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.  Since DNS-SD service names can sometimes look like
 201     * raw IPv6 addresses, we specifically look for "._tcp" in the name,
 202     * too...
 203     */
 204
 205     if (host[0] != '[' && strchr(host, ':') && !strstr(host, "._tcp"))
 206     {
 207      /*
 208       * We have a raw IPv6 address...
 209       */
 210
 211       if (strchr(host, '%'))
 212       {
 213        /*
 214         * We have a link-local address, add "[v1." prefix...
 215         */
 216
 217         if ((ptr + 4) < end)
 218         {
 219           *ptr++ = '[';
 220           *ptr++ = 'v';
 221           *ptr++ = '1';
 222           *ptr++ = '.';
 223         }
 224         else
 225           goto assemble_overflow;
 226       }
 227       else
 228       {
 229        /*
 230         * We have a normal address, add "[" prefix...
 231         */
 232
 233         if (ptr < end)
 234           *ptr++ = '[';
 235         else
 236           goto assemble_overflow;
 237       }
 238
 239      /*
 240       * Copy the rest of the IPv6 address, and terminate with "]".
 241       */
 242
 243       while (ptr < end && *host)
 244       {
 245         if (*host == '%')
 246         {
 247           *ptr++ = '+';                 /* Convert zone separator */
 248           host ++;
 249         }
 250         else
 251           *ptr++ = *host++;
 252       }
 253
 254       if (*host)
 255         goto assemble_overflow;
 256
 257       if (ptr < end)
 258         *ptr++ = ']';
 259       else
 260         goto assemble_overflow;
 261     }
 262     else
 263     {
 264      /*
 265       * Otherwise, just copy the host string...
 266       */
 267
 268       ptr = http_copy_encode(ptr, host, end, ":/?#[]@", NULL,
 269                              encoding & HTTP_URI_CODING_HOSTNAME);
 270
 271       if (!ptr)
 272         goto assemble_overflow;
 273     }
 274
 275    /*
 276     * Finish things off with the port number...
 277     */
 278
 279     if (port > 0)
 280     {
 281       snprintf(ptr, end - ptr + 1, ":%d", port);
 282       ptr += strlen(ptr);
 283
 284       if (ptr >= end)
 285         goto assemble_overflow;
 286     }
 287   }
 288
 289  /*
 290   * Last but not least, add the resource string...
 291   */
 292
 293   if (resource)
 294   {
 295     char        *query;                 /* Pointer to query string */
 296
 297
 298    /*
 299     * Copy the resource string up to the query string if present...
 300     */
 301
 302     query = strchr(resource, '?');
 303     ptr   = http_copy_encode(ptr, resource, end, NULL, "?",
 304                              encoding & HTTP_URI_CODING_RESOURCE);
 305     if (!ptr)
 306       goto assemble_overflow;
 307
 308     if (query)
 309     {
 310      /*
 311       * Copy query string without encoding...
 312       */
 313
 314       ptr = http_copy_encode(ptr, query, end, NULL, NULL,
 315                              encoding & HTTP_URI_CODING_QUERY);
 316       if (!ptr)
 317         goto assemble_overflow;
 318     }
 319   }
 320   else if (ptr < end)
 321     *ptr++ = '/';
 322   else
 323     goto assemble_overflow;
 324
 325  /*
 326   * Nul-terminate the URI buffer and return with no errors...
 327   */
 328
 329   *ptr = '\0';
 330
 331   return (HTTP_URI_OK);
 332
 333  /*
 334   * Clear the URI string and return an overflow error; I don't usually
 335   * like goto's, but in this case it makes sense...
 336   */
 337
 338   assemble_overflow:
 339
 340   *uri = '\0';
 341   return (HTTP_URI_OVERFLOW);
 342 }
 343
 344
 345 /*
 346  * 'httpAssembleURIf()' - Assemble a uniform resource identifier from its
 347  *                        components with a formatted resource.
 348  *
 349  * This function creates a formatted version of the resource string
 350  * argument "resourcef" and escapes reserved characters in the URI
 351  * depending on the value of the "encoding" argument.  You should use
 352  * this function in place of traditional string functions whenever
 353  * you need to create a URI string.
 354  *
 355  * @since CUPS 1.2@
 356  */
 357
 358 http_uri_status_t                       /* O - URI status */
 359 httpAssembleURIf(
 360     http_uri_coding_t encoding,         /* I - Encoding flags */
 361     char              *uri,             /* I - URI buffer */
 362     int               urilen,           /* I - Size of URI buffer */
 363     const char        *scheme,          /* I - Scheme name */
 364     const char        *username,        /* I - Username */
 365     const char        *host,            /* I - Hostname or address */
 366     int               port,             /* I - Port number */
 367     const char        *resourcef,       /* I - Printf-style resource */
 368     ...)                                /* I - Additional arguments as needed */
 369 {
 370   va_list       ap;                     /* Pointer to additional arguments */
 371   char          resource[1024];         /* Formatted resource string */
 372   int           bytes;                  /* Bytes in formatted string */
 373
 374
 375  /*
 376   * Range check input...
 377   */
 378
 379   if (!uri || urilen < 1 || !scheme || port < 0 || !resourcef)
 380   {
 381     if (uri)
 382       *uri = '\0';
 383
 384     return (HTTP_URI_BAD_ARGUMENTS);
 385   }
 386
 387  /*
 388   * Format the resource string and assemble the URI...
 389   */
 390
 391   va_start(ap, resourcef);
 392   bytes = vsnprintf(resource, sizeof(resource), resourcef, ap);
 393   va_end(ap);
 394
 395   if (bytes >= sizeof(resource))
 396   {
 397     *uri = '\0';
 398     return (HTTP_URI_OVERFLOW);
 399   }
 400   else
 401     return (httpAssembleURI(encoding,  uri, urilen, scheme, username, host,
 402                             port, resource));
 403 }
 404
 405
 406 /*
 407  * 'httpDecode64()' - Base64-decode a string.
 408  *
 409  * This function is deprecated. Use the httpDecode64_2() function instead
 410  * which provides buffer length arguments.
 411  *
 412  * @deprecated@
 413  */
 414
 415 char *                                  /* O - Decoded string */
 416 httpDecode64(char       *out,           /* I - String to write to */
 417              const char *in)            /* I - String to read from */
 418 {
 419   int   outlen;                         /* Output buffer length */
 420
 421
 422  /*
 423   * Use the old maximum buffer size for binary compatibility...
 424   */
 425
 426   outlen = 512;
 427
 428   return (httpDecode64_2(out, &outlen, in));
 429 }
 430
 431
 432 /*
 433  * 'httpDecode64_2()' - Base64-decode a string.
 434  *
 435  * @since CUPS 1.1.21@
 436  */
 437
 438 char *                                  /* O  - Decoded string */
 439 httpDecode64_2(char       *out,         /* I  - String to write to */
 440                int        *outlen,      /* IO - Size of output string */
 441                const char *in)          /* I  - String to read from */
 442 {
 443   int   pos,                            /* Bit position */
 444         base64;                         /* Value of this character */
 445   char  *outptr,                        /* Output pointer */
 446         *outend;                        /* End of output buffer */
 447
 448
 449  /*
 450   * Range check input...
 451   */
 452
 453   if (!out || !outlen || *outlen < 1 || !in)
 454     return (NULL);
 455
 456   if (!*in)
 457   {
 458     *out    = '\0';
 459     *outlen = 0;
 460
 461     return (out);
 462   }
 463
 464  /*
 465   * Convert from base-64 to bytes...
 466   */
 467
 468   for (outptr = out, outend = out + *outlen - 1, pos = 0; *in != '\0'; in ++)
 469   {
 470    /*
 471     * Decode this character into a number from 0 to 63...
 472     */
 473
 474     if (*in >= 'A' && *in <= 'Z')
 475       base64 = *in - 'A';
 476     else if (*in >= 'a' && *in <= 'z')
 477       base64 = *in - 'a' + 26;
 478     else if (*in >= '0' && *in <= '9')
 479       base64 = *in - '0' + 52;
 480     else if (*in == '+')
 481       base64 = 62;
 482     else if (*in == '/')
 483       base64 = 63;
 484     else if (*in == '=')
 485       break;
 486     else
 487       continue;
 488
 489    /*
 490     * Store the result in the appropriate chars...
 491     */
 492
 493     switch (pos)
 494     {
 495       case 0 :
 496           if (outptr < outend)
 497             *outptr = base64 << 2;
 498           pos ++;
 499           break;
 500       case 1 :
 501           if (outptr < outend)
 502             *outptr++ |= (base64 >> 4) & 3;
 503           if (outptr < outend)
 504             *outptr = (base64 << 4) & 255;
 505           pos ++;
 506           break;
 507       case 2 :
 508           if (outptr < outend)
 509             *outptr++ |= (base64 >> 2) & 15;
 510           if (outptr < outend)
 511             *outptr = (base64 << 6) & 255;
 512           pos ++;
 513           break;
 514       case 3 :
 515           if (outptr < outend)
 516             *outptr++ |= base64;
 517           pos = 0;
 518           break;
 519     }
 520   }
 521
 522   *outptr = '\0';
 523
 524  /*
 525   * Return the decoded string and size...
 526   */
 527
 528   *outlen = (int)(outptr - out);
 529
 530   return (out);
 531 }
 532
 533
 534 /*
 535  * 'httpEncode64()' - Base64-encode a string.
 536  *
 537  * This function is deprecated. Use the httpEncode64_2() function instead
 538  * which provides buffer length arguments.
 539  *
 540  * @deprecated@
 541  */
 542
 543 char *                                  /* O - Encoded string */
 544 httpEncode64(char       *out,           /* I - String to write to */
 545              const char *in)            /* I - String to read from */
 546 {
 547   return (httpEncode64_2(out, 512, in, (int)strlen(in)));
 548 }
 549
 550
 551 /*
 552  * 'httpEncode64_2()' - Base64-encode a string.
 553  *
 554  * @since CUPS 1.1.21@
 555  */
 556
 557 char *                                  /* O - Encoded string */
 558 httpEncode64_2(char       *out,         /* I - String to write to */
 559                int        outlen,       /* I - Size of output string */
 560                const char *in,          /* I - String to read from */
 561                int        inlen)        /* I - Size of input string */
 562 {
 563   char          *outptr,                /* Output pointer */
 564                 *outend;                /* End of output buffer */
 565   static const char base64[] =          /* Base64 characters... */
 566                 {
 567                   "ABCDEFGHIJKLMNOPQRSTUVWXYZ"
 568                   "abcdefghijklmnopqrstuvwxyz"
 569                   "0123456789"
 570                   "+/"
 571                 };
 572
 573
 574  /*
 575   * Range check input...
 576   */
 577
 578   if (!out || outlen < 1 || !in)
 579     return (NULL);
 580
 581  /*
 582   * Convert bytes to base-64...
 583   */
 584
 585   for (outptr = out, outend = out + outlen - 1; inlen > 0; in ++, inlen --)
 586   {
 587    /*
 588     * Encode the up to 3 characters as 4 Base64 numbers...
 589     */
 590
 591     if (outptr < outend)
 592       *outptr ++ = base64[(in[0] & 255) >> 2];
 593
 594     if (outptr < outend)
 595     {
 596       if (inlen > 1)
 597         *outptr ++ = base64[(((in[0] & 255) << 4) | ((in[1] & 255) >> 4)) & 63];
 598       else
 599         *outptr ++ = base64[((in[0] & 255) << 4) & 63];
 600     }
 601
 602     in ++;
 603     inlen --;
 604     if (inlen <= 0)
 605     {
 606       if (outptr < outend)
 607         *outptr ++ = '=';
 608       if (outptr < outend)
 609         *outptr ++ = '=';
 610       break;
 611     }
 612
 613     if (outptr < outend)
 614     {
 615       if (inlen > 1)
 616         *outptr ++ = base64[(((in[0] & 255) << 2) | ((in[1] & 255) >> 6)) & 63];
 617       else
 618         *outptr ++ = base64[((in[0] & 255) << 2) & 63];
 619     }
 620
 621     in ++;
 622     inlen --;
 623     if (inlen <= 0)
 624     {
 625       if (outptr < outend)
 626         *outptr ++ = '=';
 627       break;
 628     }
 629
 630     if (outptr < outend)
 631       *outptr ++ = base64[in[0] & 63];
 632   }
 633
 634   *outptr = '\0';
 635
 636  /*
 637   * Return the encoded string...
 638   */
 639
 640   return (out);
 641 }
 642
 643
 644 /*
 645  * 'httpGetDateString()' - Get a formatted date/time string from a time value.
 646  *
 647  * @deprecated@
 648  */
 649
 650 const char *                            /* O - Date/time string */
 651 httpGetDateString(time_t t)             /* I - UNIX time */
 652 {
 653   _cups_globals_t *cg = _cupsGlobals(); /* Pointer to library globals */
 654
 655
 656   return (httpGetDateString2(t, cg->http_date, sizeof(cg->http_date)));
 657 }
 658
 659
 660 /*
 661  * 'httpGetDateString2()' - Get a formatted date/time string from a time value.
 662  *
 663  * @since CUPS 1.2@
 664  */
 665
 666 const char *                            /* O - Date/time string */
 667 httpGetDateString2(time_t t,            /* I - UNIX time */
 668                    char   *s,           /* I - String buffer */
 669                    int    slen)         /* I - Size of string buffer */
 670 {
 671   struct tm     *tdate;                 /* UNIX date/time data */
 672
 673
 674   tdate = gmtime(&t);
 675   snprintf(s, slen, "%s, %02d %s %d %02d:%02d:%02d GMT",
 676            http_days[tdate->tm_wday], tdate->tm_mday,
 677            http_months[tdate->tm_mon], tdate->tm_year + 1900,
 678            tdate->tm_hour, tdate->tm_min, tdate->tm_sec);
 679
 680   return (s);
 681 }
 682
 683
 684 /*
 685  * 'httpGetDateTime()' - Get a time value from a formatted date/time string.
 686  */
 687
 688 time_t                                  /* O - UNIX time */
 689 httpGetDateTime(const char *s)          /* I - Date/time string */
 690 {
 691   int           i;                      /* Looping var */
 692   char          mon[16];                /* Abbreviated month name */
 693   int           day, year;              /* Day of month and year */
 694   int           hour, min, sec;         /* Time */
 695   int           days;                   /* Number of days since 1970 */
 696   static const int normal_days[] =      /* Days to a month, normal years */
 697                 { 0, 31, 59, 90, 120, 151, 181, 212, 243, 273, 304, 334, 365 };
 698   static const int leap_days[] =        /* Days to a month, leap years */
 699                 { 0, 31, 60, 91, 121, 152, 182, 213, 244, 274, 305, 335, 366 };
 700
 701
 702   DEBUG_printf(("httpGetDateTime(s=\"%s\")\n", s));
 703
 704  /*
 705   * Extract the date and time from the formatted string...
 706   */
 707
 708   if (sscanf(s, "%*s%d%15s%d%d:%d:%d", &day, mon, &year, &hour, &min, &sec) < 6)
 709     return (0);
 710
 711   DEBUG_printf(("    day=%d, mon=\"%s\", year=%d, hour=%d, min=%d, sec=%d\n",
 712                 day, mon, year, hour, min, sec));
 713
 714  /*
 715   * Convert the month name to a number from 0 to 11.
 716   */
 717
 718   for (i = 0; i < 12; i ++)
 719     if (!strcasecmp(mon, http_months[i]))
 720       break;
 721
 722   if (i >= 12)
 723     return (0);
 724
 725   DEBUG_printf(("    i=%d\n", i));
 726
 727  /*
 728   * Now convert the date and time to a UNIX time value in seconds since
 729   * 1970.  We can't use mktime() since the timezone may not be UTC but
 730   * the date/time string *is* UTC.
 731   */
 732
 733   if ((year & 3) == 0 && ((year % 100) != 0 || (year % 400) == 0))
 734     days = leap_days[i] + day - 1;
 735   else
 736     days = normal_days[i] + day - 1;
 737
 738   DEBUG_printf(("    days=%d\n", days));
 739
 740   days += (year - 1970) * 365 +         /* 365 days per year (normally) */
 741           ((year - 1) / 4 - 492) -      /* + leap days */
 742           ((year - 1) / 100 - 19) +     /* - 100 year days */
 743           ((year - 1) / 400 - 4);       /* + 400 year days */
 744
 745   DEBUG_printf(("    days=%d\n", days));
 746
 747   return (days * 86400 + hour * 3600 + min * 60 + sec);
 748 }
 749
 750
 751 /*
 752  * 'httpSeparate()' - Separate a Universal Resource Identifier into its
 753  *                    components.
 754  *
 755  * This function is deprecated; use the httpSeparateURI() function instead.
 756  *
 757  * @deprecated@
 758  */
 759
 760 void
 761 httpSeparate(const char *uri,           /* I - Universal Resource Identifier */
 762              char       *scheme,        /* O - Scheme [32] (http, https, etc.) */
 763              char       *username,      /* O - Username [1024] */
 764              char       *host,          /* O - Hostname [1024] */
 765              int        *port,          /* O - Port number to use */
 766              char       *resource)      /* O - Resource/filename [1024] */
 767 {
 768   httpSeparateURI(HTTP_URI_CODING_ALL, uri, scheme, 32, username,
 769                   HTTP_MAX_URI, host, HTTP_MAX_URI, port, resource,
 770                   HTTP_MAX_URI);
 771 }
 772
 773
 774 /*
 775  * 'httpSeparate2()' - Separate a Universal Resource Identifier into its
 776  *                     components.
 777  *
 778  * This function is deprecated; use the httpSeparateURI() function instead.
 779  *
 780  * @since CUPS 1.1.21@
 781  * @deprecated@
 782  */
 783
 784 void
 785 httpSeparate2(const char *uri,          /* I - Universal Resource Identifier */
 786               char       *scheme,       /* O - Scheme (http, https, etc.) */
 787               int        schemelen,     /* I - Size of scheme buffer */
 788               char       *username,     /* O - Username */
 789               int        usernamelen,   /* I - Size of username buffer */
 790               char       *host,         /* O - Hostname */
 791               int        hostlen,       /* I - Size of hostname buffer */
 792               int        *port,         /* O - Port number to use */
 793               char       *resource,     /* O - Resource/filename */
 794               int        resourcelen)   /* I - Size of resource buffer */
 795 {
 796   httpSeparateURI(HTTP_URI_CODING_ALL, uri, scheme, schemelen, username,
 797                   usernamelen, host, hostlen, port, resource, resourcelen);
 798 }
 799
 800
 801 /*
 802  * 'httpSeparateURI()' - Separate a Universal Resource Identifier into its
 803  *                       components.
 804  *
 805  * @since CUPS 1.2@
 806  */
 807
 808 http_uri_status_t                       /* O - Result of separation */
 809 httpSeparateURI(
 810     http_uri_coding_t decoding,         /* I - Decoding flags */
 811     const char        *uri,             /* I - Universal Resource Identifier */
 812     char              *scheme,          /* O - Scheme (http, https, etc.) */
 813     int               schemelen,        /* I - Size of scheme buffer */
 814     char              *username,        /* O - Username */
 815     int               usernamelen,      /* I - Size of username buffer */
 816     char              *host,            /* O - Hostname */
 817     int               hostlen,          /* I - Size of hostname buffer */
 818     int               *port,            /* O - Port number to use */
 819     char              *resource,        /* O - Resource/filename */
 820     int               resourcelen)      /* I - Size of resource buffer */
 821 {
 822   char                  *ptr,           /* Pointer into string... */
 823                         *end;           /* End of string */
 824   const char            *sep;           /* Separator character */
 825   http_uri_status_t     status;         /* Result of separation */
 826
 827
 828  /*
 829   * Initialize everything to blank...
 830   */
 831
 832   if (scheme && schemelen > 0)
 833     *scheme = '\0';
 834
 835   if (username && usernamelen > 0)
 836     *username = '\0';
 837
 838   if (host && hostlen > 0)
 839     *host = '\0';
 840
 841   if (port)
 842     *port = 0;
 843
 844   if (resource && resourcelen > 0)
 845     *resource = '\0';
 846
 847  /*
 848   * Range check input...
 849   */
 850
 851   if (!uri || !port || !scheme || schemelen <= 0 || !username ||
 852       usernamelen <= 0 || !host || hostlen <= 0 || !resource ||
 853       resourcelen <= 0)
 854     return (HTTP_URI_BAD_ARGUMENTS);
 855
 856   if (!*uri)
 857     return (HTTP_URI_BAD_URI);
 858
 859  /*
 860   * Grab the scheme portion of the URI...
 861   */
 862
 863   status = HTTP_URI_OK;
 864
 865   if (!strncmp(uri, "//", 2))
 866   {
 867    /*
 868     * Workaround for HP IPP client bug...
 869     */
 870
 871     strlcpy(scheme, "ipp", schemelen);
 872     status = HTTP_URI_MISSING_SCHEME;
 873   }
 874   else if (*uri == '/')
 875   {
 876    /*
 877     * Filename...
 878     */
 879
 880     strlcpy(scheme, "file", schemelen);
 881     status = HTTP_URI_MISSING_SCHEME;
 882   }
 883   else
 884   {
 885    /*
 886     * Standard URI with scheme...
 887     */
 888
 889     for (ptr = scheme, end = scheme + schemelen - 1;
 890          *uri && *uri != ':' && ptr < end;)
 891       if (isalnum(*uri & 255) || *uri == '-' || *uri == '+' || *uri == '.')
 892         *ptr++ = *uri++;
 893       else
 894         break;
 895
 896     *ptr = '\0';
 897
 898     if (*uri != ':')
 899     {
 900       *scheme = '\0';
 901       return (HTTP_URI_BAD_SCHEME);
 902     }
 903
 904     uri ++;
 905   }
 906
 907  /*
 908   * Set the default port number...
 909   */
 910
 911   if (!strcmp(scheme, "http"))
 912     *port = 80;
 913   else if (!strcmp(scheme, "https"))
 914     *port = 443;
 915   else if (!strcmp(scheme, "ipp"))
 916     *port = 631;
 917   else if (!strcasecmp(scheme, "lpd"))
 918     *port = 515;
 919   else if (!strcmp(scheme, "socket"))   /* Not yet registered with IANA... */
 920     *port = 9100;
 921   else if (strcmp(scheme, "file") && strcmp(scheme, "mailto"))
 922     status = HTTP_URI_UNKNOWN_SCHEME;
 923
 924  /*
 925   * Now see if we have a hostname...
 926   */
 927
 928   if (!strncmp(uri, "//", 2))
 929   {
 930    /*
 931     * Yes, extract it...
 932     */
 933
 934     uri += 2;
 935
 936    /*
 937     * Grab the username, if any...
 938     */
 939
 940     if ((sep = strpbrk(uri, "@/")) != NULL && *sep == '@')
 941     {
 942      /*
 943       * Get a username:password combo...
 944       */
 945
 946       uri = http_copy_decode(username, uri, usernamelen, "@",
 947                              decoding & HTTP_URI_CODING_USERNAME);
 948
 949       if (!uri)
 950       {
 951         *username = '\0';
 952         return (HTTP_URI_BAD_USERNAME);
 953       }
 954
 955       uri ++;
 956     }
 957
 958    /*
 959     * Then the hostname/IP address...
 960     */
 961
 962     if (*uri == '[')
 963     {
 964      /*
 965       * Grab IPv6 address...
 966       */
 967
 968       uri ++;
 969       if (!strncmp(uri, "v1.", 3))
 970         uri += 3;                       /* Skip IPvN leader... */
 971
 972       uri = http_copy_decode(host, uri, hostlen, "]",
 973                              decoding & HTTP_URI_CODING_HOSTNAME);
 974
 975       if (!uri)
 976       {
 977         *host = '\0';
 978         return (HTTP_URI_BAD_HOSTNAME);
 979       }
 980
 981      /*
 982       * Validate value...
 983       */
 984
 985       if (*uri != ']')
 986       {
 987         *host = '\0';
 988         return (HTTP_URI_BAD_HOSTNAME);
 989       }
 990
 991       uri ++;
 992
 993       for (ptr = host; *ptr; ptr ++)
 994         if (*ptr == '+')
 995         {
 996          /*
 997           * Convert zone separator to % and stop here...
 998           */
 999
1000           *ptr = '%';
1001           break;
1002         }
1003         else if (*ptr != ':' && *ptr != '.' && !isxdigit(*ptr & 255))
1004         {
1005           *host = '\0';
1006           return (HTTP_URI_BAD_HOSTNAME);
1007         }
1008     }
1009     else
1010     {
1011      /*
1012       * Validate the hostname or IPv4 address first...
1013       */
1014
1015       for (ptr = (char *)uri; *ptr; ptr ++)
1016         if (strchr(":?/", *ptr))
1017           break;
1018         else if (!strchr("abcdefghijklmnopqrstuvwxyz"
1019                          "ABCDEFGHIJKLMNOPQRSTUVWXYZ"
1020                          "0123456789"
1021                          "-._~"
1022                          "%"
1023                          "!$&'()*+,;=\\", *ptr))
1024         {
1025           *host = '\0';
1026           return (HTTP_URI_BAD_HOSTNAME);
1027         }
1028
1029      /*
1030       * Then copy the hostname or IPv4 address to the buffer...
1031       */
1032
1033       uri = http_copy_decode(host, uri, hostlen, ":?/",
1034                              decoding & HTTP_URI_CODING_HOSTNAME);
1035
1036       if (!uri)
1037       {
1038         *host = '\0';
1039         return (HTTP_URI_BAD_HOSTNAME);
1040       }
1041     }
1042
1043    /*
1044     * Validate hostname for file scheme - only empty and localhost are
1045     * acceptable.
1046     */
1047
1048     if (!strcmp(scheme, "file") && strcmp(host, "localhost") && host[0])
1049     {
1050       *host = '\0';
1051       return (HTTP_URI_BAD_HOSTNAME);
1052     }
1053
1054    /*
1055     * See if we have a port number...
1056     */
1057
1058     if (*uri == ':')
1059     {
1060      /*
1061       * Yes, collect the port number...
1062       */
1063
1064       *port = strtol(uri + 1, (char **)&uri, 10);
1065
1066       if (*uri != '/' && *uri)
1067       {
1068         *port = 0;
1069         return (HTTP_URI_BAD_PORT);
1070       }
1071     }
1072   }
1073
1074  /*
1075   * The remaining portion is the resource string...
1076   */
1077
1078   if (*uri == '?' || !*uri)
1079   {
1080    /*
1081     * Hostname but no path...
1082     */
1083
1084     status    = HTTP_URI_MISSING_RESOURCE;
1085     *resource = '/';
1086
1087    /*
1088     * Copy any query string...
1089     */
1090
1091     if (*uri == '?')
1092       uri = http_copy_decode(resource + 1, uri, resourcelen - 1, NULL,
1093                              decoding & HTTP_URI_CODING_QUERY);
1094     else
1095       resource[1] = '\0';
1096   }
1097   else
1098   {
1099     uri = http_copy_decode(resource, uri, resourcelen, "?",
1100                            decoding & HTTP_URI_CODING_RESOURCE);
1101
1102     if (uri && *uri == '?')
1103     {
1104      /*
1105       * Concatenate any query string...
1106       */
1107
1108       char *resptr = resource + strlen(resource);
1109
1110       uri = http_copy_decode(resptr, uri, resourcelen - (int)(resptr - resource),
1111                              NULL, decoding & HTTP_URI_CODING_QUERY);
1112     }
1113   }
1114
1115   if (!uri)
1116   {
1117     *resource = '\0';
1118     return (HTTP_URI_BAD_RESOURCE);
1119   }
1120
1121  /*
1122   * Return the URI separation status...
1123   */
1124
1125   return (status);
1126 }
1127
1128
1129 /*
1130  * 'httpStatus()' - Return a short string describing a HTTP status code.
1131  */
1132
1133 const char *                            /* O - String or NULL */
1134 httpStatus(http_status_t status)        /* I - HTTP status code */
1135 {
1136   switch (status)
1137   {
1138     case HTTP_CONTINUE :
1139         return ("Continue");
1140     case HTTP_SWITCHING_PROTOCOLS :
1141         return ("Switching Protocols");
1142     case HTTP_OK :
1143         return ("OK");
1144     case HTTP_CREATED :
1145         return ("Created");
1146     case HTTP_ACCEPTED :
1147         return ("Accepted");
1148     case HTTP_NO_CONTENT :
1149         return ("No Content");
1150     case HTTP_MOVED_PERMANENTLY :
1151         return ("Moved Permanently");
1152     case HTTP_SEE_OTHER :
1153         return ("See Other");
1154     case HTTP_NOT_MODIFIED :
1155         return ("Not Modified");
1156     case HTTP_BAD_REQUEST :
1157         return ("Bad Request");
1158     case HTTP_UNAUTHORIZED :
1159         return ("Unauthorized");
1160     case HTTP_FORBIDDEN :
1161         return ("Forbidden");
1162     case HTTP_NOT_FOUND :
1163         return ("Not Found");
1164     case HTTP_REQUEST_TOO_LARGE :
1165         return ("Request Entity Too Large");
1166     case HTTP_URI_TOO_LONG :
1167         return ("URI Too Long");
1168     case HTTP_UPGRADE_REQUIRED :
1169         return ("Upgrade Required");
1170     case HTTP_NOT_IMPLEMENTED :
1171         return ("Not Implemented");
1172     case HTTP_NOT_SUPPORTED :
1173         return ("Not Supported");
1174     case HTTP_EXPECTATION_FAILED :
1175         return ("Expectation Failed");
1176
1177     default :
1178         return ("Unknown");
1179   }
1180 }
1181
1182
1183 #ifndef HAVE_HSTRERROR
1184 /*
1185  * '_cups_hstrerror()' - hstrerror() emulation function for Solaris and others...
1186  */
1187
1188 const char *                            /* O - Error string */
1189 _cups_hstrerror(int error)              /* I - Error number */
1190 {
1191   static const char * const errors[] =  /* Error strings */
1192                 {
1193                   "OK",
1194                   "Host not found.",
1195                   "Try again.",
1196                   "Unrecoverable lookup error.",
1197                   "No data associated with name."
1198                 };
1199
1200
1201   if (error < 0 || error > 4)
1202     return ("Unknown hostname lookup error.");
1203   else
1204     return (errors[error]);
1205 }
1206 #endif /* !HAVE_HSTRERROR */
1207
1208
1209 /*
1210  * '_httpEncodeURI()' - Percent-encode a HTTP request URI.
1211  */
1212
1213 char *                                  /* O - Encoded URI */
1214 _httpEncodeURI(char       *dst,         /* I - Destination buffer */
1215                const char *src,         /* I - Source URI */
1216                size_t     dstsize)      /* I - Size of destination buffer */
1217 {
1218   http_copy_encode(dst, src, dst + dstsize - 1, NULL, NULL, 1);
1219   return (dst);
1220 }
1221
1222
1223 /*
1224  * 'http_copy_decode()' - Copy and decode a URI.
1225  */
1226
1227 static const char *                     /* O - New source pointer or NULL on error */
1228 http_copy_decode(char       *dst,       /* O - Destination buffer */
1229                  const char *src,       /* I - Source pointer */
1230                  int        dstsize,    /* I - Destination size */
1231                  const char *term,      /* I - Terminating characters */
1232                  int        decode)     /* I - Decode %-encoded values */
1233 {
1234   char  *ptr,                           /* Pointer into buffer */
1235         *end;                           /* End of buffer */
1236   int   quoted;                         /* Quoted character */
1237
1238
1239  /*
1240   * Copy the src to the destination until we hit a terminating character
1241   * or the end of the string.
1242   */
1243
1244   for (ptr = dst, end = dst + dstsize - 1;
1245        *src && (!term || !strchr(term, *src));
1246        src ++)
1247     if (ptr < end)
1248     {
1249       if (*src == '%' && decode)
1250       {
1251         if (isxdigit(src[1] & 255) && isxdigit(src[2] & 255))
1252         {
1253          /*
1254           * Grab a hex-encoded character...
1255           */
1256
1257           src ++;
1258           if (isalpha(*src))
1259             quoted = (tolower(*src) - 'a' + 10) << 4;
1260           else
1261             quoted = (*src - '0') << 4;
1262
1263           src ++;
1264           if (isalpha(*src))
1265             quoted |= tolower(*src) - 'a' + 10;
1266           else
1267             quoted |= *src - '0';
1268
1269           *ptr++ = quoted;
1270         }
1271         else
1272         {
1273          /*
1274           * Bad hex-encoded character...
1275           */
1276
1277           *ptr = '\0';
1278           return (NULL);
1279         }
1280       }
1281       else
1282         *ptr++ = *src;
1283     }
1284
1285   *ptr = '\0';
1286
1287   return (src);
1288 }
1289
1290
1291 /*
1292  * 'http_copy_encode()' - Copy and encode a URI.
1293  */
1294
1295 static char *                           /* O - End of current URI */
1296 http_copy_encode(char       *dst,       /* O - Destination buffer */
1297                  const char *src,       /* I - Source pointer */
1298                  char       *dstend,    /* I - End of destination buffer */
1299                  const char *reserved,  /* I - Extra reserved characters */
1300                  const char *term,      /* I - Terminating characters */
1301                  int        encode)     /* I - %-encode reserved chars? */
1302 {
1303   static const char hex[] = "0123456789ABCDEF";
1304
1305
1306   while (*src && dst < dstend)
1307   {
1308     if (term && *src == *term)
1309       return (dst);
1310
1311     if (encode && (*src == '%' || *src <= ' ' || *src & 128 ||
1312                    (reserved && strchr(reserved, *src))))
1313     {
1314      /*
1315       * Hex encode reserved characters...
1316       */
1317
1318       if ((dst + 2) >= dstend)
1319         break;
1320
1321       *dst++ = '%';
1322       *dst++ = hex[(*src >> 4) & 15];
1323       *dst++ = hex[*src & 15];
1324
1325       src ++;
1326     }
1327     else
1328       *dst++ = *src++;
1329   }
1330
1331   *dst = '\0';
1332
1333   if (*src)
1334     return (NULL);
1335   else
1336     return (dst);
1337 }
1338
1339
1340 /*
1341  * End of "$Id: http-support.c 6649 2007-07-11 21:46:42Z mike $".
1342  */