fmt.c

在x86平台上运行不可信任代码的sandbox。

/*
 * The authors of this software are Rob Pike and Ken Thompson,
 * with contributions from Mike Burrows and Sean Dorward.
 *
 *     Copyright (c) 2002-2006 by Lucent Technologies.
 *     Portions Copyright (c) 2004 Google Inc.
 * 
 * Permission to use, copy, modify, and distribute this software for any
 * purpose without fee is hereby granted, provided that this entire notice
 * is included in all copies of any software which is or includes a copy
 * or modification of this software and in all copies of the supporting
 * documentation for such software.
 * THIS SOFTWARE IS BEING PROVIDED "AS IS", WITHOUT ANY EXPRESS OR IMPLIED
 * WARRANTY.  IN PARTICULAR, NEITHER THE AUTHORS NOR LUCENT TECHNOLOGIES 
 * NOR GOOGLE INC MAKE ANY REPRESENTATION OR WARRANTY OF ANY KIND CONCERNING 
 * THE MERCHANTABILITY OF THIS SOFTWARE OR ITS FITNESS FOR ANY PARTICULAR PURPOSE.
 */
#include <stdio.h>
#include <math.h>
#include <float.h>
#include <string.h>
#include <stdlib.h>
#include <errno.h>
#include <stdarg.h>
#include <assert.h>
#include <ctype.h>
#include <unistd.h>
#include "u.h"
#include "utf.h"
#include "fmt.h"

#define PLAN9PORT /* Get Plan 9 verbs */

/*
 * compiler directive on Plan 9
 */
#ifndef USED
#define USED(x) if(x);else
#endif

/*
 * nil cannot be ((void*)0) on ANSI C,
 * because it is used for function pointers
 */
#undef	nil
#define	nil	0

#undef	nelem
#define	nelem(x)	(sizeof (x)/sizeof (x)[0])

/* Copyright (c) 2002-2006 Lucent Technologies; see LICENSE */

/*
 * dofmt -- format to a buffer
 * the number of characters formatted is returned,
 * or -1 if there was an error.
 * if the buffer is ever filled, flush is called.
 * it should reset the buffer and return whether formatting should continue.
 */

typedef int (*Fmts)(Fmt*);

typedef struct Quoteinfo Quoteinfo;
struct Quoteinfo
{
	int	quoted;		/* if set, string must be quoted */
	int	nrunesin;	/* number of input runes that can be accepted */
	int	nbytesin;	/* number of input bytes that can be accepted */
	int	nrunesout;	/* number of runes that will be generated */
	int	nbytesout;	/* number of bytes that will be generated */
};

/* Edit .+1,/^$/ |cfn |grep -v static | grep __ */
double       __Inf(int sign);
double       __NaN(void);
int          __badfmt(Fmt *f);
int          __charfmt(Fmt *f);
int          __countfmt(Fmt *f);
int          __efgfmt(Fmt *fmt);
int          __errfmt(Fmt *f);
int          __flagfmt(Fmt *f);
int          __fmtFdFlush(Fmt *f);
int          __fmtcpy(Fmt *f, const void *vm, int n, int sz);
void*        __fmtdispatch(Fmt *f, void *fmt, int isrunes);
void *       __fmtflush(Fmt *f, void *t, int len);
void         __fmtlock(void);
int          __fmtpad(Fmt *f, int n);
double       __fmtpow10(int n);
int          __fmtrcpy(Fmt *f, const void *vm, int n);
void         __fmtunlock(void);
int          __ifmt(Fmt *f);
int          __isInf(double d, int sign);
int          __isNaN(double d);
int          __needsep(int*, char**);
int          __needsquotes(char *s, int *quotelenp);
int          __percentfmt(Fmt *f);
void         __quotesetup(char *s, Rune *r, int nin, int nout, Quoteinfo *q, int sharp, int runesout);
int          __quotestrfmt(int runesin, Fmt *f);
int          __rfmtpad(Fmt *f, int n);
int          __runefmt(Fmt *f);
int          __runeneedsquotes(Rune *r, int *quotelenp);
int          __runesfmt(Fmt *f);
int          __strfmt(Fmt *f);

#define FMTCHAR(f, t, s, c)\
	do{\
	if(t + 1 > (char*)s){\
		t = (char*)__fmtflush(f, t, 1);\
		if(t != nil)\
			s = (char*)f->stop;\
		else\
			return -1;\
	}\
	*t++ = c;\
	}while(0)

#define FMTRCHAR(f, t, s, c)\
	do{\
	if(t + 1 > (Rune*)s){\
		t = (Rune*)__fmtflush(f, t, sizeof(Rune));\
		if(t != nil)\
			s = (Rune*)f->stop;\
		else\
			return -1;\
	}\
	*t++ = c;\
	}while(0)

#define FMTRUNE(f, t, s, r)\
	do{\
	Rune _rune;\
	int _runelen;\
	if(t + UTFmax > (char*)s && t + (_runelen = runelen(r)) > (char*)s){\
		t = (char*)__fmtflush(f, t, _runelen);\
		if(t != nil)\
			s = (char*)f->stop;\
		else\
			return -1;\
	}\
	if(r < Runeself)\
		*t++ = r;\
	else{\
		_rune = r;\
		t += runetochar(t, &_rune);\
	}\
	}while(0)

#ifdef va_copy
#	define VA_COPY(a,b) va_copy(a,b)
#	define VA_END(a) va_end(a)
#else
#	define VA_COPY(a,b) (a) = (b)
#	define VA_END(a)
#endif


/* ---------- end preamble -------- */

/* -------------- charstod.c --------------- */
/* Copyright (c) 2002-2006 Lucent Technologies; see LICENSE */
// #include <stdarg.h>
// #include <string.h>
// #include "plan9.h"
// #include "fmt.h"
// #include "fmtdef.h"

/*
 * Reads a floating-point number by interpreting successive characters
 * returned by (*f)(vp).  The last call it makes to f terminates the
 * scan, so is not a character in the number.  It may therefore be
 * necessary to back up the input stream up one byte after calling charstod.
 */

double
fmtcharstod(int(*f)(void*), void *vp)
{
	double num, dem;
	int neg, eneg, dig, exp, c;

	num = 0;
	neg = 0;
	dig = 0;
	exp = 0;
	eneg = 0;

	c = (*f)(vp);
	while(c == ' ' || c == '\t')
		c = (*f)(vp);
	if(c == '-' || c == '+'){
		if(c == '-')
			neg = 1;
		c = (*f)(vp);
	}
	while(c >= '0' && c <= '9'){
		num = num*10 + c-'0';
		c = (*f)(vp);
	}
	if(c == '.')
		c = (*f)(vp);
	while(c >= '0' && c <= '9'){
		num = num*10 + c-'0';
		dig++;
		c = (*f)(vp);
	}
	if(c == 'e' || c == 'E'){
		c = (*f)(vp);
		if(c == '-' || c == '+'){
			if(c == '-'){
				dig = -dig;
				eneg = 1;
			}
			c = (*f)(vp);
		}
		while(c >= '0' && c <= '9'){
			exp = exp*10 + c-'0';
			c = (*f)(vp);
		}
	}
	exp -= dig;
	if(exp < 0){
		exp = -exp;
		eneg = !eneg;
	}
	dem = __fmtpow10(exp);
	if(eneg)
		num /= dem;
	else
		num *= dem;
	if(neg)
		return -num;
	return num;
}
/* -------------- dofmt.c --------------- */
/* Copyright (c) 2002-2006 Lucent Technologies; see LICENSE */
/* Copyright (c) 2004 Google Inc.; see LICENSE */

// #include <stdarg.h>
// #include <string.h>
// #include "plan9.h"
// #include "fmt.h"
// #include "fmtdef.h"

/* format the output into f->to and return the number of characters fmted  */
int
dofmt(Fmt *f, char *fmt)
{
	Rune rune, *rt, *rs;
	int r;
	char *t, *s;
	int n, nfmt;

	nfmt = f->nfmt;
	for(;;){
		if(f->runes){
			rt = (Rune*)f->to;
			rs = (Rune*)f->stop;
			while((r = *(uchar*)fmt) && r != '%'){
				if(r < Runeself)
					fmt++;
				else{
					fmt += chartorune(&rune, fmt);
					r = rune;
				}
				FMTRCHAR(f, rt, rs, r);
			}
			fmt++;
			f->nfmt += rt - (Rune *)f->to;
			f->to = rt;
			if(!r)
				return f->nfmt - nfmt;
			f->stop = rs;
		}else{
			t = (char*)f->to;
			s = (char*)f->stop;
			while((r = *(uchar*)fmt) && r != '%'){
				if(r < Runeself){
					FMTCHAR(f, t, s, r);
					fmt++;
				}else{
					n = chartorune(&rune, fmt);
					if(t + n > s){
						t = (char*)__fmtflush(f, t, n);
						if(t != nil)
							s = (char*)f->stop;
						else
							return -1;
					}
					while(n--)
						*t++ = *fmt++;
				}
			}
			fmt++;
			f->nfmt += t - (char *)f->to;
			f->to = t;
			if(!r)
				return f->nfmt - nfmt;
			f->stop = s;
		}

		fmt = (char*)__fmtdispatch(f, fmt, 0);
		if(fmt == nil)
			return -1;
	}
}

void *
__fmtflush(Fmt *f, void *t, int len)
{
	if(f->runes)
		f->nfmt += (Rune*)t - (Rune*)f->to;
	else
		f->nfmt += (char*)t - (char *)f->to;
	f->to = t;
	if(f->flush == 0 || (*f->flush)(f) == 0 || (char*)f->to + len > (char*)f->stop){
		f->stop = f->to;
		return nil;
	}
	return f->to;
}

/*
 * put a formatted block of memory sz bytes long of n runes into the output buffer,
 * left/right justified in a field of at least f->width characters (if FmtWidth is set)
 */
int
__fmtpad(Fmt *f, int n)
{
	char *t, *s;
	int i;

	t = (char*)f->to;
	s = (char*)f->stop;
	for(i = 0; i < n; i++)
		FMTCHAR(f, t, s, ' ');
	f->nfmt += t - (char *)f->to;
	f->to = t;
	return 0;
}

int
__rfmtpad(Fmt *f, int n)
{
	Rune *t, *s;
	int i;

	t = (Rune*)f->to;
	s = (Rune*)f->stop;
	for(i = 0; i < n; i++)
		FMTRCHAR(f, t, s, ' ');
	f->nfmt += t - (Rune *)f->to;
	f->to = t;
	return 0;
}

int
__fmtcpy(Fmt *f, const void *vm, int n, int sz)
{
	Rune *rt, *rs, r;
	char *t, *s, *m, *me;
	ulong fl;
	int nc, w;

	m = (char*)vm;
	me = m + sz;
	fl = f->flags;
	w = 0;
	if(fl & FmtWidth)
		w = f->width;
	if((fl & FmtPrec) && n > f->prec)
		n = f->prec;
	if(f->runes){
		if(!(fl & FmtLeft) && __rfmtpad(f, w - n) < 0)
			return -1;
		rt = (Rune*)f->to;
		rs = (Rune*)f->stop;
		for(nc = n; nc > 0; nc--){
			r = *(uchar*)m;
			if(r < Runeself)
				m++;
			else if((me - m) >= UTFmax || fullrune(m, me-m))
				m += chartorune(&r, m);
			else
				break;
			FMTRCHAR(f, rt, rs, r);
		}
		f->nfmt += rt - (Rune *)f->to;
		f->to = rt;
		if(fl & FmtLeft && __rfmtpad(f, w - n) < 0)
			return -1;
	}else{
		if(!(fl & FmtLeft) && __fmtpad(f, w - n) < 0)
			return -1;
		t = (char*)f->to;
		s = (char*)f->stop;
		for(nc = n; nc > 0; nc--){
			r = *(uchar*)m;
			if(r < Runeself)
				m++;
			else if((me - m) >= UTFmax || fullrune(m, me-m))
				m += chartorune(&r, m);
			else
				break;
			FMTRUNE(f, t, s, r);
		}
		f->nfmt += t - (char *)f->to;
		f->to = t;
		if(fl & FmtLeft && __fmtpad(f, w - n) < 0)
			return -1;
	}
	return 0;
}

int
__fmtrcpy(Fmt *f, const void *vm, int n)
{
	Rune r, *m, *me, *rt, *rs;
	char *t, *s;
	ulong fl;
	int w;

	m = (Rune*)vm;
	fl = f->flags;
	w = 0;
	if(fl & FmtWidth)
		w = f->width;
	if((fl & FmtPrec) && n > f->prec)
		n = f->prec;
	if(f->runes){
		if(!(fl & FmtLeft) && __rfmtpad(f, w - n) < 0)
			return -1;
		rt = (Rune*)f->to;
		rs = (Rune*)f->stop;
		for(me = m + n; m < me; m++)
			FMTRCHAR(f, rt, rs, *m);
		f->nfmt += rt - (Rune *)f->to;
		f->to = rt;
		if(fl & FmtLeft && __rfmtpad(f, w - n) < 0)
			return -1;
	}else{
		if(!(fl & FmtLeft) && __fmtpad(f, w - n) < 0)
			return -1;
		t = (char*)f->to;
		s = (char*)f->stop;
		for(me = m + n; m < me; m++){
			r = *m;
			FMTRUNE(f, t, s, r);
		}
		f->nfmt += t - (char *)f->to;
		f->to = t;
		if(fl & FmtLeft && __fmtpad(f, w - n) < 0)
			return -1;
	}
	return 0;
}

/* fmt out one character */
int
__charfmt(Fmt *f)
{
	char x[1];

	x[0] = va_arg(f->args, int);
	f->prec = 1;
	return __fmtcpy(f, (const char*)x, 1, 1);
}

/* fmt out one rune */
int
__runefmt(Fmt *f)
{
	Rune x[1];

	x[0] = va_arg(f->args, int);
	return __fmtrcpy(f, (const void*)x, 1);
}

/* public helper routine: fmt out a null terminated string already in hand */
int
fmtstrcpy(Fmt *f, char *s)
{
	int i, j;

	if(!s)
		return __fmtcpy(f, "<nil>", 5, 5);
	/* if precision is specified, make sure we don't wander off the end */
	if(f->flags & FmtPrec){
#ifdef PLAN9PORT
		Rune r;
		i = 0;
		for(j=0; j<f->prec && s[i]; j++)
			i += chartorune(&r, s+i);
#else
		/* ANSI requires precision in bytes, not Runes */
		for(i=0; i<f->prec; i++)
			if(s[i] == 0)
				break;
		j = utfnlen(s, i);	/* won't print partial at end */
#endif
		return __fmtcpy(f, s, j, i);
	}
	return __fmtcpy(f, s, utflen(s), strlen(s));
}

/* fmt out a null terminated utf string */
int
__strfmt(Fmt *f)
{
	char *s;

	s = va_arg(f->args, char *);
	return fmtstrcpy(f, s);
}

/* public helper routine: fmt out a null terminated rune string already in hand */
int
fmtrunestrcpy(Fmt *f, Rune *s)
{
	Rune *e;
	int n, p;

	if(!s)
		return __fmtcpy(f, "<nil>", 5, 5);
	/* if precision is specified, make sure we don't wander off the end */
	if(f->flags & FmtPrec){
		p = f->prec;
		for(n = 0; n < p; n++)
			if(s[n] == 0)
				break;
	}else{
		for(e = s; *e; e++)
			;
		n = e - s;
	}
	return __fmtrcpy(f, s, n);
}

/* fmt out a null terminated rune string */
int
__runesfmt(Fmt *f)
{
	Rune *s;

	s = va_arg(f->args, Rune *);
	return fmtrunestrcpy(f, s);
}

/* fmt a % */
int
__percentfmt(Fmt *f)
{
	Rune x[1];

	x[0] = f->r;
	f->prec = 1;
	return __fmtrcpy(f, (const void*)x, 1);
}

/* fmt an integer */
int
__ifmt(Fmt *f)
{
	char buf[140], *p, *conv;
	/* 140: for 64 bits of binary + 3-byte sep every 4 digits */
	uvlong vu;
	ulong u;
	int neg, base, i, n, fl, w, isv;
	int ndig, len, excess, bytelen;
	char *grouping;
	char *thousands;

	neg = 0;
	fl = f->flags;
	isv = 0;
	vu = 0;
	u = 0;
#ifndef PLAN9PORT
	/*
	 * Unsigned verbs for ANSI C
	 */
	switch(f->r){
	case 'o':
	case 'p':
	case 'u':
	case 'x':
	case 'X':
		fl |= FmtUnsigned;
		fl &= ~(FmtSign|FmtSpace);
		break;
	}
#endif
	if(f->r == 'p'){
		u = (ulong)va_arg(f->args, void*);
		f->r = 'x';
		fl |= FmtUnsigned;
	}else if(fl & FmtVLong){
		isv = 1;
		if(fl & FmtUnsigned)
			vu = va_arg(f->args, uvlong);
		else
			vu = va_arg(f->args, vlong);
	}else if(fl & FmtLong){
		if(fl & FmtUnsigned)
			u = va_arg(f->args, ulong);
		else
			u = va_arg(f->args, long);
	}else if(fl & FmtByte){
		if(fl & FmtUnsigned)
			u = (uchar)va_arg(f->args, int);
		else
			u = (char)va_arg(f->args, int);
	}else if(fl & FmtShort){
		if(fl & FmtUnsigned)
			u = (ushort)va_arg(f->args, int);
		else
			u = (short)va_arg(f->args, int);
	}else{
		if(fl & FmtUnsigned)
			u = va_arg(f->args, uint);
		else
			u = va_arg(f->args, int);
	}
	conv = "0123456789abcdef";
	grouping = "\4";	/* for hex, octal etc. (undefined by spec but nice) */
	thousands = f->thousands;
	switch(f->r){
	case 'd':
	case 'i':
	case 'u':
		base = 10;
		grouping = f->grouping;
		break;
	case 'X':
		conv = "0123456789ABCDEF";
		/* fall through */
	case 'x':
		base = 16;
		thousands = ":";
		break;
	case 'b':
		base = 2;
		thousands = ":";
		break;
	case 'o':
		base = 8;
		break;
	default:
		return -1;
	}
	if(!(fl & FmtUnsigned)){
		if(isv && (vlong)vu < 0){
			vu = -(vlong)vu;
			neg = 1;
		}else if(!isv && (long)u < 0){
			u = -(long)u;
			neg = 1;
		}
	}
	p = buf + sizeof buf - 1;
	n = 0;	/* in runes */
	excess = 0;	/* number of bytes > number runes */
	ndig = 0;
	len = utflen(thousands);
	bytelen = strlen(thousands);
	if(isv){
		while(vu){
			i = vu % base;
			vu /= base;
			if((fl & FmtComma) && n % 4 == 3){
				*p-- = ',';
				n++;
			}
			if((fl & FmtApost) && __needsep(&ndig, &grouping)){
				n += len;
				excess += bytelen - len;
				p -= bytelen;
				memmove(p+1, thousands, bytelen);
			}
			*p-- = conv[i];
			n++;
		}
	}else{
		while(u){
			i = u % base;
			u /= base;
			if((fl & FmtComma) && n % 4 == 3){
				*p-- = ',';
				n++;
			}
			if((fl & FmtApost) && __needsep(&ndig, &grouping)){
				n += len;