Home | History | Annotate | Line # | Download | only in gen
unvis.c revision 1.28
      1 /*	$NetBSD: unvis.c,v 1.28 2005/09/13 01:44:09 christos Exp $	*/
      2 
      3 /*-
      4  * Copyright (c) 1989, 1993
      5  *	The Regents of the University of California.  All rights reserved.
      6  *
      7  * Redistribution and use in source and binary forms, with or without
      8  * modification, are permitted provided that the following conditions
      9  * are met:
     10  * 1. Redistributions of source code must retain the above copyright
     11  *    notice, this list of conditions and the following disclaimer.
     12  * 2. Redistributions in binary form must reproduce the above copyright
     13  *    notice, this list of conditions and the following disclaimer in the
     14  *    documentation and/or other materials provided with the distribution.
     15  * 3. Neither the name of the University nor the names of its contributors
     16  *    may be used to endorse or promote products derived from this software
     17  *    without specific prior written permission.
     18  *
     19  * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
     20  * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
     21  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
     22  * ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
     23  * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
     24  * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
     25  * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
     26  * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
     27  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
     28  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
     29  * SUCH DAMAGE.
     30  */
     31 
     32 #include <sys/cdefs.h>
     33 #if defined(LIBC_SCCS) && !defined(lint)
     34 #if 0
     35 static char sccsid[] = "@(#)unvis.c	8.1 (Berkeley) 6/4/93";
     36 #else
     37 __RCSID("$NetBSD: unvis.c,v 1.28 2005/09/13 01:44:09 christos Exp $");
     38 #endif
     39 #endif /* LIBC_SCCS and not lint */
     40 
     41 #include "namespace.h"
     42 #include <sys/types.h>
     43 
     44 #include <assert.h>
     45 #include <ctype.h>
     46 #include <stdio.h>
     47 #include <vis.h>
     48 
     49 #ifdef __weak_alias
     50 __weak_alias(strunvis,_strunvis)
     51 #endif
     52 
     53 #if !HAVE_VIS
     54 /*
     55  * decode driven by state machine
     56  */
     57 #define	S_GROUND	0	/* haven't seen escape char */
     58 #define	S_START		1	/* start decoding special sequence */
     59 #define	S_META		2	/* metachar started (M) */
     60 #define	S_META1		3	/* metachar more, regular char (-) */
     61 #define	S_CTRL		4	/* control char started (^) */
     62 #define	S_OCTAL2	5	/* octal digit 2 */
     63 #define	S_OCTAL3	6	/* octal digit 3 */
     64 #define	S_HEX1		7	/* hex digit */
     65 #define	S_HEX2		8	/* hex digit 2 */
     66 
     67 #define	isoctal(c)	(((u_char)(c)) >= '0' && ((u_char)(c)) <= '7')
     68 #define xtod(c)		(isdigit(c) ? (c - '0') : ((tolower(c) - 'a') + 10))
     69 
     70 /*
     71  * unvis - decode characters previously encoded by vis
     72  */
     73 int
     74 unvis(cp, c, astate, flag)
     75 	char *cp;
     76 	int c;
     77 	int *astate, flag;
     78 {
     79 	unsigned char uc = (unsigned char)c;
     80 
     81 	_DIAGASSERT(cp != NULL);
     82 	_DIAGASSERT(astate != NULL);
     83 
     84 	if (flag & UNVIS_END) {
     85 		if (*astate == S_OCTAL2 || *astate == S_OCTAL3
     86 		    || *astate == S_HEX2) {
     87 			*astate = S_GROUND;
     88 			return (UNVIS_VALID);
     89 		}
     90 		return (*astate == S_GROUND ? UNVIS_NOCHAR : UNVIS_SYNBAD);
     91 	}
     92 
     93 	switch (*astate) {
     94 
     95 	case S_GROUND:
     96 		*cp = 0;
     97 		if (c == '\\') {
     98 			*astate = S_START;
     99 			return (0);
    100 		}
    101 		if ((flag & VIS_HTTPSTYLE) && c == '%') {
    102 			*astate = S_HEX1;
    103 			return (0);
    104 		}
    105 		*cp = c;
    106 		return (UNVIS_VALID);
    107 
    108 	case S_START:
    109 		switch(c) {
    110 		case '\\':
    111 			*cp = c;
    112 			*astate = S_GROUND;
    113 			return (UNVIS_VALID);
    114 		case '0': case '1': case '2': case '3':
    115 		case '4': case '5': case '6': case '7':
    116 			*cp = (c - '0');
    117 			*astate = S_OCTAL2;
    118 			return (0);
    119 		case 'M':
    120 			*cp = (char)0200;
    121 			*astate = S_META;
    122 			return (0);
    123 		case '^':
    124 			*astate = S_CTRL;
    125 			return (0);
    126 		case 'n':
    127 			*cp = '\n';
    128 			*astate = S_GROUND;
    129 			return (UNVIS_VALID);
    130 		case 'r':
    131 			*cp = '\r';
    132 			*astate = S_GROUND;
    133 			return (UNVIS_VALID);
    134 		case 'b':
    135 			*cp = '\b';
    136 			*astate = S_GROUND;
    137 			return (UNVIS_VALID);
    138 		case 'a':
    139 			*cp = '\007';
    140 			*astate = S_GROUND;
    141 			return (UNVIS_VALID);
    142 		case 'v':
    143 			*cp = '\v';
    144 			*astate = S_GROUND;
    145 			return (UNVIS_VALID);
    146 		case 't':
    147 			*cp = '\t';
    148 			*astate = S_GROUND;
    149 			return (UNVIS_VALID);
    150 		case 'f':
    151 			*cp = '\f';
    152 			*astate = S_GROUND;
    153 			return (UNVIS_VALID);
    154 		case 's':
    155 			*cp = ' ';
    156 			*astate = S_GROUND;
    157 			return (UNVIS_VALID);
    158 		case 'E':
    159 			*cp = '\033';
    160 			*astate = S_GROUND;
    161 			return (UNVIS_VALID);
    162 		case '\n':
    163 			/*
    164 			 * hidden newline
    165 			 */
    166 			*astate = S_GROUND;
    167 			return (UNVIS_NOCHAR);
    168 		case '$':
    169 			/*
    170 			 * hidden marker
    171 			 */
    172 			*astate = S_GROUND;
    173 			return (UNVIS_NOCHAR);
    174 		}
    175 		*astate = S_GROUND;
    176 		return (UNVIS_SYNBAD);
    177 
    178 	case S_META:
    179 		if (c == '-')
    180 			*astate = S_META1;
    181 		else if (c == '^')
    182 			*astate = S_CTRL;
    183 		else {
    184 			*astate = S_GROUND;
    185 			return (UNVIS_SYNBAD);
    186 		}
    187 		return (0);
    188 
    189 	case S_META1:
    190 		*astate = S_GROUND;
    191 		*cp |= c;
    192 		return (UNVIS_VALID);
    193 
    194 	case S_CTRL:
    195 		if (c == '?')
    196 			*cp |= 0177;
    197 		else
    198 			*cp |= c & 037;
    199 		*astate = S_GROUND;
    200 		return (UNVIS_VALID);
    201 
    202 	case S_OCTAL2:	/* second possible octal digit */
    203 		if (isoctal(uc)) {
    204 			/*
    205 			 * yes - and maybe a third
    206 			 */
    207 			*cp = (*cp << 3) + (c - '0');
    208 			*astate = S_OCTAL3;
    209 			return (0);
    210 		}
    211 		/*
    212 		 * no - done with current sequence, push back passed char
    213 		 */
    214 		*astate = S_GROUND;
    215 		return (UNVIS_VALIDPUSH);
    216 
    217 	case S_OCTAL3:	/* third possible octal digit */
    218 		*astate = S_GROUND;
    219 		if (isoctal(uc)) {
    220 			*cp = (*cp << 3) + (c - '0');
    221 			return (UNVIS_VALID);
    222 		}
    223 		/*
    224 		 * we were done, push back passed char
    225 		 */
    226 		return (UNVIS_VALIDPUSH);
    227 
    228 	case S_HEX1:
    229 		if (isxdigit(uc)) {
    230 			*cp = xtod(uc);
    231 			*astate = S_HEX2;
    232 			return (0);
    233 		}
    234 		/*
    235 		 * no - done with current sequence, push back passed char
    236 		 */
    237 		*astate = S_GROUND;
    238 		return (UNVIS_VALIDPUSH);
    239 
    240 	case S_HEX2:
    241 		*astate = S_GROUND;
    242 		if (isxdigit(uc)) {
    243 			*cp = xtod(uc) | (*cp << 4);
    244 			return (UNVIS_VALID);
    245 		}
    246 		return (UNVIS_VALIDPUSH);
    247 
    248 	default:
    249 		/*
    250 		 * decoder in unknown state - (probably uninitialized)
    251 		 */
    252 		*astate = S_GROUND;
    253 		return (UNVIS_SYNBAD);
    254 	}
    255 }
    256 
    257 /*
    258  * strunvis - decode src into dst
    259  *
    260  *	Number of chars decoded into dst is returned, -1 on error.
    261  *	Dst is null terminated.
    262  */
    263 
    264 int
    265 strunvisx(dst, src, flag)
    266 	char *dst;
    267 	const char *src;
    268 	int flag;
    269 {
    270 	char c;
    271 	char *start = dst;
    272 	int state = 0;
    273 
    274 	_DIAGASSERT(src != NULL);
    275 	_DIAGASSERT(dst != NULL);
    276 
    277 	while ((c = *src++) != '\0') {
    278  again:
    279 		switch (unvis(dst, c, &state, flag)) {
    280 		case UNVIS_VALID:
    281 			dst++;
    282 			break;
    283 		case UNVIS_VALIDPUSH:
    284 			dst++;
    285 			goto again;
    286 		case 0:
    287 		case UNVIS_NOCHAR:
    288 			break;
    289 		default:
    290 			return (-1);
    291 		}
    292 	}
    293 	if (unvis(dst, c, &state, UNVIS_END) == UNVIS_VALID)
    294 		dst++;
    295 	*dst = '\0';
    296 	return (dst - start);
    297 }
    298 
    299 int
    300 strunvis(dst, src)
    301 	char *dst;
    302 	const char *src;
    303 {
    304 	return strunvisx(dst, src, 0);
    305 }
    306 #endif
    307