WINGs/string.c

   1 /*
   2  * Until FreeBSD gets their act together;
   3  * http://www.mail-archive.com/freebsd-hackers@freebsd.org/msg69469.html
   4  */
   5 #if defined( FREEBSD )
   6 #   undef _XOPEN_SOURCE
   7 #endif
   8
   9 #include "wconfig.h"
  10
  11 #include <string.h>
  12 #include <stdlib.h>
  13 #include <assert.h>
  14 #include <ctype.h>
  15 #ifdef HAVE_BSD_STRING_H
  16 #include <bsd/string.h>
  17 #endif
  18
  19 #include "WUtil.h"
  20
  21 #define PRC_ALPHA       0
  22 #define PRC_BLANK       1
  23 #define PRC_ESCAPE      2
  24 #define PRC_DQUOTE      3
  25 #define PRC_EOS         4
  26 #define PRC_SQUOTE      5
  27
  28 typedef struct {
  29         short nstate;
  30         short output;
  31 } DFA;
  32
  33 static DFA mtable[9][6] = {
  34         {{3, 1}, {0, 0}, {4, 0}, {1, 0}, {8, 0}, {6, 0}},
  35         {{1, 1}, {1, 1}, {2, 0}, {3, 0}, {5, 0}, {1, 1}},
  36         {{1, 1}, {1, 1}, {1, 1}, {1, 1}, {5, 0}, {1, 1}},
  37         {{3, 1}, {5, 0}, {4, 0}, {1, 0}, {5, 0}, {6, 0}},
  38         {{3, 1}, {3, 1}, {3, 1}, {3, 1}, {5, 0}, {3, 1}},
  39         {{-1, -1}, {0, 0}, {0, 0}, {0, 0}, {0, 0}, {0, 0}},     /* final state */
  40         {{6, 1}, {6, 1}, {7, 0}, {6, 1}, {5, 0}, {3, 0}},
  41         {{6, 1}, {6, 1}, {6, 1}, {6, 1}, {5, 0}, {6, 1}},
  42         {{-1, -1}, {0, 0}, {0, 0}, {0, 0}, {0, 0}, {0, 0}},     /* final state */
  43 };
  44
  45 char *wtokennext(char *word, char **next)
  46 {
  47         char *ptr;
  48         char *ret, *t;
  49         int state, ctype;
  50
  51         t = ret = wmalloc(strlen(word) + 1);
  52         ptr = word;
  53
  54         state = 0;
  55         while (1) {
  56                 if (*ptr == 0)
  57                         ctype = PRC_EOS;
  58                 else if (*ptr == '\\')
  59                         ctype = PRC_ESCAPE;
  60                 else if (*ptr == '"')
  61                         ctype = PRC_DQUOTE;
  62                 else if (*ptr == '\'')
  63                         ctype = PRC_SQUOTE;
  64                 else if (*ptr == ' ' || *ptr == '\t')
  65                         ctype = PRC_BLANK;
  66                 else
  67                         ctype = PRC_ALPHA;
  68
  69                 if (mtable[state][ctype].output) {
  70                         *t = *ptr;
  71                         t++;
  72                         *t = 0;
  73                 }
  74                 state = mtable[state][ctype].nstate;
  75                 ptr++;
  76                 if (mtable[state][0].output < 0) {
  77                         break;
  78                 }
  79         }
  80
  81         if (*ret == 0) {
  82                 wfree(ret);
  83                 ret = NULL;
  84         }
  85
  86         if (ctype == PRC_EOS)
  87                 *next = NULL;
  88         else
  89                 *next = ptr;
  90
  91         return ret;
  92 }
  93
  94 /* separate a string in tokens, taking " and ' into account */
  95 void wtokensplit(char *command, char ***argv, int *argc)
  96 {
  97         char *token, *line;
  98         int count;
  99
 100         count = 0;
 101         line = command;
 102         do {
 103                 token = wtokennext(line, &line);
 104                 if (token) {
 105                         if (count == 0)
 106                                 *argv = wmalloc(sizeof(**argv));
 107                         else
 108                                 *argv = wrealloc(*argv, (count + 1) * sizeof(**argv));
 109                         (*argv)[count++] = token;
 110                 }
 111         } while (token != NULL && line != NULL);
 112
 113         *argc = count;
 114 }
 115
 116 char *wtokenjoin(char **list, int count)
 117 {
 118         int i, j;
 119         char *flat_string, *wspace;
 120
 121         j = 0;
 122         for (i = 0; i < count; i++) {
 123                 if (list[i] != NULL && list[i][0] != 0) {
 124                         j += strlen(list[i]);
 125                         if (strpbrk(list[i], " \t"))
 126                                 j += 2;
 127                 }
 128         }
 129
 130         flat_string = wmalloc(j + count + 1);
 131
 132         for (i = 0; i < count; i++) {
 133                 if (list[i] != NULL && list[i][0] != 0) {
 134                         if (i > 0 &&
 135                             wstrlcat(flat_string, " ", j + count + 1) >= j + count + 1)
 136                                         goto error;
 137
 138                         wspace = strpbrk(list[i], " \t");
 139
 140                         if (wspace &&
 141                             wstrlcat(flat_string, "\"", j + count + 1) >= j + count + 1)
 142                                         goto error;
 143
 144                         if (wstrlcat(flat_string, list[i], j + count + 1) >= j + count + 1)
 145                                 goto error;
 146
 147                         if (wspace &&
 148                             wstrlcat(flat_string, "\"", j + count + 1) >= j + count + 1)
 149                                         goto error;
 150                 }
 151         }
 152
 153         return flat_string;
 154
 155 error:
 156         wfree(flat_string);
 157
 158         return NULL;
 159 }
 160
 161 void wtokenfree(char **tokens, int count)
 162 {
 163         while (count--)
 164                 wfree(tokens[count]);
 165         wfree(tokens);
 166 }
 167
 168 char *wtrimspace(const char *s)
 169 {
 170         const char *t;
 171
 172         if (s == NULL)
 173                 return NULL;
 174
 175         while (isspace(*s) && *s)
 176                 s++;
 177         t = s + strlen(s) - 1;
 178         while (t > s && isspace(*t))
 179                 t--;
 180
 181         return wstrndup(s, t - s + 1);
 182 }
 183
 184 char *wstrdup(const char *str)
 185 {
 186         assert(str != NULL);
 187
 188         return strcpy(wmalloc(strlen(str) + 1), str);
 189 }
 190
 191 char *wstrndup(const char *str, size_t len)
 192 {
 193         char *copy;
 194
 195         assert(str != NULL);
 196
 197         len = WMIN(len, strlen(str));
 198         copy = strncpy(wmalloc(len + 1), str, len);
 199         copy[len] = 0;
 200
 201         return copy;
 202 }
 203
 204 char *wstrconcat(const char *str1, const char *str2)
 205 {
 206         char *str;
 207         size_t slen;
 208
 209         if (!str1 && str2)
 210                 return wstrdup(str2);
 211         else if (str1 && !str2)
 212                 return wstrdup(str1);
 213         else if (!str1 && !str2)
 214                 return NULL;
 215
 216         slen = strlen(str1) + strlen(str2) + 1;
 217         str = wmalloc(slen);
 218         if (wstrlcpy(str, str1, slen) >= slen ||
 219             wstrlcat(str, str2, slen) >= slen) {
 220                 wfree(str);
 221                 return NULL;
 222         }
 223
 224         return str;
 225 }
 226
 227 char *wstrappend(char *dst, const char *src)
 228 {
 229         size_t slen;
 230
 231         if (!src || *src == 0)
 232                 return dst;
 233         else if (!dst)
 234                 return wstrdup(src);
 235
 236         slen = strlen(dst) + strlen(src) + 1;
 237         dst = wrealloc(dst, slen);
 238         strcat(dst, src);
 239
 240         return dst;
 241 }
 242
 243
 244 #ifdef HAVE_STRLCAT
 245 size_t
 246 wstrlcat(char *dst, const char *src, size_t siz)
 247 {
 248         return strlcat(dst, src, siz);
 249 }
 250 #else
 251 /*      $OpenBSD: strlcat.c,v 1.13 2005/08/08 08:05:37 espie Exp $      */
 252
 253 /*
 254  * Copyright (c) 1998 Todd C. Miller <Todd.Miller@courtesan.com>
 255  *
 256  * Permission to use, copy, modify, and distribute this software for any
 257  * purpose with or without fee is hereby granted, provided that the above
 258  * copyright notice and this permission notice appear in all copies.
 259  *
 260  * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES
 261  * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
 262  * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR
 263  * ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
 264  * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
 265  * ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
 266  * OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
 267  */
 268
 269 /*
 270  * Appends src to string dst of size siz (unlike strncat, siz is the
 271  * full size of dst, not space left).  At most siz-1 characters
 272  * will be copied.  Always NUL terminates (unless siz <= strlen(dst)).
 273  * Returns strlen(src) + MIN(siz, strlen(initial dst)).
 274  * If retval >= siz, truncation occurred.
 275  */
 276 size_t
 277 wstrlcat(char *dst, const char *src, size_t siz)
 278 {
 279         char *d = dst;
 280         const char *s = src;
 281         size_t n = siz;
 282         size_t dlen;
 283
 284         /* Find the end of dst and adjust bytes left but don't go past end */
 285         while (n-- != 0 && *d != '\0')
 286                 d++;
 287         dlen = d - dst;
 288         n = siz - dlen;
 289
 290         if (n == 0)
 291                 return(dlen + strlen(s));
 292         while (*s != '\0') {
 293                 if (n != 1) {
 294                         *d++ = *s;
 295                         n--;
 296                 }
 297                 s++;
 298         }
 299         *d = '\0';
 300
 301         return(dlen + (s - src));       /* count does not include NUL */
 302 }
 303 #endif /* HAVE_STRLCAT */
 304
 305 #ifdef HAVE_STRLCPY
 306 size_t
 307 wstrlcpy(char *dst, const char *src, size_t siz)
 308 {
 309         return strlcpy(dst, src, siz);
 310 }
 311 #else
 312
 313 /*      $OpenBSD: strlcpy.c,v 1.11 2006/05/05 15:27:38 millert Exp $    */
 314
 315 /*
 316  * Copyright (c) 1998 Todd C. Miller <Todd.Miller@courtesan.com>
 317  *
 318  * Permission to use, copy, modify, and distribute this software for any
 319  * purpose with or without fee is hereby granted, provided that the above
 320  * copyright notice and this permission notice appear in all copies.
 321  *
 322  * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES
 323  * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
 324  * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR
 325  * ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
 326  * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
 327  * ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
 328  * OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
 329  */
 330
 331 /*
 332  * Copy src to string dst of size siz.  At most siz-1 characters
 333  * will be copied.  Always NUL terminates (unless siz == 0).
 334  * Returns strlen(src); if retval >= siz, truncation occurred.
 335  */
 336 size_t
 337 wstrlcpy(char *dst, const char *src, size_t siz)
 338 {
 339         char *d = dst;
 340         const char *s = src;
 341         size_t n = siz;
 342
 343         /* Copy as many bytes as will fit */
 344         if (n != 0) {
 345                 while (--n != 0) {
 346                         if ((*d++ = *s++) == '\0')
 347                                 break;
 348                 }
 349         }
 350
 351         /* Not enough room in dst, add NUL and traverse rest of src */
 352         if (n == 0) {
 353                 if (siz != 0)
 354                         *d = '\0';              /* NUL-terminate dst */
 355                 while (*s++)
 356                         ;
 357         }
 358
 359         return(s - src - 1);    /* count does not include NUL */
 360 }
 361 #endif /* HAVE_STRLCPY */
 362
 363 /* transform `s' so that the result is safe to pass to the shell as an argument.
 364  * returns a newly allocated string.
 365  * with very heavy inspirations from NetBSD's shquote(3).
 366  */
 367 char *wshellquote(const char *s)
 368 {
 369         char *p, *r, *last, *ret;
 370         size_t slen;
 371         int needs_quoting;
 372
 373         if (!s)
 374                 return NULL;
 375
 376         needs_quoting = !*s;                    /* the empty string does need quoting */
 377
 378         /* do not quote if consists only of the following characters */
 379         for (p = (char *)s; *p && !needs_quoting; p++) {
 380                 needs_quoting = !(isalnum(*p) || (*p == '+') || (*p == '/') ||
 381                                  (*p == '.') || (*p == ',') || (*p == '-'));
 382         }
 383
 384         if (!needs_quoting)
 385                 return wstrdup(s);
 386
 387         for (slen = 0, p = (char *)s; *p; p++)          /* count space needed (worst case) */
 388                 slen += *p == '\'' ? 4 : 1;             /* every single ' becomes ''\' */
 389
 390         slen += 2 /* leading + trailing "'" */ + 1 /* NULL */;
 391
 392         ret = r = wmalloc(slen);
 393         p = (char *)s;
 394         last = p;
 395
 396         if (*p != '\'')                                 /* if string doesn't already begin with "'" */
 397                 *r++ ='\'';                             /* start putting it in quotes */
 398
 399         while (*p) {
 400                 last = p;
 401                 if (*p == '\'') {                       /* turn each ' into ''\' */
 402                         if (p != s)                     /* except if it's the first ', in which case */
 403                                 *r++ = '\'';            /* only escape it */
 404                         *r++ = '\\';
 405                         *r++ = '\'';
 406                         while (*++p && *p == '\'') {    /* keep turning each consecutive 's into \' */
 407                                 *r++ = '\\';
 408                                 *r++ = '\'';
 409                         }
 410                         if (*p)                         /* if more input follows, terminate */
 411                                 *r++ = '\'';            /* what we have so far */
 412                 } else {
 413                         *r++ = *p++;
 414                 }
 415         }
 416
 417         if (*last != '\'')                              /* if the last one isn't already a ' */
 418                 *r++ = '\'';                            /* terminate the whole shebang */
 419
 420         *r = '\0';
 421
 422         return ret;                                     /* technically, we lose (but not leak) a couple of */
 423                                                         /* bytes (twice the number of consecutive 's in the */
 424                                                         /* input or so), but since these are relatively rare */
 425                                                         /* and short-lived strings, not sure if a trip to */
 426                                                         /* wstrdup+wfree worths the gain. */
 427 }