cups/http-support.c

   1 /*
   2  * "$Id: http-support.c 6649 2007-07-11 21:46:42Z mike $"
   3  *
   4  *   HTTP support routines for the Common UNIX Printing System (CUPS) scheduler.
   5  *
   6  *   Copyright 2007 by Apple Inc.
   7  *   Copyright 1997-2007 by Easy Software Products, all rights reserved.
   8  *
   9  *   These coded instructions, statements, and computer programs are the
  10  *   property of Apple Inc. and are protected by Federal copyright
  11  *   law.  Distribution and use rights are outlined in the file "LICENSE.txt"
  12  *   which should have been included with this file.  If this file is
  13  *   file is missing or damaged, see the license at "http://www.cups.org/".
  14  *
  15  *   This file is subject to the Apple OS-Developed Software exception.
  16  *
  17  * Contents:
  18  *
  19  *   httpAssembleURI()    - Assemble a uniform resource identifier from its
  20  *                          components.
  21  *   httpAssembleURIf()   - Assemble a uniform resource identifier from its
  22  *                          components with a formatted resource.
  23  *   httpDecode64()       - Base64-decode a string.
  24  *   httpDecode64_2()     - Base64-decode a string.
  25  *   httpEncode64()       - Base64-encode a string.
  26  *   httpEncode64_2()     - Base64-encode a string.
  27  *   httpGetDateString()  - Get a formatted date/time string from a time value.
  28  *   httpGetDateString2() - Get a formatted date/time string from a time value.
  29  *   httpGetDateTime()    - Get a time value from a formatted date/time string.
  30  *   httpSeparate()       - Separate a Universal Resource Identifier into its
  31  *                          components.
  32  *   httpSeparate2()      - Separate a Universal Resource Identifier into its
  33  *                          components.
  34  *   httpSeparateURI()    - Separate a Universal Resource Identifier into its
  35  *                          components.
  36  *   httpStatus()         - Return a short string describing a HTTP status code.
  37  *   _cups_hstrerror()    - hstrerror() emulation function for Solaris and
  38  *                          others...
  39  *   _httpEncodeURI()     - Percent-encode a HTTP request URI.
  40  *   http_copy_decode()   - Copy and decode a URI.
  41  *   http_copy_encode()   - Copy and encode a URI.
  42  */
  43
  44 /*
  45  * Include necessary headers...
  46  */
  47
  48 #include "debug.h"
  49 #include "globals.h"
  50 #include <stdlib.h>
  51
  52
  53 /*
  54  * Local globals...
  55  */
  56
  57 static const char * const http_days[7] =
  58                         {
  59                           "Sun",
  60                           "Mon",
  61                           "Tue",
  62                           "Wed",
  63                           "Thu",
  64                           "Fri",
  65                           "Sat"
  66                         };
  67 static const char * const http_months[12] =
  68                         {
  69                           "Jan",
  70                           "Feb",
  71                           "Mar",
  72                           "Apr",
  73                           "May",
  74                           "Jun",
  75                           "Jul",
  76                           "Aug",
  77                           "Sep",
  78                           "Oct",
  79                           "Nov",
  80                           "Dec"
  81                         };
  82
  83
  84 /*
  85  * Local functions...
  86  */
  87
  88 static const char       *http_copy_decode(char *dst, const char *src,
  89                                           int dstsize, const char *term,
  90                                           int decode);
  91 static char             *http_copy_encode(char *dst, const char *src,
  92                                           char *dstend, const char *reserved,
  93                                           const char *term, int encode);
  94
  95
  96 /*
  97  * 'httpAssembleURI()' - Assemble a uniform resource identifier from its
  98  *                       components.
  99  *
 100  * This function escapes reserved characters in the URI depending on the
 101  * value of the "encoding" argument.  You should use this function in
 102  * place of traditional string functions whenever you need to create a
 103  * URI string.
 104  *
 105  * @since CUPS 1.2@
 106  */
 107
 108 http_uri_status_t                       /* O - URI status */
 109 httpAssembleURI(
 110     http_uri_coding_t encoding,         /* I - Encoding flags */
 111     char              *uri,             /* I - URI buffer */
 112     int               urilen,           /* I - Size of URI buffer */
 113     const char        *scheme,          /* I - Scheme name */
 114     const char        *username,        /* I - Username */
 115     const char        *host,            /* I - Hostname or address */
 116     int               port,             /* I - Port number */
 117     const char        *resource)        /* I - Resource */
 118 {
 119   char          *ptr,                   /* Pointer into URI buffer */
 120                 *end;                   /* End of URI buffer */
 121
 122
 123  /*
 124   * Range check input...
 125   */
 126
 127   if (!uri || urilen < 1 || !scheme || port < 0)
 128   {
 129     if (uri)
 130       *uri = '\0';
 131
 132     return (HTTP_URI_BAD_ARGUMENTS);
 133   }
 134
 135  /*
 136   * Assemble the URI starting with the scheme...
 137   */
 138
 139   end = uri + urilen - 1;
 140   ptr = http_copy_encode(uri, scheme, end, NULL, NULL, 0);
 141
 142   if (!ptr)
 143     goto assemble_overflow;
 144
 145   if (!strcmp(scheme, "mailto"))
 146   {
 147    /*
 148     * mailto: only has :, no //...
 149     */
 150
 151     if (ptr < end)
 152       *ptr++ = ':';
 153     else
 154       goto assemble_overflow;
 155   }
 156   else
 157   {
 158    /*
 159     * Schemes other than mailto: all have //...
 160     */
 161
 162     if ((ptr + 2) < end)
 163     {
 164       *ptr++ = ':';
 165       *ptr++ = '/';
 166       *ptr++ = '/';
 167     }
 168     else
 169       goto assemble_overflow;
 170   }
 171
 172  /*
 173   * Next the username and hostname, if any...
 174   */
 175
 176   if (host)
 177   {
 178     if (username && *username)
 179     {
 180      /*
 181       * Add username@ first...
 182       */
 183
 184       ptr = http_copy_encode(ptr, username, end, "/?@", NULL,
 185                              encoding & HTTP_URI_CODING_USERNAME);
 186
 187       if (!ptr)
 188         goto assemble_overflow;
 189
 190       if (ptr < end)
 191         *ptr++ = '@';
 192       else
 193         goto assemble_overflow;
 194     }
 195
 196    /*
 197     * Then add the hostname.  Since IPv6 is a particular pain to deal
 198     * with, we have several special cases to deal with...  If we get
 199     * an IPv6 address with brackets around it, assume it is already in
 200     * URI format...
 201     */
 202
 203     if (host[0] != '[' && strchr(host, ':'))
 204     {
 205      /*
 206       * We have an IPv6 address...
 207       */
 208
 209       if (strchr(host, '%'))
 210       {
 211        /*
 212         * We have a link-local address, add "[v1." prefix...
 213         */
 214
 215         if ((ptr + 4) < end)
 216         {
 217           *ptr++ = '[';
 218           *ptr++ = 'v';
 219           *ptr++ = '1';
 220           *ptr++ = '.';
 221         }
 222         else
 223           goto assemble_overflow;
 224       }
 225       else
 226       {
 227        /*
 228         * We have a normal address, add "[" prefix...
 229         */
 230
 231         if (ptr < end)
 232           *ptr++ = '[';
 233         else
 234           goto assemble_overflow;
 235       }
 236
 237      /*
 238       * Copy the rest of the IPv6 address, and terminate with "]".
 239       */
 240
 241       while (ptr < end && *host)
 242       {
 243         if (*host == '%')
 244         {
 245           *ptr++ = '+';                 /* Convert zone separator */
 246           host ++;
 247         }
 248         else
 249           *ptr++ = *host++;
 250       }
 251
 252       if (*host)
 253         goto assemble_overflow;
 254
 255       if (ptr < end)
 256         *ptr++ = ']';
 257       else
 258         goto assemble_overflow;
 259     }
 260     else
 261     {
 262      /*
 263       * Otherwise, just copy the host string...
 264       */
 265
 266       ptr = http_copy_encode(ptr, host, end, NULL, NULL,
 267                              encoding & HTTP_URI_CODING_HOSTNAME);
 268
 269       if (!ptr)
 270         goto assemble_overflow;
 271     }
 272
 273    /*
 274     * Finish things off with the port number...
 275     */
 276
 277     if (port > 0)
 278     {
 279       snprintf(ptr, end - ptr + 1, ":%d", port);
 280       ptr += strlen(ptr);
 281
 282       if (ptr >= end)
 283         goto assemble_overflow;
 284     }
 285   }
 286
 287  /*
 288   * Last but not least, add the resource string...
 289   */
 290
 291   if (resource)
 292   {
 293     char        *query;                 /* Pointer to query string */
 294
 295
 296    /*
 297     * Copy the resource string up to the query string if present...
 298     */
 299
 300     query = strchr(resource, '?');
 301     ptr   = http_copy_encode(ptr, resource, end, NULL, "?",
 302                              encoding & HTTP_URI_CODING_RESOURCE);
 303     if (!ptr)
 304       goto assemble_overflow;
 305
 306     if (query)
 307     {
 308      /*
 309       * Copy query string without encoding...
 310       */
 311
 312       ptr = http_copy_encode(ptr, query, end, NULL, NULL,
 313                              encoding & HTTP_URI_CODING_QUERY);
 314       if (!ptr)
 315         goto assemble_overflow;
 316     }
 317   }
 318   else if (ptr < end)
 319     *ptr++ = '/';
 320   else
 321     goto assemble_overflow;
 322
 323  /*
 324   * Nul-terminate the URI buffer and return with no errors...
 325   */
 326
 327   *ptr = '\0';
 328
 329   return (HTTP_URI_OK);
 330
 331  /*
 332   * Clear the URI string and return an overflow error; I don't usually
 333   * like goto's, but in this case it makes sense...
 334   */
 335
 336   assemble_overflow:
 337
 338   *uri = '\0';
 339   return (HTTP_URI_OVERFLOW);
 340 }
 341
 342
 343 /*
 344  * 'httpAssembleURIf()' - Assemble a uniform resource identifier from its
 345  *                        components with a formatted resource.
 346  *
 347  * This function creates a formatted version of the resource string
 348  * argument "resourcef" and escapes reserved characters in the URI
 349  * depending on the value of the "encoding" argument.  You should use
 350  * this function in place of traditional string functions whenever
 351  * you need to create a URI string.
 352  *
 353  * @since CUPS 1.2@
 354  */
 355
 356 http_uri_status_t                       /* O - URI status */
 357 httpAssembleURIf(
 358     http_uri_coding_t encoding,         /* I - Encoding flags */
 359     char              *uri,             /* I - URI buffer */
 360     int               urilen,           /* I - Size of URI buffer */
 361     const char        *scheme,          /* I - Scheme name */
 362     const char        *username,        /* I - Username */
 363     const char        *host,            /* I - Hostname or address */
 364     int               port,             /* I - Port number */
 365     const char        *resourcef,       /* I - Printf-style resource */
 366     ...)                                /* I - Additional arguments as needed */
 367 {
 368   va_list       ap;                     /* Pointer to additional arguments */
 369   char          resource[1024];         /* Formatted resource string */
 370   int           bytes;                  /* Bytes in formatted string */
 371
 372
 373  /*
 374   * Range check input...
 375   */
 376
 377   if (!uri || urilen < 1 || !scheme || port < 0 || !resourcef)
 378   {
 379     if (uri)
 380       *uri = '\0';
 381
 382     return (HTTP_URI_BAD_ARGUMENTS);
 383   }
 384
 385  /*
 386   * Format the resource string and assemble the URI...
 387   */
 388
 389   va_start(ap, resourcef);
 390   bytes = vsnprintf(resource, sizeof(resource), resourcef, ap);
 391   va_end(ap);
 392
 393   if (bytes >= sizeof(resource))
 394   {
 395     *uri = '\0';
 396     return (HTTP_URI_OVERFLOW);
 397   }
 398   else
 399     return (httpAssembleURI(encoding,  uri, urilen, scheme, username, host,
 400                             port, resource));
 401 }
 402
 403
 404 /*
 405  * 'httpDecode64()' - Base64-decode a string.
 406  *
 407  * This function is deprecated. Use the httpDecode64_2() function instead
 408  * which provides buffer length arguments.
 409  *
 410  * @deprecated@
 411  */
 412
 413 char *                                  /* O - Decoded string */
 414 httpDecode64(char       *out,           /* I - String to write to */
 415              const char *in)            /* I - String to read from */
 416 {
 417   int   outlen;                         /* Output buffer length */
 418
 419
 420  /*
 421   * Use the old maximum buffer size for binary compatibility...
 422   */
 423
 424   outlen = 512;
 425
 426   return (httpDecode64_2(out, &outlen, in));
 427 }
 428
 429
 430 /*
 431  * 'httpDecode64_2()' - Base64-decode a string.
 432  *
 433  * @since CUPS 1.1.21@
 434  */
 435
 436 char *                                  /* O  - Decoded string */
 437 httpDecode64_2(char       *out,         /* I  - String to write to */
 438                int        *outlen,      /* IO - Size of output string */
 439                const char *in)          /* I  - String to read from */
 440 {
 441   int   pos,                            /* Bit position */
 442         base64;                         /* Value of this character */
 443   char  *outptr,                        /* Output pointer */
 444         *outend;                        /* End of output buffer */
 445
 446
 447  /*
 448   * Range check input...
 449   */
 450
 451   if (!out || !outlen || *outlen < 1 || !in)
 452     return (NULL);
 453
 454   if (!*in)
 455   {
 456     *out    = '\0';
 457     *outlen = 0;
 458
 459     return (out);
 460   }
 461
 462  /*
 463   * Convert from base-64 to bytes...
 464   */
 465
 466   for (outptr = out, outend = out + *outlen - 1, pos = 0; *in != '\0'; in ++)
 467   {
 468    /*
 469     * Decode this character into a number from 0 to 63...
 470     */
 471
 472     if (*in >= 'A' && *in <= 'Z')
 473       base64 = *in - 'A';
 474     else if (*in >= 'a' && *in <= 'z')
 475       base64 = *in - 'a' + 26;
 476     else if (*in >= '0' && *in <= '9')
 477       base64 = *in - '0' + 52;
 478     else if (*in == '+')
 479       base64 = 62;
 480     else if (*in == '/')
 481       base64 = 63;
 482     else if (*in == '=')
 483       break;
 484     else
 485       continue;
 486
 487    /*
 488     * Store the result in the appropriate chars...
 489     */
 490
 491     switch (pos)
 492     {
 493       case 0 :
 494           if (outptr < outend)
 495             *outptr = base64 << 2;
 496           pos ++;
 497           break;
 498       case 1 :
 499           if (outptr < outend)
 500             *outptr++ |= (base64 >> 4) & 3;
 501           if (outptr < outend)
 502             *outptr = (base64 << 4) & 255;
 503           pos ++;
 504           break;
 505       case 2 :
 506           if (outptr < outend)
 507             *outptr++ |= (base64 >> 2) & 15;
 508           if (outptr < outend)
 509             *outptr = (base64 << 6) & 255;
 510           pos ++;
 511           break;
 512       case 3 :
 513           if (outptr < outend)
 514             *outptr++ |= base64;
 515           pos = 0;
 516           break;
 517     }
 518   }
 519
 520   *outptr = '\0';
 521
 522  /*
 523   * Return the decoded string and size...
 524   */
 525
 526   *outlen = (int)(outptr - out);
 527
 528   return (out);
 529 }
 530
 531
 532 /*
 533  * 'httpEncode64()' - Base64-encode a string.
 534  *
 535  * This function is deprecated. Use the httpEncode64_2() function instead
 536  * which provides buffer length arguments.
 537  *
 538  * @deprecated@
 539  */
 540
 541 char *                                  /* O - Encoded string */
 542 httpEncode64(char       *out,           /* I - String to write to */
 543              const char *in)            /* I - String to read from */
 544 {
 545   return (httpEncode64_2(out, 512, in, (int)strlen(in)));
 546 }
 547
 548
 549 /*
 550  * 'httpEncode64_2()' - Base64-encode a string.
 551  *
 552  * @since CUPS 1.1.21@
 553  */
 554
 555 char *                                  /* O - Encoded string */
 556 httpEncode64_2(char       *out,         /* I - String to write to */
 557                int        outlen,       /* I - Size of output string */
 558                const char *in,          /* I - String to read from */
 559                int        inlen)        /* I - Size of input string */
 560 {
 561   char          *outptr,                /* Output pointer */
 562                 *outend;                /* End of output buffer */
 563   static const char base64[] =          /* Base64 characters... */
 564                 {
 565                   "ABCDEFGHIJKLMNOPQRSTUVWXYZ"
 566                   "abcdefghijklmnopqrstuvwxyz"
 567                   "0123456789"
 568                   "+/"
 569                 };
 570
 571
 572  /*
 573   * Range check input...
 574   */
 575
 576   if (!out || outlen < 1 || !in)
 577     return (NULL);
 578
 579  /*
 580   * Convert bytes to base-64...
 581   */
 582
 583   for (outptr = out, outend = out + outlen - 1; inlen > 0; in ++, inlen --)
 584   {
 585    /*
 586     * Encode the up to 3 characters as 4 Base64 numbers...
 587     */
 588
 589     if (outptr < outend)
 590       *outptr ++ = base64[(in[0] & 255) >> 2];
 591
 592     if (outptr < outend)
 593     {
 594       if (inlen > 1)
 595         *outptr ++ = base64[(((in[0] & 255) << 4) | ((in[1] & 255) >> 4)) & 63];
 596       else
 597         *outptr ++ = base64[((in[0] & 255) << 4) & 63];
 598     }
 599
 600     in ++;
 601     inlen --;
 602     if (inlen <= 0)
 603     {
 604       if (outptr < outend)
 605         *outptr ++ = '=';
 606       if (outptr < outend)
 607         *outptr ++ = '=';
 608       break;
 609     }
 610
 611     if (outptr < outend)
 612     {
 613       if (inlen > 1)
 614         *outptr ++ = base64[(((in[0] & 255) << 2) | ((in[1] & 255) >> 6)) & 63];
 615       else
 616         *outptr ++ = base64[((in[0] & 255) << 2) & 63];
 617     }
 618
 619     in ++;
 620     inlen --;
 621     if (inlen <= 0)
 622     {
 623       if (outptr < outend)
 624         *outptr ++ = '=';
 625       break;
 626     }
 627
 628     if (outptr < outend)
 629       *outptr ++ = base64[in[0] & 63];
 630   }
 631
 632   *outptr = '\0';
 633
 634  /*
 635   * Return the encoded string...
 636   */
 637
 638   return (out);
 639 }
 640
 641
 642 /*
 643  * 'httpGetDateString()' - Get a formatted date/time string from a time value.
 644  *
 645  * @deprecated@
 646  */
 647
 648 const char *                            /* O - Date/time string */
 649 httpGetDateString(time_t t)             /* I - UNIX time */
 650 {
 651   _cups_globals_t *cg = _cupsGlobals(); /* Pointer to library globals */
 652
 653
 654   return (httpGetDateString2(t, cg->http_date, sizeof(cg->http_date)));
 655 }
 656
 657
 658 /*
 659  * 'httpGetDateString2()' - Get a formatted date/time string from a time value.
 660  *
 661  * @since CUPS 1.2@
 662  */
 663
 664 const char *                            /* O - Date/time string */
 665 httpGetDateString2(time_t t,            /* I - UNIX time */
 666                    char   *s,           /* I - String buffer */
 667                    int    slen)         /* I - Size of string buffer */
 668 {
 669   struct tm     *tdate;                 /* UNIX date/time data */
 670
 671
 672   tdate = gmtime(&t);
 673   snprintf(s, slen, "%s, %02d %s %d %02d:%02d:%02d GMT",
 674            http_days[tdate->tm_wday], tdate->tm_mday,
 675            http_months[tdate->tm_mon], tdate->tm_year + 1900,
 676            tdate->tm_hour, tdate->tm_min, tdate->tm_sec);
 677
 678   return (s);
 679 }
 680
 681
 682 /*
 683  * 'httpGetDateTime()' - Get a time value from a formatted date/time string.
 684  */
 685
 686 time_t                                  /* O - UNIX time */
 687 httpGetDateTime(const char *s)          /* I - Date/time string */
 688 {
 689   int           i;                      /* Looping var */
 690   char          mon[16];                /* Abbreviated month name */
 691   int           day, year;              /* Day of month and year */
 692   int           hour, min, sec;         /* Time */
 693   int           days;                   /* Number of days since 1970 */
 694   static const int normal_days[] =      /* Days to a month, normal years */
 695                 { 0, 31, 59, 90, 120, 151, 181, 212, 243, 273, 304, 334, 365 };
 696   static const int leap_days[] =        /* Days to a month, leap years */
 697                 { 0, 31, 60, 91, 121, 152, 182, 213, 244, 274, 305, 335, 366 };
 698
 699
 700   DEBUG_printf(("httpGetDateTime(s=\"%s\")\n", s));
 701
 702  /*
 703   * Extract the date and time from the formatted string...
 704   */
 705
 706   if (sscanf(s, "%*s%d%15s%d%d:%d:%d", &day, mon, &year, &hour, &min, &sec) < 6)
 707     return (0);
 708
 709   DEBUG_printf(("    day=%d, mon=\"%s\", year=%d, hour=%d, min=%d, sec=%d\n",
 710                 day, mon, year, hour, min, sec));
 711
 712  /*
 713   * Convert the month name to a number from 0 to 11.
 714   */
 715
 716   for (i = 0; i < 12; i ++)
 717     if (!strcasecmp(mon, http_months[i]))
 718       break;
 719
 720   if (i >= 12)
 721     return (0);
 722
 723   DEBUG_printf(("    i=%d\n", i));
 724
 725  /*
 726   * Now convert the date and time to a UNIX time value in seconds since
 727   * 1970.  We can't use mktime() since the timezone may not be UTC but
 728   * the date/time string *is* UTC.
 729   */
 730
 731   if ((year & 3) == 0 && ((year % 100) != 0 || (year % 400) == 0))
 732     days = leap_days[i] + day - 1;
 733   else
 734     days = normal_days[i] + day - 1;
 735
 736   DEBUG_printf(("    days=%d\n", days));
 737
 738   days += (year - 1970) * 365 +         /* 365 days per year (normally) */
 739           ((year - 1) / 4 - 492) -      /* + leap days */
 740           ((year - 1) / 100 - 19) +     /* - 100 year days */
 741           ((year - 1) / 400 - 4);       /* + 400 year days */
 742
 743   DEBUG_printf(("    days=%d\n", days));
 744
 745   return (days * 86400 + hour * 3600 + min * 60 + sec);
 746 }
 747
 748
 749 /*
 750  * 'httpSeparate()' - Separate a Universal Resource Identifier into its
 751  *                    components.
 752  *
 753  * This function is deprecated; use the httpSeparateURI() function instead.
 754  *
 755  * @deprecated@
 756  */
 757
 758 void
 759 httpSeparate(const char *uri,           /* I - Universal Resource Identifier */
 760              char       *scheme,        /* O - Scheme [32] (http, https, etc.) */
 761              char       *username,      /* O - Username [1024] */
 762              char       *host,          /* O - Hostname [1024] */
 763              int        *port,          /* O - Port number to use */
 764              char       *resource)      /* O - Resource/filename [1024] */
 765 {
 766   httpSeparateURI(HTTP_URI_CODING_ALL, uri, scheme, 32, username,
 767                   HTTP_MAX_URI, host, HTTP_MAX_URI, port, resource,
 768                   HTTP_MAX_URI);
 769 }
 770
 771
 772 /*
 773  * 'httpSeparate2()' - Separate a Universal Resource Identifier into its
 774  *                     components.
 775  *
 776  * This function is deprecated; use the httpSeparateURI() function instead.
 777  *
 778  * @since CUPS 1.1.21@
 779  * @deprecated@
 780  */
 781
 782 void
 783 httpSeparate2(const char *uri,          /* I - Universal Resource Identifier */
 784               char       *scheme,       /* O - Scheme (http, https, etc.) */
 785               int        schemelen,     /* I - Size of scheme buffer */
 786               char       *username,     /* O - Username */
 787               int        usernamelen,   /* I - Size of username buffer */
 788               char       *host,         /* O - Hostname */
 789               int        hostlen,       /* I - Size of hostname buffer */
 790               int        *port,         /* O - Port number to use */
 791               char       *resource,     /* O - Resource/filename */
 792               int        resourcelen)   /* I - Size of resource buffer */
 793 {
 794   httpSeparateURI(HTTP_URI_CODING_ALL, uri, scheme, schemelen, username,
 795                   usernamelen, host, hostlen, port, resource, resourcelen);
 796 }
 797
 798
 799 /*
 800  * 'httpSeparateURI()' - Separate a Universal Resource Identifier into its
 801  *                       components.
 802  *
 803  * @since CUPS 1.2@
 804  */
 805
 806 http_uri_status_t                       /* O - Result of separation */
 807 httpSeparateURI(
 808     http_uri_coding_t decoding,         /* I - Decoding flags */
 809     const char        *uri,             /* I - Universal Resource Identifier */
 810     char              *scheme,          /* O - Scheme (http, https, etc.) */
 811     int               schemelen,        /* I - Size of scheme buffer */
 812     char              *username,        /* O - Username */
 813     int               usernamelen,      /* I - Size of username buffer */
 814     char              *host,            /* O - Hostname */
 815     int               hostlen,          /* I - Size of hostname buffer */
 816     int               *port,            /* O - Port number to use */
 817     char              *resource,        /* O - Resource/filename */
 818     int               resourcelen)      /* I - Size of resource buffer */
 819 {
 820   char                  *ptr,           /* Pointer into string... */
 821                         *end;           /* End of string */
 822   const char            *sep;           /* Separator character */
 823   http_uri_status_t     status;         /* Result of separation */
 824
 825
 826  /*
 827   * Initialize everything to blank...
 828   */
 829
 830   if (scheme && schemelen > 0)
 831     *scheme = '\0';
 832
 833   if (username && usernamelen > 0)
 834     *username = '\0';
 835
 836   if (host && hostlen > 0)
 837     *host = '\0';
 838
 839   if (port)
 840     *port = 0;
 841
 842   if (resource && resourcelen > 0)
 843     *resource = '\0';
 844
 845  /*
 846   * Range check input...
 847   */
 848
 849   if (!uri || !port || !scheme || schemelen <= 0 || !username ||
 850       usernamelen <= 0 || !host || hostlen <= 0 || !resource ||
 851       resourcelen <= 0)
 852     return (HTTP_URI_BAD_ARGUMENTS);
 853
 854   if (!*uri)
 855     return (HTTP_URI_BAD_URI);
 856
 857  /*
 858   * Grab the scheme portion of the URI...
 859   */
 860
 861   status = HTTP_URI_OK;
 862
 863   if (!strncmp(uri, "//", 2))
 864   {
 865    /*
 866     * Workaround for HP IPP client bug...
 867     */
 868
 869     strlcpy(scheme, "ipp", schemelen);
 870     status = HTTP_URI_MISSING_SCHEME;
 871   }
 872   else if (*uri == '/')
 873   {
 874    /*
 875     * Filename...
 876     */
 877
 878     strlcpy(scheme, "file", schemelen);
 879     status = HTTP_URI_MISSING_SCHEME;
 880   }
 881   else
 882   {
 883    /*
 884     * Standard URI with scheme...
 885     */
 886
 887     for (ptr = scheme, end = scheme + schemelen - 1;
 888          *uri && *uri != ':' && ptr < end;)
 889       if (isalnum(*uri & 255) || *uri == '-' || *uri == '+' || *uri == '.')
 890         *ptr++ = *uri++;
 891       else
 892         break;
 893
 894     *ptr = '\0';
 895
 896     if (*uri != ':')
 897     {
 898       *scheme = '\0';
 899       return (HTTP_URI_BAD_SCHEME);
 900     }
 901
 902     uri ++;
 903   }
 904
 905  /*
 906   * Set the default port number...
 907   */
 908
 909   if (!strcmp(scheme, "http"))
 910     *port = 80;
 911   else if (!strcmp(scheme, "https"))
 912     *port = 443;
 913   else if (!strcmp(scheme, "ipp"))
 914     *port = 631;
 915   else if (!strcasecmp(scheme, "lpd"))
 916     *port = 515;
 917   else if (!strcmp(scheme, "socket"))   /* Not yet registered with IANA... */
 918     *port = 9100;
 919   else if (strcmp(scheme, "file") && strcmp(scheme, "mailto"))
 920     status = HTTP_URI_UNKNOWN_SCHEME;
 921
 922  /*
 923   * Now see if we have a hostname...
 924   */
 925
 926   if (!strncmp(uri, "//", 2))
 927   {
 928    /*
 929     * Yes, extract it...
 930     */
 931
 932     uri += 2;
 933
 934    /*
 935     * Grab the username, if any...
 936     */
 937
 938     if ((sep = strpbrk(uri, "@/")) != NULL && *sep == '@')
 939     {
 940      /*
 941       * Get a username:password combo...
 942       */
 943
 944       uri = http_copy_decode(username, uri, usernamelen, "@",
 945                              decoding & HTTP_URI_CODING_USERNAME);
 946
 947       if (!uri)
 948       {
 949         *username = '\0';
 950         return (HTTP_URI_BAD_USERNAME);
 951       }
 952
 953       uri ++;
 954     }
 955
 956    /*
 957     * Then the hostname/IP address...
 958     */
 959
 960     if (*uri == '[')
 961     {
 962      /*
 963       * Grab IPv6 address...
 964       */
 965
 966       uri ++;
 967       if (!strncmp(uri, "v1.", 3))
 968         uri += 3;                       /* Skip IPvN leader... */
 969
 970       uri = http_copy_decode(host, uri, hostlen, "]",
 971                              decoding & HTTP_URI_CODING_HOSTNAME);
 972
 973       if (!uri)
 974       {
 975         *host = '\0';
 976         return (HTTP_URI_BAD_HOSTNAME);
 977       }
 978
 979      /*
 980       * Validate value...
 981       */
 982
 983       if (*uri != ']')
 984       {
 985         *host = '\0';
 986         return (HTTP_URI_BAD_HOSTNAME);
 987       }
 988
 989       uri ++;
 990
 991       for (ptr = host; *ptr; ptr ++)
 992         if (*ptr == '+')
 993         {
 994          /*
 995           * Convert zone separator to % and stop here...
 996           */
 997
 998           *ptr = '%';
 999           break;
1000         }
1001         else if (*ptr != ':' && *ptr != '.' && !isxdigit(*ptr & 255))
1002         {
1003           *host = '\0';
1004           return (HTTP_URI_BAD_HOSTNAME);
1005         }
1006     }
1007     else
1008     {
1009      /*
1010       * Validate the hostname or IPv4 address first...
1011       */
1012
1013       for (ptr = (char *)uri; *ptr; ptr ++)
1014         if (strchr(":?/", *ptr))
1015           break;
1016         else if (!strchr("abcdefghijklmnopqrstuvwxyz"
1017                          "ABCDEFGHIJKLMNOPQRSTUVWXYZ"
1018                          "0123456789"
1019                          "-._~"
1020                          "%"
1021                          "!$&'()*+,;=\\", *ptr))
1022         {
1023           *host = '\0';
1024           return (HTTP_URI_BAD_HOSTNAME);
1025         }
1026
1027      /*
1028       * Then copy the hostname or IPv4 address to the buffer...
1029       */
1030
1031       uri = http_copy_decode(host, uri, hostlen, ":?/",
1032                              decoding & HTTP_URI_CODING_HOSTNAME);
1033
1034       if (!uri)
1035       {
1036         *host = '\0';
1037         return (HTTP_URI_BAD_HOSTNAME);
1038       }
1039     }
1040
1041    /*
1042     * Validate hostname for file scheme - only empty and localhost are
1043     * acceptable.
1044     */
1045
1046     if (!strcmp(scheme, "file") && strcmp(host, "localhost") && host[0])
1047     {
1048       *host = '\0';
1049       return (HTTP_URI_BAD_HOSTNAME);
1050     }
1051
1052    /*
1053     * See if we have a port number...
1054     */
1055
1056     if (*uri == ':')
1057     {
1058      /*
1059       * Yes, collect the port number...
1060       */
1061
1062       *port = strtol(uri + 1, (char **)&uri, 10);
1063
1064       if (*uri != '/' && *uri)
1065       {
1066         *port = 0;
1067         return (HTTP_URI_BAD_PORT);
1068       }
1069     }
1070   }
1071
1072  /*
1073   * The remaining portion is the resource string...
1074   */
1075
1076   if (*uri == '?' || !*uri)
1077   {
1078    /*
1079     * Hostname but no path...
1080     */
1081
1082     status    = HTTP_URI_MISSING_RESOURCE;
1083     *resource = '/';
1084
1085    /*
1086     * Copy any query string...
1087     */
1088
1089     if (*uri == '?')
1090       uri = http_copy_decode(resource + 1, uri, resourcelen - 1, NULL,
1091                              decoding & HTTP_URI_CODING_QUERY);
1092     else
1093       resource[1] = '\0';
1094   }
1095   else
1096   {
1097     uri = http_copy_decode(resource, uri, resourcelen, "?",
1098                            decoding & HTTP_URI_CODING_RESOURCE);
1099
1100     if (uri && *uri == '?')
1101     {
1102      /*
1103       * Concatenate any query string...
1104       */
1105
1106       char *resptr = resource + strlen(resource);
1107
1108       uri = http_copy_decode(resptr, uri, resourcelen - (int)(resptr - resource),
1109                              NULL, decoding & HTTP_URI_CODING_QUERY);
1110     }
1111   }
1112
1113   if (!uri)
1114   {
1115     *resource = '\0';
1116     return (HTTP_URI_BAD_RESOURCE);
1117   }
1118
1119  /*
1120   * Return the URI separation status...
1121   */
1122
1123   return (status);
1124 }
1125
1126
1127 /*
1128  * 'httpStatus()' - Return a short string describing a HTTP status code.
1129  */
1130
1131 const char *                            /* O - String or NULL */
1132 httpStatus(http_status_t status)        /* I - HTTP status code */
1133 {
1134   switch (status)
1135   {
1136     case HTTP_CONTINUE :
1137         return ("Continue");
1138     case HTTP_SWITCHING_PROTOCOLS :
1139         return ("Switching Protocols");
1140     case HTTP_OK :
1141         return ("OK");
1142     case HTTP_CREATED :
1143         return ("Created");
1144     case HTTP_ACCEPTED :
1145         return ("Accepted");
1146     case HTTP_NO_CONTENT :
1147         return ("No Content");
1148     case HTTP_MOVED_PERMANENTLY :
1149         return ("Moved Permanently");
1150     case HTTP_SEE_OTHER :
1151         return ("See Other");
1152     case HTTP_NOT_MODIFIED :
1153         return ("Not Modified");
1154     case HTTP_BAD_REQUEST :
1155         return ("Bad Request");
1156     case HTTP_UNAUTHORIZED :
1157         return ("Unauthorized");
1158     case HTTP_FORBIDDEN :
1159         return ("Forbidden");
1160     case HTTP_NOT_FOUND :
1161         return ("Not Found");
1162     case HTTP_REQUEST_TOO_LARGE :
1163         return ("Request Entity Too Large");
1164     case HTTP_URI_TOO_LONG :
1165         return ("URI Too Long");
1166     case HTTP_UPGRADE_REQUIRED :
1167         return ("Upgrade Required");
1168     case HTTP_NOT_IMPLEMENTED :
1169         return ("Not Implemented");
1170     case HTTP_NOT_SUPPORTED :
1171         return ("Not Supported");
1172     case HTTP_EXPECTATION_FAILED :
1173         return ("Expectation Failed");
1174
1175     default :
1176         return ("Unknown");
1177   }
1178 }
1179
1180
1181 #ifndef HAVE_HSTRERROR
1182 /*
1183  * '_cups_hstrerror()' - hstrerror() emulation function for Solaris and others...
1184  */
1185
1186 const char *                            /* O - Error string */
1187 _cups_hstrerror(int error)              /* I - Error number */
1188 {
1189   static const char * const errors[] =  /* Error strings */
1190                 {
1191                   "OK",
1192                   "Host not found.",
1193                   "Try again.",
1194                   "Unrecoverable lookup error.",
1195                   "No data associated with name."
1196                 };
1197
1198
1199   if (error < 0 || error > 4)
1200     return ("Unknown hostname lookup error.");
1201   else
1202     return (errors[error]);
1203 }
1204 #endif /* !HAVE_HSTRERROR */
1205
1206
1207 /*
1208  * '_httpEncodeURI()' - Percent-encode a HTTP request URI.
1209  */
1210
1211 char *                                  /* O - Encoded URI */
1212 _httpEncodeURI(char       *dst,         /* I - Destination buffer */
1213                const char *src,         /* I - Source URI */
1214                size_t     dstsize)      /* I - Size of destination buffer */
1215 {
1216   http_copy_encode(dst, src, dst + dstsize - 1, NULL, NULL, 1);
1217   return (dst);
1218 }
1219
1220
1221 /*
1222  * 'http_copy_decode()' - Copy and decode a URI.
1223  */
1224
1225 static const char *                     /* O - New source pointer or NULL on error */
1226 http_copy_decode(char       *dst,       /* O - Destination buffer */
1227                  const char *src,       /* I - Source pointer */
1228                  int        dstsize,    /* I - Destination size */
1229                  const char *term,      /* I - Terminating characters */
1230                  int        decode)     /* I - Decode %-encoded values */
1231 {
1232   char  *ptr,                           /* Pointer into buffer */
1233         *end;                           /* End of buffer */
1234   int   quoted;                         /* Quoted character */
1235
1236
1237  /*
1238   * Copy the src to the destination until we hit a terminating character
1239   * or the end of the string.
1240   */
1241
1242   for (ptr = dst, end = dst + dstsize - 1;
1243        *src && (!term || !strchr(term, *src));
1244        src ++)
1245     if (ptr < end)
1246     {
1247       if (*src == '%' && decode)
1248       {
1249         if (isxdigit(src[1] & 255) && isxdigit(src[2] & 255))
1250         {
1251          /*
1252           * Grab a hex-encoded character...
1253           */
1254
1255           src ++;
1256           if (isalpha(*src))
1257             quoted = (tolower(*src) - 'a' + 10) << 4;
1258           else
1259             quoted = (*src - '0') << 4;
1260
1261           src ++;
1262           if (isalpha(*src))
1263             quoted |= tolower(*src) - 'a' + 10;
1264           else
1265             quoted |= *src - '0';
1266
1267           *ptr++ = quoted;
1268         }
1269         else
1270         {
1271          /*
1272           * Bad hex-encoded character...
1273           */
1274
1275           *ptr = '\0';
1276           return (NULL);
1277         }
1278       }
1279       else
1280         *ptr++ = *src;
1281     }
1282
1283   *ptr = '\0';
1284
1285   return (src);
1286 }
1287
1288
1289 /*
1290  * 'http_copy_encode()' - Copy and encode a URI.
1291  */
1292
1293 static char *                           /* O - End of current URI */
1294 http_copy_encode(char       *dst,       /* O - Destination buffer */
1295                  const char *src,       /* I - Source pointer */
1296                  char       *dstend,    /* I - End of destination buffer */
1297                  const char *reserved,  /* I - Extra reserved characters */
1298                  const char *term,      /* I - Terminating characters */
1299                  int        encode)     /* I - %-encode reserved chars? */
1300 {
1301   static const char hex[] = "0123456789ABCDEF";
1302
1303
1304   while (*src && dst < dstend)
1305   {
1306     if (term && *src == *term)
1307       return (dst);
1308
1309     if (encode && (*src == '%' || *src <= ' ' || *src & 128 ||
1310                    (reserved && strchr(reserved, *src))))
1311     {
1312      /*
1313       * Hex encode reserved characters...
1314       */
1315
1316       if ((dst + 2) >= dstend)
1317         break;
1318
1319       *dst++ = '%';
1320       *dst++ = hex[(*src >> 4) & 15];
1321       *dst++ = hex[*src & 15];
1322
1323       src ++;
1324     }
1325     else
1326       *dst++ = *src++;
1327   }
1328
1329   *dst = '\0';
1330
1331   if (*src)
1332     return (NULL);
1333   else
1334     return (dst);
1335 }
1336
1337
1338 /*
1339  * End of "$Id: http-support.c 6649 2007-07-11 21:46:42Z mike $".
1340  */