
 	ap->base = 2;
 	ap->fn = fmt2;
@@ -196,7 +209,7 @@
 xd(char *name, int title)
 {
 	int fd;
-	int i, star;
+	int i, star, nsee, nleft;
 	Arg *ap;
 	Biobuf *bp;
 
@@ -215,19 +228,27 @@
 		xprint("%s\n", name);
 	addr = 0;
 	star = 0;
-	while((ndata=Bread(bp, data, 16)) >= 0){
-		if(ndata < 16)
-			for(i=ndata; i<16; i++)
+	nsee = 16;
+	nleft = 0;
+	/* read 32 but see only 16 so that runes are happy */
+	while((ndata=Bread(bp, data + nleft, 32 - nleft)) >= 0){
+		ndata += nleft;
+		nleft = 0;
+		nread = ndata;
+		if(ndata>nsee)
+			ndata = nsee;
+		else if(ndata<nsee)
+			for(i=ndata; i<nsee; i++)
 				data[i] = 0;
 		if(swizzle)
 			swizz();
-		if(ndata==16 && repeats){
+		if(ndata==nsee && repeats){
 			if(addr>0 && data[0]==odata[0]){
-				for(i=1; i<16; i++)
+				for(i=1; i<nsee; i++)
 					if(data[i] != odata[i])
 						break;
-				if(i == 16){
-					addr += 16;
+				if(i == nsee){
+					addr += nsee;
 					if(star == 0){
 						star++;
 						xprint("*\n", 0);
@@ -235,7 +256,7 @@
 					continue;
 				}
 			}
-			for(i=0; i<16; i++)
+			for(i=0; i<nsee; i++)
 				odata[i] = data[i];
 			star = 0;
 		}
@@ -247,74 +268,44 @@
 				Bflush(&bout);
 		}
 		addr += ndata;
-		if(ndata<16){
+		if(ndata<nsee){
 			xprint(afmt[0][abase], addr);
 			xprint("\n", 0);
 			if(flush)
 				Bflush(&bout);
 			break;
 		}
+		if(nread>nsee){
+			nleft = nread - nsee;
+			memmove(data, data + nsee, nleft);
+		}
 	}
 	Bterm(bp);
 	return 0;
 }
 
-void
-swizz2(void)
-{
-	uchar *p, *q;
-	int i;
-	uchar swdata[16];
-
-	p = data;
-	q = swdata;
-	for(i=0; i<16; i++)
-		*q++ = *p++;
-	p = data;
-	q = swdata;
-	for(i=0; i<8; i++){
-		p[0] = q[1];
-		p[1] = q[0];
-		p += 2;
-		q += 2;
-	}
-}
+uchar order[] = {7, 6, 5, 4, 3, 2, 1, 0};
 
 void
-swizz4(void)
+swizz(void)
 {
-	uchar *p, *q;
-	int i;
+	uchar *p, *q, *t;
+	int i, j, n;
 	uchar swdata[16];
 
 	p = data;
 	q = swdata;
+	t = order + (8-swizzle);
 	for(i=0; i<16; i++)
 		*q++ = *p++;
 	p = data;
 	q = swdata;
-	for(i=0; i<4; i++){
-		p[0] = q[3];
-		p[1] = q[2];
-		p[2] = q[1];
-		p[3] = q[0];
-		p += 4;
-		q += 4;
-	}
-}
-
-void
-swizz(void)
-{
-	switch(swizzle){
-	case 2:
-		swizz2();
-		break;
-	case 4:
-		swizz4();
-		break;
-	default:
-		sysfatal("swizz");
+	n = 16/swizzle;
+	for(i=0; i<n; i++){
+		for(j = 0; j < swizzle; j++)
+			p[j] = q[t[j]];
+		p += j;
+		q += j;
 	}
 }
 
@@ -330,7 +321,7 @@
 fmt1(char *f)
 {
 	int i;
-	for(i=0; i<ndata; i+=sizeof(unsigned short))
+	for(i=0; i<ndata; i+=sizeof(ushort))
 		xprint(f, (data[i]<<8)|data[i+1]);
 }
 
@@ -338,7 +329,7 @@
 fmt2(char *f)
 {
 	int i;
-	for(i=0; i<ndata; i+=sizeof(unsigned long))
+	for(i=0; i<ndata; i+=sizeof(ulong))
 		xprint(f, (data[i]<<24)|(data[i+1]<<16)|(data[i+2]<<8)|data[i+3]);
 }
 
@@ -346,8 +337,9 @@
 fmt3(char *f)
 {
 	int i;
-	unsigned long long v;
-	for(i=0; i<ndata; i+=sizeof(unsigned long long)){
+	uvlong v;
+
+	for(i=0; i<ndata; i+=sizeof(uvlong)){
 		v = (data[i]<<24)|(data[i+1]<<16)|(data[i+2]<<8)|data[i+3];
 		v <<= 32;
 		v |= (data[i+4]<<24)|(data[i+1+4]<<16)|(data[i+2+4]<<8)|data[i+3+4];
@@ -359,32 +351,70 @@
 }
 
 void
+onefmtc(uchar c)
+{
+	switch(c){
+	case '\t':
+		xprint(cfmt[1][2], "\\t");
+		break;
+	case '\r':
+		xprint(cfmt[1][2], "\\r");
+		break;
+	case '\n':
+		xprint(cfmt[1][2], "\\n");
+		break;
+	case '\b':
+		xprint(cfmt[1][2], "\\b");
+		break;
+	default:
+		if(c>=0x7F || ' '>c)
+			xprint(cfmt[2][2], c);
+		else
+			xprint(cfmt[0][2], c);
+		break;
+	}
+}
+
+void
 fmtc(char *f)
 {
 	int i;
 
 	USED(f);
 	for(i=0; i<ndata; i++)
-		switch(data[i]){
-		case '\t':
-			xprint(cfmt[1][2], "\\t");
-			break;
-		case '\r':
-			xprint(cfmt[1][2], "\\r");
-			break;
-		case '\n':
-			xprint(cfmt[1][2], "\\n");
-			break;
-		case '\b':
-			xprint(cfmt[1][2], "\\b");
-			break;
-		default:
-			if(data[i]>=0x7F || ' '>data[i])
-				xprint(cfmt[2][2], data[i]);
-			else
-				xprint(cfmt[0][2], data[i]);
-			break;
+		onefmtc(data[i]);
+}
+
+void
+fmtr(char *f)
+{
+	int i, w, cw;
+	Rune r;
+	static int nstart;
+
+	USED(f);
+	if(nstart)	
+		xprint("%*c", 3*nstart, ' ');
+	for(i=nstart; i<ndata; )
+		if(data[i] < Runeself)
+			onefmtc(data[i++]);
+		else{
+			w = chartorune(&r, (char *)data+i);
+			if(w == 1 || i + w>nread)
+				onefmtc(data[i++]);
+			else{
+				cw = w;
+				if(i + w>ndata)
+					cw = ndata - i;
+				xprint(rfmt[0][0], r);	
+				xprint("%*c", 3*cw-3, ' ');
+				i += w;
+			}
 		}
+	if(i > ndata)
+		nstart = i - ndata;
+	else
+		nstart = 0;
 }
 
 void
--- /sys/man/1/xd	Thu Aug  1 18:06:17 2013
+++ /sys/man/1/xd	Thu Aug  1 18:06:17 2013
@@ -58,7 +58,7 @@
 .PD
 .PP
 Other options are
-.TP \w'\fL-a\fIstyle\fLXX'u
+.TP \w'\fL-a\fIstyle\fLX'u
 .B -c
 Format as
 .B 1x
@@ -66,17 +66,26 @@
 .SM ASCII
 representations or C escape sequences where possible.
 .TP
+.B -R
+Format as
+.B 1x
+but print
+.B Rune
+representations or C escape sequences where possible.
+.TP
 .BI -a style
 Print file addresses in the given style (and size 4).
 .TP
 .B -u
 (Unbuffered) Flush the output buffer after each 16-byte sequence.
 .TP
-.B -s
-Reverse (swab) the order of bytes in each group of 4 before printing.
+.BI -s " n"
+Reverse (swab) the order of bytes in each group of
+.I n
+before printing.  By default
+.I n
+is 4.  Legal values range from 1 to 8.
 .TP
-.B -S
-Reverse (swab) the order of bytes in each group of 2 before printing.
 .B -r
 Print repeating groups of identical 16-byte sequences as the first group
 followed by an asterisk.