tconvert to 4-byte UTF-8 and 32-bit Rune - plan9port - [fork] Plan 9 from user space
(HTM) git clone git://src.adamsgaard.dk/plan9port
(DIR) Log
(DIR) Files
(DIR) Refs
(DIR) README
(DIR) LICENSE
---
(DIR) commit 0cadb4301d18724e7513d7489cb5bebd262c82f1
(DIR) parent 4dbf255619efac4f0a00e4216d6c999128910df2
(HTM) Author: Russ Cox <russcox@gmail.com>
Date: Fri, 11 Sep 2009 17:03:06 -0400
convert to 4-byte UTF-8 and 32-bit Rune
http://codereview.appspot.com/116075
Diffstat:
M include/fmt.h | 2 +-
M include/utf.h | 7 ++++---
M src/cmd/9term/wind.c | 2 +-
M src/cmd/acme/regx.c | 4 ++--
M src/cmd/sam/cmd.c | 2 +-
M src/cmd/sam/regexp.c | 4 ++--
M src/cmd/sed.c | 4 ++--
M src/cmd/tr.c | 2 +-
M src/cmd/troff/mbwc.c | 4 ++--
M src/cmd/unicode.c | 6 +++---
M src/lib9/fmt/dofmt.c | 11 ++++++-----
M src/lib9/utf/rune.c | 78 +++++++++++++++++++++++--------
M src/libbio/bgetrune.c | 2 +-
M src/libbio/bputrune.c | 2 +-
14 files changed, 86 insertions(+), 44 deletions(-)
---
(DIR) diff --git a/include/fmt.h b/include/fmt.h
t@@ -30,7 +30,7 @@ struct Fmt{
void *farg; /* to make flush a closure */
int nfmt; /* num chars formatted so far */
va_list args; /* args passed to dofmt */
- int r; /* % format Rune */
+ Rune r; /* % format Rune */
int width;
int prec;
unsigned long flags;
(DIR) diff --git a/include/utf.h b/include/utf.h
t@@ -4,14 +4,15 @@
extern "C" {
#endif
-typedef unsigned short Rune; /* 16 bits */
+typedef unsigned int Rune; /* 32 bits */
enum
{
- UTFmax = 3, /* maximum bytes per rune */
+ UTFmax = 4, /* maximum bytes per rune */
Runesync = 0x80, /* cannot represent part of a UTF sequence (<) */
Runeself = 0x80, /* rune and UTF sequences are the same (<) */
- Runeerror = 0xFFFD /* decoding error in UTF */
+ Runeerror = 0xFFFD, /* decoding error in UTF */
+ Runemax = 0x10FFFF /* maximum rune value */
};
/* Edit .+1,/^$/ | cfn $PLAN9/src/lib9/utf/?*.c | grep -v static |grep -v __ */
(DIR) diff --git a/src/cmd/9term/wind.c b/src/cmd/9term/wind.c
t@@ -193,7 +193,7 @@ winctl(void *arg)
Rune *rp, *bp, *up, *kbdr;
uint qh;
int nr, nb, c, wid, i, npart, initial, lastb;
- char *s, *t, part[3];
+ char *s, *t, part[UTFmax];
Window *w;
Mousestate *mp, m;
enum { WKey, WMouse, WMouseread, WCtl, WCwrite, WCread, WWread, NWALT };
(DIR) diff --git a/src/cmd/acme/regx.c b/src/cmd/acme/regx.c
t@@ -488,7 +488,7 @@ bldcclass(void)
exprp++; /* eat '-' */
if((c2 = nextrec()) == ']')
goto Error;
- classp[n+0] = 0xFFFF;
+ classp[n+0] = Runemax;
classp[n+1] = c1;
classp[n+2] = c2;
n += 3;
t@@ -510,7 +510,7 @@ classmatch(int classno, int c, int negate)
p = class[classno];
while(*p){
- if(*p == 0xFFFF){
+ if(*p == Runemax){
if(p[1]<=c && c<=p[2])
return !negate;
p += 3;
(DIR) diff --git a/src/cmd/sam/cmd.c b/src/cmd/sam/cmd.c
t@@ -71,7 +71,7 @@ int
inputc(void)
{
int n, nbuf;
- char buf[3];
+ char buf[UTFmax];
Rune r;
Again:
(DIR) diff --git a/src/cmd/sam/regexp.c b/src/cmd/sam/regexp.c
t@@ -494,7 +494,7 @@ bldcclass(void)
exprp++; /* eat '-' */
if((c2 = nextrec()) == ']')
goto Error;
- classp[n+0] = 0xFFFF;
+ classp[n+0] = Runemax;
classp[n+1] = c1;
classp[n+2] = c2;
n += 3;
t@@ -516,7 +516,7 @@ classmatch(int classno, int c, int negate)
p = class[classno];
while(*p){
- if(*p == 0xFFFF){
+ if(*p == Runemax){
if(p[1]<=c && c<=p[2])
return !negate;
p += 3;
(DIR) diff --git a/src/cmd/sed.c b/src/cmd/sed.c
t@@ -615,7 +615,7 @@ compsub(Rune *rhs, Rune *end)
while ((r = *cp++) != '\0') {
if(r == '\\') {
if (rhs < end)
- *rhs++ = 0xFFFF;
+ *rhs++ = Runemax;
else
return 0;
r = *cp++;
t@@ -1050,7 +1050,7 @@ dosub(Rune *rhsbuf)
sp = place(sp, loc1, loc2);
continue;
}
- if (c == 0xFFFF && (c = *rp++) >= '1' && c < MAXSUB+'0') {
+ if (c == Runemax && (c = *rp++) >= '1' && c < MAXSUB+'0') {
n = c-'0';
if (subexp[n].s.rsp && subexp[n].e.rep) {
sp = place(sp, subexp[n].s.rsp, subexp[n].e.rep);
(DIR) diff --git a/src/cmd/tr.c b/src/cmd/tr.c
t@@ -15,7 +15,7 @@ uchar bits[] = { 1, 2, 4, 8, 16, 32, 64, 128 };
#define CLEARBIT(a,c) ((a)[(c)/8] &= ~bits[(c)&07])
#define BITSET(a,c) ((a)[(c)/8] & bits[(c)&07])
-#define MAXRUNE 0xFFFF
+#define MAXRUNE Runemax
uchar f[(MAXRUNE+1)/8];
uchar t[(MAXRUNE+1)/8];
(DIR) diff --git a/src/cmd/troff/mbwc.c b/src/cmd/troff/mbwc.c
t@@ -152,9 +152,9 @@ wcstombs(char *s, const wchar_t *pwcs, size_t n)
if(p+d <= pe+3) {
*p++ = buf[0];
if(d > 1) {
- *p++ = buf[2];
+ *p++ = buf[1];
if(d > 2)
- *p++ = buf[3];
+ *p++ = buf[2];
}
}
if(c == 0)
(DIR) diff --git a/src/cmd/unicode.c b/src/cmd/unicode.c
t@@ -51,13 +51,13 @@ range(char *argv[])
return "bad range";
}
min = strtoul(q, &q, 16);
- if(min<0 || min>0xFFFF || *q!='-')
+ if(min<0 || min>Runemax || *q!='-')
goto err;
q++;
if(strchr(hex, *q) == 0)
goto err;
max = strtoul(q, &q, 16);
- if(max<0 || max>0xFFFF || max<min || *q!=0)
+ if(max<0 || max>Runemax || max<min || *q!=0)
goto err;
i = 0;
do{
t@@ -111,7 +111,7 @@ chars(char *argv[])
return "bad char";
}
m = strtoul(q, &q, 16);
- if(m<0 || m>0xFFFF || *q!=0)
+ if(m<0 || m>Runemax || *q!=0)
goto err;
Bprint(&bout, "%C", m);
if(!text)
(DIR) diff --git a/src/lib9/fmt/dofmt.c b/src/lib9/fmt/dofmt.c
t@@ -605,12 +605,13 @@ __flagfmt(Fmt *f)
int
__badfmt(Fmt *f)
{
- char x[3];
+ char x[2+UTFmax];
+ int n;
x[0] = '%';
- x[1] = f->r;
- x[2] = '%';
- f->prec = 3;
- __fmtcpy(f, (const void*)x, 3, 3);
+ n = 1 + runetochar(x+1, &f->r);
+ x[n++] = '%';
+ f->prec = n;
+ __fmtcpy(f, (const void*)x, n, n);
return 0;
}
(DIR) diff --git a/src/lib9/utf/rune.c b/src/lib9/utf/rune.c
t@@ -23,16 +23,19 @@ enum
Bit2 = 5,
Bit3 = 4,
Bit4 = 3,
+ Bit5 = 2,
T1 = ((1<<(Bit1+1))-1) ^ 0xFF, /* 0000 0000 */
Tx = ((1<<(Bitx+1))-1) ^ 0xFF, /* 1000 0000 */
T2 = ((1<<(Bit2+1))-1) ^ 0xFF, /* 1100 0000 */
T3 = ((1<<(Bit3+1))-1) ^ 0xFF, /* 1110 0000 */
T4 = ((1<<(Bit4+1))-1) ^ 0xFF, /* 1111 0000 */
+ T5 = ((1<<(Bit5+1))-1) ^ 0xFF, /* 1111 1000 */
- Rune1 = (1<<(Bit1+0*Bitx))-1, /* 0000 0000 0111 1111 */
- Rune2 = (1<<(Bit2+1*Bitx))-1, /* 0000 0111 1111 1111 */
- Rune3 = (1<<(Bit3+2*Bitx))-1, /* 1111 1111 1111 1111 */
+ Rune1 = (1<<(Bit1+0*Bitx))-1, /* 0000 0000 0000 0000 0111 1111 */
+ Rune2 = (1<<(Bit2+1*Bitx))-1, /* 0000 0000 0000 0111 1111 1111 */
+ Rune3 = (1<<(Bit3+2*Bitx))-1, /* 0000 0000 1111 1111 1111 1111 */
+ Rune4 = (1<<(Bit4+3*Bitx))-1, /* 0011 1111 1111 1111 1111 1111 */
Maskx = (1<<Bitx)-1, /* 0011 1111 */
Testx = Maskx ^ 0xFF, /* 1100 0000 */
t@@ -43,7 +46,7 @@ enum
int
chartorune(Rune *rune, char *str)
{
- int c, c1, c2;
+ int c, c1, c2, c3;
long l;
/*
t@@ -89,6 +92,25 @@ chartorune(Rune *rune, char *str)
}
/*
+ * four character sequence
+ * 10000-10FFFF => T4 Tx Tx Tx
+ */
+ if(UTFmax >= 4) {
+ c3 = *(uchar*)(str+3) ^ Tx;
+ if(c3 & Testx)
+ goto bad;
+ if(c < T5) {
+ l = ((((((c << Bitx) | c1) << Bitx) | c2) << Bitx) | c3) & Rune4;
+ if(l <= Rune3)
+ goto bad;
+ if(l > Runemax)
+ goto bad;
+ *rune = l;
+ return 4;
+ }
+ }
+
+ /*
* bad decoding
*/
bad:
t@@ -113,7 +135,7 @@ runetochar(char *str, Rune *rune)
/*
* two character sequence
- * 0080-07FF => T2 Tx
+ * 00080-007FF => T2 Tx
*/
if(c <= Rune2) {
str[0] = T2 | (c >> 1*Bitx);
t@@ -123,12 +145,26 @@ runetochar(char *str, Rune *rune)
/*
* three character sequence
- * 0800-FFFF => T3 Tx Tx
+ * 00800-0FFFF => T3 Tx Tx
*/
- str[0] = T3 | (c >> 2*Bitx);
- str[1] = Tx | ((c >> 1*Bitx) & Maskx);
- str[2] = Tx | (c & Maskx);
- return 3;
+ if(c > Runemax)
+ c = Runeerror;
+ if(c <= Rune3) {
+ str[0] = T3 | (c >> 2*Bitx);
+ str[1] = Tx | ((c >> 1*Bitx) & Maskx);
+ str[2] = Tx | (c & Maskx);
+ return 3;
+ }
+
+ /*
+ * four character sequence
+ * 010000-1FFFFF => T4 Tx Tx Tx
+ */
+ str[0] = T4 | (c >> 3*Bitx);
+ str[1] = Tx | ((c >> 2*Bitx) & Maskx);
+ str[2] = Tx | ((c >> 1*Bitx) & Maskx);
+ str[3] = Tx | (c & Maskx);
+ return 4;
}
int
t@@ -155,7 +191,10 @@ runenlen(Rune *r, int nrune)
if(c <= Rune2)
nb += 2;
else
+ if(c <= Rune3 || c > Runemax)
nb += 3;
+ else
+ nb += 4;
}
return nb;
}
t@@ -165,13 +204,14 @@ fullrune(char *str, int n)
{
int c;
- if(n > 0) {
- c = *(uchar*)str;
- if(c < Tx)
- return 1;
- if(n > 1)
- if(c < T3 || n > 2)
- return 1;
- }
- return 0;
+ if(n <= 0)
+ return 0;
+ c = *(uchar*)str;
+ if(c < Tx)
+ return 1;
+ if(c < T3)
+ return n >= 2;
+ if(UTFmax == 3 || c < T4)
+ return n >= 3;
+ return n >= 4;
}
(DIR) diff --git a/src/libbio/bgetrune.c b/src/libbio/bgetrune.c
t@@ -7,7 +7,7 @@ Bgetrune(Biobuf *bp)
{
int c, i;
Rune rune;
- char str[4];
+ char str[UTFmax];
c = Bgetc(bp);
if(c < Runeself) { /* one char */
(DIR) diff --git a/src/libbio/bputrune.c b/src/libbio/bputrune.c
t@@ -6,7 +6,7 @@ int
Bputrune(Biobuf *bp, long c)
{
Rune rune;
- char str[4];
+ char str[UTFmax];
int n;
rune = c;