src/url.cc

   1
   2 /*
   3  * $Id: url.cc,v 1.66 1997/11/05 05:29:40 wessels Exp $
   4  *
   5  * DEBUG: section 23    URL Parsing
   6  * AUTHOR: Duane Wessels
   7  *
   8  * SQUID Internet Object Cache  http://squid.nlanr.net/Squid/
   9  * --------------------------------------------------------
  10  *
  11  *  Squid is the result of efforts by numerous individuals from the
  12  *  Internet community.  Development is led by Duane Wessels of the
  13  *  National Laboratory for Applied Network Research and funded by
  14  *  the National Science Foundation.
  15  *
  16  *  This program is free software; you can redistribute it and/or modify
  17  *  it under the terms of the GNU General Public License as published by
  18  *  the Free Software Foundation; either version 2 of the License, or
  19  *  (at your option) any later version.
  20  *
  21  *  This program is distributed in the hope that it will be useful,
  22  *  but WITHOUT ANY WARRANTY; without even the implied warranty of
  23  *  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  24  *  GNU General Public License for more details.
  25  *
  26  *  You should have received a copy of the GNU General Public License
  27  *  along with this program; if not, write to the Free Software
  28  *  Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
  29  *
  30  */
  31
  32 #include "squid.h"
  33
  34 const char *RequestMethodStr[] =
  35 {
  36     "NONE",
  37     "GET",
  38     "POST",
  39     "PUT",
  40     "HEAD",
  41     "CONNECT",
  42     "TRACE",
  43     "PURGE"
  44 };
  45
  46 const char *ProtocolStr[] =
  47 {
  48     "NONE",
  49     "http",
  50     "ftp",
  51     "gopher",
  52     "wais",
  53     "cache_object",
  54     "TOTAL"
  55 };
  56
  57 static int url_acceptable[256];
  58 static const char *const hex = "0123456789abcdef";
  59
  60 /* convert %xx in url string to a character
  61  * Allocate a new string and return a pointer to converted string */
  62
  63 char *
  64 url_convert_hex(char *org_url, int allocate)
  65 {
  66     static char code[] = "00";
  67     char *url = NULL;
  68     char *s = NULL;
  69     char *t = NULL;
  70
  71     url = allocate ? (char *) xstrdup(org_url) : org_url;
  72
  73     if ((int) strlen(url) < 3 || !strchr(url, '%'))
  74         return url;
  75
  76     for (s = t = url; *(s + 2); s++) {
  77         if (*s == '%') {
  78             code[0] = *(++s);
  79             code[1] = *(++s);
  80             *t++ = (char) strtol(code, NULL, 16);
  81         } else {
  82             *t++ = *s;
  83         }
  84     }
  85     do {
  86         *t++ = *s;
  87     } while (*s++);
  88     return url;
  89 }
  90
  91
  92 /* INIT Acceptable table.
  93  * Borrow from libwww2 with Mosaic2.4 Distribution   */
  94 void
  95 urlInitialize(void)
  96 {
  97     unsigned int i;
  98     char *good =
  99     "abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ0123456789./-_$";
 100     debug(23, 5) ("urlInitialize: Initializing...\n");
 101     for (i = 0; i < 256; i++)
 102         url_acceptable[i] = 0;
 103     for (; *good; good++)
 104         url_acceptable[(unsigned int) *good] = 1;
 105 }
 106
 107
 108 /* Encode prohibited char in string */
 109 /* return the pointer to new (allocated) string */
 110 char *
 111 url_escape(const char *url)
 112 {
 113     const char *p;
 114     char *q;
 115     char *tmpline = xcalloc(1, MAX_URL);
 116
 117     q = tmpline;
 118     for (p = url; *p; p++) {
 119         if (url_acceptable[(int) (*p)])
 120             *q++ = *p;
 121         else {
 122             *q++ = '%';         /* Means hex coming */
 123             *q++ = hex[(int) ((*p) >> 4)];
 124             *q++ = hex[(int) ((*p) & 15)];
 125         }
 126     }
 127     *q++ = '\0';
 128     return tmpline;
 129 }
 130
 131 method_t
 132 urlParseMethod(const char *s)
 133 {
 134     if (strcasecmp(s, "GET") == 0) {
 135         return METHOD_GET;
 136     } else if (strcasecmp(s, "POST") == 0) {
 137         return METHOD_POST;
 138     } else if (strcasecmp(s, "PUT") == 0) {
 139         return METHOD_PUT;
 140     } else if (strcasecmp(s, "HEAD") == 0) {
 141         return METHOD_HEAD;
 142     } else if (strcasecmp(s, "CONNECT") == 0) {
 143         return METHOD_CONNECT;
 144     } else if (strcasecmp(s, "TRACE") == 0) {
 145         return METHOD_TRACE;
 146     } else if (strcasecmp(s, "PURGE") == 0) {
 147         return METHOD_PURGE;
 148     }
 149     return METHOD_NONE;
 150 }
 151
 152
 153 protocol_t
 154 urlParseProtocol(const char *s)
 155 {
 156     if (strncasecmp(s, "http", 4) == 0)
 157         return PROTO_HTTP;
 158     if (strncasecmp(s, "ftp", 3) == 0)
 159         return PROTO_FTP;
 160 #ifndef NO_FTP_FOR_FILE
 161     if (strncasecmp(s, "file", 4) == 0)
 162         return PROTO_FTP;
 163 #endif
 164     if (strncasecmp(s, "gopher", 6) == 0)
 165         return PROTO_GOPHER;
 166     if (strncasecmp(s, "wais", 4) == 0)
 167         return PROTO_WAIS;
 168     if (strncasecmp(s, "cache_object", 12) == 0)
 169         return PROTO_CACHEOBJ;
 170     return PROTO_NONE;
 171 }
 172
 173
 174 int
 175 urlDefaultPort(protocol_t p)
 176 {
 177     switch (p) {
 178     case PROTO_HTTP:
 179         return 80;
 180     case PROTO_FTP:
 181         return 21;
 182     case PROTO_GOPHER:
 183         return 70;
 184     case PROTO_WAIS:
 185         return 210;
 186     case PROTO_CACHEOBJ:
 187         return CACHE_HTTP_PORT;
 188     default:
 189         return 0;
 190     }
 191 }
 192
 193 request_t *
 194 urlParse(method_t method, char *url)
 195 {
 196     LOCAL_ARRAY(char, proto, MAX_URL);
 197     LOCAL_ARRAY(char, login, MAX_URL);
 198     LOCAL_ARRAY(char, host, MAX_URL);
 199     LOCAL_ARRAY(char, urlpath, MAX_URL);
 200     request_t *request = NULL;
 201     char *t = NULL;
 202     int port;
 203     protocol_t protocol = PROTO_NONE;
 204     int l;
 205     proto[0] = host[0] = urlpath[0] = login[0] = '\0';
 206
 207     if ((l = strlen(url)) + Config.appendDomainLen > (MAX_URL - 1)) {
 208         /* terminate so it doesn't overflow other buffers */
 209         *(url + (MAX_URL >> 1)) = '\0';
 210         debug(23, 0) ("urlParse: URL too large (%d bytes)\n", l);
 211         return NULL;
 212     }
 213     if (method == METHOD_CONNECT) {
 214         port = CONNECT_PORT;
 215         if (sscanf(url, "%[^:]:%d", host, &port) < 1)
 216             return NULL;
 217     } else {
 218         if (sscanf(url, "%[^:]://%[^/]%s", proto, host, urlpath) < 2)
 219             return NULL;
 220         protocol = urlParseProtocol(proto);
 221         port = urlDefaultPort(protocol);
 222         /* Is there any login informaiton? */
 223         if ((t = strrchr(host, '@'))) {
 224             strcpy(login, host);
 225             t = strrchr(login, '@');
 226             *t = 0;
 227             strcpy(host, t + 1);
 228         }
 229         if ((t = strrchr(host, ':'))) {
 230             *t++ = '\0';
 231             if (*t != '\0')
 232                 port = atoi(t);
 233         }
 234     }
 235     for (t = host; *t; t++)
 236         *t = tolower(*t);
 237     /* remove trailing dots from hostnames */
 238     while ((l = strlen(host)) > 0 && host[--l] == '.')
 239         host[l] = '\0';
 240     if (Config.appendDomain && !strchr(host, '.'))
 241         strncat(host, Config.appendDomain, SQUIDHOSTNAMELEN);
 242     if (port == 0) {
 243         debug(23, 0) ("urlParse: Invalid port == 0\n");
 244         return NULL;
 245     }
 246 #ifdef HARDCODE_DENY_PORTS
 247     /* These ports are filtered in the default squid.conf, but
 248      * maybe someone wants them hardcoded... */
 249     if (port == 7 || port == 9 || port = 19) {
 250         debug(23, 0) ("urlParse: Deny access to port %d\n", port);
 251         return NULL;
 252     }
 253 #endif
 254 #ifdef REMOVE_FTP_TRAILING_SLASHES
 255     /* remove trailing slashes from FTP URLs */
 256     if (protocol == PROTO_FTP) {
 257         t = urlpath + strlen(urlpath);
 258         while (t > urlpath && *(--t) == '/')
 259             *t = '\0';
 260     }
 261 #endif
 262     request = get_free_request_t();
 263     request->method = method;
 264     request->protocol = protocol;
 265     xstrncpy(request->host, host, SQUIDHOSTNAMELEN);
 266     xstrncpy(request->login, login, MAX_LOGIN_SZ);
 267     request->port = (u_short) port;
 268     xstrncpy(request->urlpath, urlpath, MAX_URL);
 269     request->max_age = -1;
 270     request->max_forwards = -1;
 271     return request;
 272 }
 273
 274 char *
 275 urlCanonical(const request_t * request, char *buf)
 276 {
 277     LOCAL_ARRAY(char, urlbuf, MAX_URL);
 278     LOCAL_ARRAY(char, portbuf, 32);
 279     if (buf == NULL)
 280         buf = urlbuf;
 281     switch (request->method) {
 282     case METHOD_CONNECT:
 283         snprintf(buf, MAX_URL, "%s:%d", request->host, request->port);
 284         break;
 285     default:
 286         portbuf[0] = '\0';
 287         if (request->port != urlDefaultPort(request->protocol))
 288             snprintf(portbuf, 32, ":%d", request->port);
 289         snprintf(buf, MAX_URL, "%s://%s%s%s%s%s",
 290             ProtocolStr[request->protocol],
 291             request->login,
 292             *request->login ? "@" : null_string,
 293             request->host,
 294             portbuf,
 295             request->urlpath);
 296         break;
 297     }
 298     return buf;
 299 }
 300
 301 char *
 302 urlCanonicalClean(const request_t * request)
 303 {
 304     LOCAL_ARRAY(char, buf, MAX_URL);
 305     LOCAL_ARRAY(char, portbuf, 32);
 306     char *t;
 307     switch (request->method) {
 308     case METHOD_CONNECT:
 309         snprintf(buf, MAX_URL, "%s:%d", request->host, request->port);
 310         break;
 311     default:
 312         portbuf[0] = '\0';
 313         if (request->port != urlDefaultPort(request->protocol))
 314             snprintf(portbuf, 32, ":%d", request->port);
 315         snprintf(buf, MAX_URL, "%s://%s%s%s",
 316             ProtocolStr[request->protocol],
 317             request->host,
 318             portbuf,
 319             request->urlpath);
 320         if ((t = strchr(buf, '?')))
 321             *t = '\0';
 322         break;
 323     }
 324     return buf;
 325 }
 326
 327 char *
 328 urlClean(char *dirty)
 329 {
 330     char *clean;
 331     request_t *r = urlParse(METHOD_GET, dirty);
 332     if (r == NULL)
 333         return dirty;
 334     clean = urlCanonicalClean(r);
 335     put_free_request_t(r);
 336     return clean;
 337 }
 338
 339
 340 request_t *
 341 requestLink(request_t * request)
 342 {
 343     request->link_count++;
 344     return request;
 345 }
 346
 347 void
 348 requestUnlink(request_t * request)
 349 {
 350     if (request == NULL)
 351         return;
 352     request->link_count--;
 353     if (request->link_count)
 354         return;
 355     safe_free(request->headers);
 356     put_free_request_t(request);
 357 }
 358
 359 int
 360 matchDomainName(const char *domain, const char *host)
 361 {
 362     int offset;
 363     if ((offset = strlen(host) - strlen(domain)) < 0)
 364         return 0;               /* host too short */
 365     if (strcasecmp(domain, host + offset) != 0)
 366         return 0;               /* no match at all */
 367     if (*domain == '.')
 368         return 1;
 369     if (*(host + offset - 1) == '.')
 370         return 1;
 371     if (offset == 0)
 372         return 1;
 373     return 0;
 374 }
 375
 376 int
 377 urlCheckRequest(const request_t * r)
 378 {
 379     int rc = 0;
 380     if (r->method == METHOD_CONNECT)
 381         return 1;
 382     if (r->method == METHOD_TRACE)
 383         return 1;
 384     if (r->method == METHOD_PURGE)
 385         return 1;
 386     switch (r->protocol) {
 387     case PROTO_HTTP:
 388     case PROTO_CACHEOBJ:
 389         rc = 1;
 390         break;
 391     case PROTO_FTP:
 392     case PROTO_GOPHER:
 393     case PROTO_WAIS:
 394         if (r->method == METHOD_GET)
 395             rc = 1;
 396         break;
 397     default:
 398         break;
 399     }
 400     return rc;
 401 }