mdw@git.distorted.org.uk Git - mLib/blob - str.c

   1 /* -*-c-*-
   2  *
   3  * $Id: str.c,v 1.4 2000/10/08 09:43:34 mdw Exp $
   4  *
   5  * Functions for hacking with strings
   6  *
   7  * (c) 1999 Straylight/Edgeware
   8  */
   9
  10 /*----- Licensing notice --------------------------------------------------*
  11  *
  12  * This file is part of the mLib utilities library.
  13  *
  14  * mLib is free software; you can redistribute it and/or modify
  15  * it under the terms of the GNU Library General Public License as
  16  * published by the Free Software Foundation; either version 2 of the
  17  * License, or (at your option) any later version.
  18  *
  19  * mLib is distributed in the hope that it will be useful,
  20  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  21  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  22  * GNU Library General Public License for more details.
  23  *
  24  * You should have received a copy of the GNU Library General Public
  25  * License along with mLib; if not, write to the Free
  26  * Software Foundation, Inc., 59 Temple Place - Suite 330, Boston,
  27  * MA 02111-1307, USA.
  28  */
  29
  30 /*----- Revision history --------------------------------------------------*
  31  *
  32  * $Log: str.c,v $
  33  * Revision 1.4  2000/10/08 09:43:34  mdw
  34  * New quoted string handling and simple pattern matching.
  35  *
  36  * Revision 1.3  1999/12/22 15:41:14  mdw
  37  * Skip past trailing whitespace in str_getword.
  38  *
  39  * Revision 1.2  1999/05/26 20:52:57  mdw
  40  * Add new `rest' argument for `str_split'.
  41  *
  42  * Revision 1.1  1999/05/17 20:37:01  mdw
  43  * Some trivial string hacks.
  44  *
  45  */
  46
  47 /*----- Header files ------------------------------------------------------*/
  48
  49 #include <ctype.h>
  50 #include <stdio.h>
  51 #include <stdlib.h>
  52 #include <string.h>
  53
  54 #include "str.h"
  55
  56 /*----- Main code ---------------------------------------------------------*/
  57
  58 /* --- @str_qword@ --- *
  59  *
  60  * Arguments:   @char **pp@ = address of pointer into string
  61  *              @unsigned f@ = various flags
  62  *
  63  * Returns:     Pointer to the next space-separated possibly-quoted word from
  64  *              the string, or null.
  65  *
  66  * Use:         Fetches the next word from a string.  If the flag
  67  *              @STRF_QUOTE@ is set, the `\' character acts as an escape, and
  68  *              single and double quotes protect whitespace.
  69  */
  70
  71 #define STRF_QUOTE 1u
  72
  73 char *str_qword(char **pp, unsigned f)
  74 {
  75   char *p = *pp, *q, *qq;
  76   int st = 0, pst = 0;
  77
  78   /* --- Preliminaries --- */
  79
  80   if (!p)
  81     return (0);
  82   while (isspace((unsigned char)*p))
  83     p++;
  84   if (!*p) {
  85     *pp = 0;
  86     return (0);
  87   }
  88
  89   /* --- Main work --- */
  90
  91   for (q = qq = p; *q; q++) {
  92     switch (st) {
  93       case '\\':
  94         *qq++ = *q;
  95         st = pst;
  96         break;
  97       case '\'':
  98       case '\"':
  99         if (*q == st)
 100           st = pst = 0;
 101         else if (*q == '\\')
 102           st = '\\';
 103         else
 104           *qq++ = *q;
 105         break;
 106       default:
 107         if (isspace((unsigned char)*q)) {
 108           do q++; while (*q && isspace((unsigned char)*q));
 109           goto done;
 110         } else if (!(f & STRF_QUOTE))
 111           goto stdchar;
 112         switch (*q) {
 113           case '\\':
 114             st = '\\';
 115             break;
 116           case '\'':
 117           case '\"':
 118             st = pst = *q;
 119             break;
 120           default:
 121           stdchar:
 122             *qq++ = *q;
 123             break;
 124         }
 125     }
 126   }
 127
 128   /* --- Finished --- */
 129
 130 done:
 131   *pp = *q ? q : 0;
 132   *qq++ = 0;
 133   return (p);
 134 }
 135
 136 /* --- @str_qsplit@ --- *
 137  *
 138  * Arguments:   @char *p@ = pointer to string
 139  *              @char *v[]@ = pointer to array to fill in
 140  *              @size_t c@ = count of strings to fill in
 141  *              @char **rest@ = where to store the remainder of the string
 142  *              @unsigned f@ = flags for @str_qword@
 143  *
 144  * Returns:     Number of strings filled in.
 145  *
 146  * Use:         Fills an array with pointers to the individual words of a
 147  *              string.  The string is modified in place to contain zero
 148  *              bytes at the word boundaries, and the words have leading
 149  *              and trailing space stripped off.  No more than @c@ words
 150  *              are read; the actual number is returned as the value of the
 151  *              function.  Unused slots in the array are populated with
 152  *              null bytes.  If there's any string left, the address of the
 153  *              remainder is stored in @rest@ (if it's non-null); otherwise
 154  *              @rest@ is set to a null pointer.
 155  */
 156
 157 size_t str_qsplit(char *p, char *v[], size_t c, char **rest, unsigned f)
 158 {
 159   size_t n = 0;
 160   char *q;
 161
 162   while (c && (q = str_qword(&p, f)) != 0) {
 163     *v++ = q;
 164     c--;
 165     n++;
 166   }
 167   while (c) {
 168     *v++ = 0;
 169     c--;
 170   }
 171   if (rest)
 172     *rest = p;
 173   return (n);
 174 }
 175
 176 /* --- @str_getword@ --- *
 177  *
 178  * Arguments:   @char **pp@ = address of pointer into string
 179  *
 180  * Returns:     Pointer to the next space-separated word from the string,
 181  *              or null.
 182  *
 183  * Use:         Parses off space-separated words from a string.  This is a
 184  *              compatibility veneer over @str_qword@.
 185  */
 186
 187 char *str_getword(char **pp)
 188 {
 189   return (str_qword(pp, 0));
 190 }
 191
 192 /* --- @str_split@ --- *
 193  *
 194  * Arguments:   @char *p@ = pointer to string
 195  *              @char *v[]@ = pointer to array to fill in
 196  *              @size_t c@ = count of strings to fill in
 197  *              @char **rest@ = where to store the remainder of the string
 198  *
 199  * Returns:     Number of strings filled in.
 200  *
 201  * Use:         Fills an array with pointers to the individual words of a
 202  *              string.  This is a compatibility veneer over @str_qsplit@.
 203  */
 204
 205 size_t str_split(char *p, char *v[], size_t c, char **rest)
 206 {
 207   return (str_qsplit(p, v, c, rest, 0));
 208 }
 209
 210 /* --- @str_match@ --- *
 211  *
 212  * Arguments:   @const char *p@ = pointer to pattern string
 213  *              @const char *s@ = string to compare with
 214  *
 215  * Returns:     Nonzero if the pattern matches the string.
 216  *
 217  * Use:         Does simple wildcard matching.  This is quite nasty and more
 218  *              than a little slow.  Supports metacharacters `*', `?' and
 219  *              '['.
 220  */
 221
 222 int str_match(const char *p, const char *s)
 223 {
 224   for (;;) {
 225     char pch = *p++, pche, sch;
 226     int sense;
 227
 228     switch (pch) {
 229       case '?':
 230         if (!*s)
 231           return (0);
 232         s++;
 233         break;
 234       case '*':
 235         if (!*p)
 236           return (1);
 237         while (*s) {
 238           if (str_match(p, s))
 239             return (1);
 240           s++;
 241         }
 242         return (0);
 243       case '[':
 244         if (!*s)
 245           return (0);
 246         sch = *s++;
 247         pch = *p++;
 248         sense = 1;
 249         if (pch == '^' || pch == '!') {
 250           sense = !sense;
 251           pch = *p++;
 252         }
 253         if (pch == ']') {
 254           if (*p == '-' && p[1] && p[1] != ']') {
 255             pche = p[1];
 256             p += 2;
 257             if (pch <= sch && sch <= pche)
 258               goto class_match;
 259           } else if (pch == sch)
 260             goto class_match;
 261           pch = *p++;
 262         }
 263         for (;; pch = *p++) {
 264           if (!pch || pch == ']')
 265             goto class_nomatch;
 266           if (*p == '-' && p[1] && p[1] != ']') {
 267             pche = p[1];
 268             p += 2;
 269             if (pch <= sch && sch <= pche)
 270               goto class_match;
 271           } else if (pch == sch)
 272             goto class_match;
 273         }
 274       class_match:
 275         if (!sense)
 276           return (0);
 277         for (;;) {
 278           pch = *p++;
 279           if (!pch)
 280             return (0);
 281           if (pch == ']')
 282             break;
 283           if (*p == '-' && p[1] && p[1] != ']')
 284             p += 2;
 285         }
 286         break;
 287       class_nomatch:
 288         if (sense)
 289           return (0);
 290         break;
 291       case '\\':
 292         pch = *p++;
 293       default:
 294         if (pch != *s)
 295           return (0);
 296         if (!pch)
 297           return (1);
 298         s++;
 299         break;
 300     }
 301   }
 302 }
 303
 304 /* --- @str_sanitize@ --- *
 305  *
 306  * Arguments:   @char *d@ = destination buffer
 307  *              @const char *p@ = pointer to source string
 308  *              @size_t sz@ = size of destination buffer
 309  *
 310  * Returns:     ---
 311  *
 312  * Use:         Writes a string into a buffer, being careful not to overflow
 313  *              the buffer, to null terminate the result, and to prevent
 314  *              nasty nonprintable characters ending up in the buffer.
 315  */
 316
 317 void str_sanitize(char *d, const char *p, size_t sz)
 318 {
 319   if (!sz)
 320     return;
 321   sz--;
 322   while (*p && sz) {
 323     int ch = *p++;
 324     if (!isgraph((unsigned char)ch))
 325       ch = '_';
 326     *d++ = ch;
 327     sz--;
 328   }
 329   *d++ = 0;
 330 }
 331
 332 /*----- That's all, folks -------------------------------------------------*/