cups/http-support.c

   1 /*
   2  * "$Id: http-support.c 177 2006-06-21 00:20:03Z jlovell $"
   3  *
   4  *   HTTP support routines for the Common UNIX Printing System (CUPS) scheduler.
   5  *
   6  *   Copyright 1997-2006 by Easy Software Products, all rights reserved.
   7  *
   8  *   These coded instructions, statements, and computer programs are the
   9  *   property of Easy Software Products and are protected by Federal
  10  *   copyright law.  Distribution and use rights are outlined in the file
  11  *   "LICENSE.txt" which should have been included with this file.  If this
  12  *   file is missing or damaged please contact Easy Software Products
  13  *   at:
  14  *
  15  *       Attn: CUPS Licensing Information
  16  *       Easy Software Products
  17  *       44141 Airport View Drive, Suite 204
  18  *       Hollywood, Maryland 20636 USA
  19  *
  20  *       Voice: (301) 373-9600
  21  *       EMail: cups-info@cups.org
  22  *         WWW: http://www.cups.org
  23  *
  24  *   This file is subject to the Apple OS-Developed Software exception.
  25  *
  26  * Contents:
  27  *
  28  *   httpAssembleURI()    - Assemble a uniform resource identifier from its
  29  *                          components.
  30  *   httpAssembleURIf()   - Assemble a uniform resource identifier from its
  31  *                          components with a formatted resource.
  32  *   httpDecode64()       - Base64-decode a string.
  33  *   httpDecode64_2()     - Base64-decode a string.
  34  *   httpEncode64()       - Base64-encode a string.
  35  *   httpEncode64_2()     - Base64-encode a string.
  36  *   httpGetDateString()  - Get a formatted date/time string from a time value.
  37  *   httpGetDateString2() - Get a formatted date/time string from a time value.
  38  *   httpGetDateTime()    - Get a time value from a formatted date/time string.
  39  *   httpSeparate()       - Separate a Universal Resource Identifier into its
  40  *                          components.
  41  *   httpSeparate2()      - Separate a Universal Resource Identifier into its
  42  *                          components.
  43  *   httpSeparateURI()    - Separate a Universal Resource Identifier into its
  44  *                          components.
  45  *   httpStatus()         - Return a short string describing a HTTP status code.
  46  *   _cups_hstrerror()    - hstrerror() emulation function for Solaris and
  47  *                          others...
  48  *   http_copy_decode()   - Copy and decode a URI.
  49  *   http_copy_encode()   - Copy and encode a URI.
  50  */
  51
  52 /*
  53  * Include necessary headers...
  54  */
  55
  56 #include "debug.h"
  57 #include "globals.h"
  58 #include <stdlib.h>
  59
  60
  61 /*
  62  * Local globals...
  63  */
  64
  65 static const char * const http_days[7] =
  66                         {
  67                           "Sun",                          "Mon",
  68                           "Tue",
  69                           "Wed",
  70                           "Thu",
  71                           "Fri",
  72                           "Sat"
  73                         };
  74 static const char * const http_months[12] =
  75                         {
  76                           "Jan",
  77                           "Feb",
  78                           "Mar",
  79                           "Apr",
  80                           "May",
  81                           "Jun",
  82                           "Jul",
  83                           "Aug",
  84                           "Sep",
  85                           "Oct",
  86                           "Nov",
  87                           "Dec"
  88                         };
  89
  90
  91 /*
  92  * Local functions...
  93  */
  94
  95 static const char       *http_copy_decode(char *dst, const char *src,
  96                                           int dstsize, const char *term,
  97                                           int decode);
  98 static char             *http_copy_encode(char *dst, const char *src,
  99                                           char *dstend, const char *reserved,
 100                                           const char *term, int encode);
 101
 102
 103 /*
 104  * 'httpAssembleURI()' - Assemble a uniform resource identifier from its
 105  *                       components.
 106  *
 107  * This function escapes reserved characters in the URI depending on the
 108  * value of the "encoding" argument.  You should use this function in
 109  * place of traditional string functions whenever you need to create a
 110  * URI string.
 111  *
 112  * @since CUPS 1.2@
 113  */
 114
 115 http_uri_status_t                       /* O - URI status */
 116 httpAssembleURI(
 117     http_uri_coding_t encoding,         /* I - Encoding flags */
 118     char              *uri,             /* I - URI buffer */
 119     int               urilen,           /* I - Size of URI buffer */
 120     const char        *scheme,          /* I - Scheme name */
 121     const char        *username,        /* I - Username */
 122     const char        *host,            /* I - Hostname or address */
 123     int               port,             /* I - Port number */
 124     const char        *resource)        /* I - Resource */
 125 {
 126   char          *ptr,                   /* Pointer into URI buffer */
 127                 *end;                   /* End of URI buffer */
 128
 129
 130  /*
 131   * Range check input...
 132   */
 133
 134   if (!uri || urilen < 1 || !scheme || port < 0)
 135   {
 136     if (uri)
 137       *uri = '\0';
 138
 139     return (HTTP_URI_BAD_ARGUMENTS);
 140   }
 141
 142  /*
 143   * Assemble the URI starting with the scheme...
 144   */
 145
 146   end = uri + urilen - 1;
 147   ptr = http_copy_encode(uri, scheme, end, NULL, NULL, 0);
 148
 149   if (!ptr)
 150     goto assemble_overflow;
 151
 152   if (!strcmp(scheme, "mailto:"))
 153   {
 154    /*
 155     * mailto: only has :, no //...
 156     */
 157
 158     if (ptr < end)
 159       *ptr++ = ':';
 160     else
 161       goto assemble_overflow;
 162   }
 163   else
 164   {
 165    /*
 166     * Schemes other than mailto: all have //...
 167     */
 168
 169     if ((ptr + 2) < end)
 170     {
 171       *ptr++ = ':';
 172       *ptr++ = '/';
 173       *ptr++ = '/';
 174     }
 175     else
 176       goto assemble_overflow;
 177   }
 178
 179  /*
 180   * Next the username and hostname, if any...
 181   */
 182
 183   if (host)
 184   {
 185     if (username && *username)
 186     {
 187      /*
 188       * Add username@ first...
 189       */
 190
 191       ptr = http_copy_encode(ptr, username, end, "/?@", NULL,
 192                              encoding & HTTP_URI_CODING_USERNAME);
 193
 194       if (!ptr)
 195         goto assemble_overflow;
 196
 197       if (ptr < end)
 198         *ptr++ = '@';
 199       else
 200         goto assemble_overflow;
 201     }
 202
 203    /*
 204     * Then add the hostname.  Since IPv6 is a particular pain to deal
 205     * with, we have several special cases to deal with...  If we get
 206     * an IPv6 address with brackets around it, assume it is already in
 207     * URI format...
 208     */
 209
 210     if (host[0] != '[' && strchr(host, ':'))
 211     {
 212      /*
 213       * We have an IPv6 address...
 214       */
 215
 216       if (strchr(host, '%'))
 217       {
 218        /*
 219         * We have a link-local address, add "[v1." prefix...
 220         */
 221
 222         if ((ptr + 4) < end)
 223         {
 224           *ptr++ = '[';
 225           *ptr++ = 'v';
 226           *ptr++ = '1';
 227           *ptr++ = '.';
 228         }
 229         else
 230           goto assemble_overflow;
 231       }
 232       else
 233       {
 234        /*
 235         * We have a normal address, add "[" prefix...
 236         */
 237
 238         if (ptr < end)
 239           *ptr++ = '[';
 240         else
 241           goto assemble_overflow;
 242       }
 243
 244      /*
 245       * Copy the rest of the IPv6 address, and terminate with "]".
 246       */
 247
 248       while (ptr < end && *host)
 249       {
 250         if (*host == '%')
 251         {
 252           *ptr++ = '+';                 /* Convert zone separator */
 253           host ++;
 254         }
 255         else
 256           *ptr++ = *host++;
 257       }
 258
 259       if (*host)
 260         goto assemble_overflow;
 261
 262       if (ptr < end)
 263         *ptr++ = ']';
 264       else
 265         goto assemble_overflow;
 266     }
 267     else
 268     {
 269      /*
 270       * Otherwise, just copy the host string...
 271       */
 272
 273       ptr = http_copy_encode(ptr, host, end, NULL, NULL,
 274                              encoding & HTTP_URI_CODING_HOSTNAME);
 275
 276       if (!ptr)
 277         goto assemble_overflow;
 278     }
 279
 280    /*
 281     * Finish things off with the port number...
 282     */
 283
 284     if (port > 0)
 285     {
 286       snprintf(ptr, end - ptr + 1, ":%d", port);
 287       ptr += strlen(ptr);
 288
 289       if (ptr >= end)
 290         goto assemble_overflow;
 291     }
 292   }
 293
 294  /*
 295   * Last but not least, add the resource string...
 296   */
 297
 298   if (resource)
 299   {
 300     char        *query;                 /* Pointer to query string */
 301
 302
 303    /*
 304     * Copy the resource string up to the query string if present...
 305     */
 306
 307     query = strchr(resource, '?');
 308     ptr   = http_copy_encode(ptr, resource, end, NULL, "?",
 309                              encoding & HTTP_URI_CODING_RESOURCE);
 310     if (!ptr)
 311       goto assemble_overflow;
 312
 313     if (query)
 314     {
 315      /*
 316       * Copy query string without encoding...
 317       */
 318
 319       ptr = http_copy_encode(ptr, query, end, NULL, NULL,
 320                              encoding & HTTP_URI_CODING_QUERY);
 321       if (!ptr)
 322         goto assemble_overflow;
 323     }
 324   }
 325   else if (ptr < end)
 326     *ptr++ = '/';
 327   else
 328     goto assemble_overflow;
 329
 330  /*
 331   * Nul-terminate the URI buffer and return with no errors...
 332   */
 333
 334   *ptr = '\0';
 335
 336   return (HTTP_URI_OK);
 337
 338  /*
 339   * Clear the URI string and return an overflow error; I don't usually
 340   * like goto's, but in this case it makes sense...
 341   */
 342
 343   assemble_overflow:
 344
 345   *uri = '\0';
 346   return (HTTP_URI_OVERFLOW);
 347 }
 348
 349
 350 /*
 351  * 'httpAssembleURIf()' - Assemble a uniform resource identifier from its
 352  *                        components with a formatted resource.
 353  *
 354  * This function creates a formatted version of the resource string
 355  * argument "resourcef" and escapes reserved characters in the URI
 356  * depending on the value of the "encoding" argument.  You should use
 357  * this function in place of traditional string functions whenever
 358  * you need to create a URI string.
 359  *
 360  * @since CUPS 1.2@
 361  */
 362
 363 http_uri_status_t                       /* O - URI status */
 364 httpAssembleURIf(
 365     http_uri_coding_t encoding,         /* I - Encoding flags */
 366     char              *uri,             /* I - URI buffer */
 367     int               urilen,           /* I - Size of URI buffer */
 368     const char        *scheme,          /* I - Scheme name */
 369     const char        *username,        /* I - Username */
 370     const char        *host,            /* I - Hostname or address */
 371     int               port,             /* I - Port number */
 372     const char        *resourcef,       /* I - Printf-style resource */
 373     ...)                                /* I - Additional arguments as needed */
 374 {
 375   va_list       ap;                     /* Pointer to additional arguments */
 376   char          resource[1024];         /* Formatted resource string */
 377   int           bytes;                  /* Bytes in formatted string */
 378
 379
 380  /*
 381   * Range check input...
 382   */
 383
 384   if (!uri || urilen < 1 || !scheme || port < 0 || !resourcef)
 385   {
 386     if (uri)
 387       *uri = '\0';
 388
 389     return (HTTP_URI_BAD_ARGUMENTS);
 390   }
 391
 392  /*
 393   * Format the resource string and assemble the URI...
 394   */
 395
 396   va_start(ap, resourcef);
 397   bytes = vsnprintf(resource, sizeof(resource), resourcef, ap);
 398   va_end(ap);
 399
 400   if (bytes >= sizeof(resource))
 401   {
 402     *uri = '\0';
 403     return (HTTP_URI_OVERFLOW);
 404   }
 405   else
 406     return (httpAssembleURI(encoding,  uri, urilen, scheme, username, host,
 407                             port, resource));
 408 }
 409
 410
 411 /*
 412  * 'httpDecode64()' - Base64-decode a string.
 413  *
 414  * This function is deprecated. Use the httpDecode64_2() function instead
 415  * which provides buffer length arguments.
 416  *
 417  * @deprecated@
 418  */
 419
 420 char *                                  /* O - Decoded string */
 421 httpDecode64(char       *out,           /* I - String to write to */
 422              const char *in)            /* I - String to read from */
 423 {
 424   int   outlen;                         /* Output buffer length */
 425
 426
 427  /*
 428   * Use the old maximum buffer size for binary compatibility...
 429   */
 430
 431   outlen = 512;
 432
 433   return (httpDecode64_2(out, &outlen, in));
 434 }
 435
 436
 437 /*
 438  * 'httpDecode64_2()' - Base64-decode a string.
 439  *
 440  * @since CUPS 1.1.21@
 441  */
 442
 443 char *                                  /* O  - Decoded string */
 444 httpDecode64_2(char       *out,         /* I  - String to write to */
 445                int        *outlen,      /* IO - Size of output string */
 446                const char *in)          /* I  - String to read from */
 447 {
 448   int   pos,                            /* Bit position */
 449         base64;                         /* Value of this character */
 450   char  *outptr,                        /* Output pointer */
 451         *outend;                        /* End of output buffer */
 452
 453
 454  /*
 455   * Range check input...
 456   */
 457
 458   if (!out || !outlen || *outlen < 1 || !in)
 459     return (NULL);
 460
 461   if (!*in)
 462   {
 463     *out    = '\0';
 464     *outlen = 0;
 465
 466     return (out);
 467   }
 468
 469  /*
 470   * Convert from base-64 to bytes...
 471   */
 472
 473   for (outptr = out, outend = out + *outlen - 1, pos = 0; *in != '\0'; in ++)
 474   {
 475    /*
 476     * Decode this character into a number from 0 to 63...
 477     */
 478
 479     if (*in >= 'A' && *in <= 'Z')
 480       base64 = *in - 'A';
 481     else if (*in >= 'a' && *in <= 'z')
 482       base64 = *in - 'a' + 26;
 483     else if (*in >= '0' && *in <= '9')
 484       base64 = *in - '0' + 52;
 485     else if (*in == '+')
 486       base64 = 62;
 487     else if (*in == '/')
 488       base64 = 63;
 489     else if (*in == '=')
 490       break;
 491     else
 492       continue;
 493
 494    /*
 495     * Store the result in the appropriate chars...
 496     */
 497
 498     switch (pos)
 499     {
 500       case 0 :
 501           if (outptr < outend)
 502             *outptr = base64 << 2;
 503           pos ++;
 504           break;
 505       case 1 :
 506           if (outptr < outend)
 507             *outptr++ |= (base64 >> 4) & 3;
 508           if (outptr < outend)
 509             *outptr = (base64 << 4) & 255;
 510           pos ++;
 511           break;
 512       case 2 :
 513           if (outptr < outend)
 514             *outptr++ |= (base64 >> 2) & 15;
 515           if (outptr < outend)
 516             *outptr = (base64 << 6) & 255;
 517           pos ++;
 518           break;
 519       case 3 :
 520           if (outptr < outend)
 521             *outptr++ |= base64;
 522           pos = 0;
 523           break;
 524     }
 525   }
 526
 527   *outptr = '\0';
 528
 529  /*
 530   * Return the decoded string and size...
 531   */
 532
 533   *outlen = (int)(outptr - out);
 534
 535   return (out);
 536 }
 537
 538
 539 /*
 540  * 'httpEncode64()' - Base64-encode a string.
 541  *
 542  * This function is deprecated. Use the httpEncode64_2() function instead
 543  * which provides buffer length arguments.
 544  *
 545  * @deprecated@
 546  */
 547
 548 char *                                  /* O - Encoded string */
 549 httpEncode64(char       *out,           /* I - String to write to */
 550              const char *in)            /* I - String to read from */
 551 {
 552   return (httpEncode64_2(out, 512, in, strlen(in)));
 553 }
 554
 555
 556 /*
 557  * 'httpEncode64_2()' - Base64-encode a string.
 558  *
 559  * @since CUPS 1.1.21@
 560  */
 561
 562 char *                                  /* O - Encoded string */
 563 httpEncode64_2(char       *out,         /* I - String to write to */
 564                int        outlen,       /* I - Size of output string */
 565                const char *in,          /* I - String to read from */
 566                int        inlen)        /* I - Size of input string */
 567 {
 568   char          *outptr,                /* Output pointer */
 569                 *outend;                /* End of output buffer */
 570   static const char base64[] =          /* Base64 characters... */
 571                 {
 572                   "ABCDEFGHIJKLMNOPQRSTUVWXYZ"
 573                   "abcdefghijklmnopqrstuvwxyz"
 574                   "0123456789"
 575                   "+/"
 576                 };
 577
 578
 579  /*
 580   * Range check input...
 581   */
 582
 583   if (!out || outlen < 1 || !in)
 584     return (NULL);
 585
 586  /*
 587   * Convert bytes to base-64...
 588   */
 589
 590   for (outptr = out, outend = out + outlen - 1; inlen > 0; in ++, inlen --)
 591   {
 592    /*
 593     * Encode the up to 3 characters as 4 Base64 numbers...
 594     */
 595
 596     if (outptr < outend)
 597       *outptr ++ = base64[(in[0] & 255) >> 2];
 598     if (outptr < outend)
 599       *outptr ++ = base64[(((in[0] & 255) << 4) | ((in[1] & 255) >> 4)) & 63];
 600
 601     in ++;
 602     inlen --;
 603     if (inlen <= 0)
 604     {
 605       if (outptr < outend)
 606         *outptr ++ = '=';
 607       if (outptr < outend)
 608         *outptr ++ = '=';
 609       break;
 610     }
 611
 612     if (outptr < outend)
 613       *outptr ++ = base64[(((in[0] & 255) << 2) | ((in[1] & 255) >> 6)) & 63];
 614
 615     in ++;
 616     inlen --;
 617     if (inlen <= 0)
 618     {
 619       if (outptr < outend)
 620         *outptr ++ = '=';
 621       break;
 622     }
 623
 624     if (outptr < outend)
 625       *outptr ++ = base64[in[0] & 63];
 626   }
 627
 628   *outptr = '\0';
 629
 630  /*
 631   * Return the encoded string...
 632   */
 633
 634   return (out);
 635 }
 636
 637
 638 /*
 639  * 'httpGetDateString()' - Get a formatted date/time string from a time value.
 640  *
 641  * @deprecated@
 642  */
 643
 644 const char *                            /* O - Date/time string */
 645 httpGetDateString(time_t t)             /* I - UNIX time */
 646 {
 647   _cups_globals_t *cg = _cupsGlobals(); /* Pointer to library globals */
 648
 649
 650   return (httpGetDateString2(t, cg->http_date, sizeof(cg->http_date)));
 651 }
 652
 653
 654 /*
 655  * 'httpGetDateString2()' - Get a formatted date/time string from a time value.
 656  *
 657  * @since CUPS 1.2@
 658  */
 659
 660 const char *                            /* O - Date/time string */
 661 httpGetDateString2(time_t t,            /* I - UNIX time */
 662                    char   *s,           /* I - String buffer */
 663                    int    slen)         /* I - Size of string buffer */
 664 {
 665   struct tm     *tdate;                 /* UNIX date/time data */
 666
 667
 668   tdate = gmtime(&t);
 669   snprintf(s, slen, "%s, %02d %s %d %02d:%02d:%02d GMT",
 670            http_days[tdate->tm_wday], tdate->tm_mday,
 671            http_months[tdate->tm_mon], tdate->tm_year + 1900,
 672            tdate->tm_hour, tdate->tm_min, tdate->tm_sec);
 673
 674   return (s);
 675 }
 676
 677
 678 /*
 679  * 'httpGetDateTime()' - Get a time value from a formatted date/time string.
 680  */
 681
 682 time_t                                  /* O - UNIX time */
 683 httpGetDateTime(const char *s)          /* I - Date/time string */
 684 {
 685   int           i;                      /* Looping var */
 686   char          mon[16];                /* Abbreviated month name */
 687   int           day, year;              /* Day of month and year */
 688   int           hour, min, sec;         /* Time */
 689   int           days;                   /* Number of days since 1970 */
 690   static const int normal_days[] =      /* Days to a month, normal years */
 691                 { 0, 31, 59, 90, 120, 151, 181, 212, 243, 273, 304, 334, 365 };
 692   static const int leap_days[] =        /* Days to a month, leap years */
 693                 { 0, 31, 60, 91, 121, 152, 182, 213, 244, 274, 305, 335, 366 };
 694
 695
 696   DEBUG_printf(("httpGetDateTime(s=\"%s\")\n", s));
 697
 698  /*
 699   * Extract the date and time from the formatted string...
 700   */
 701
 702   if (sscanf(s, "%*s%d%15s%d%d:%d:%d", &day, mon, &year, &hour, &min, &sec) < 6)
 703     return (0);
 704
 705   DEBUG_printf(("    day=%d, mon=\"%s\", year=%d, hour=%d, min=%d, sec=%d\n",
 706                 day, mon, year, hour, min, sec));
 707
 708  /*
 709   * Convert the month name to a number from 0 to 11.
 710   */
 711
 712   for (i = 0; i < 12; i ++)
 713     if (!strcasecmp(mon, http_months[i]))
 714       break;
 715
 716   if (i >= 12)
 717     return (0);
 718
 719   DEBUG_printf(("    i=%d\n", i));
 720
 721  /*
 722   * Now convert the date and time to a UNIX time value in seconds since
 723   * 1970.  We can't use mktime() since the timezone may not be UTC but
 724   * the date/time string *is* UTC.
 725   */
 726
 727   if ((year & 3) == 0 && ((year % 100) != 0 || (year % 400) == 0))
 728     days = leap_days[i] + day - 1;
 729   else
 730     days = normal_days[i] + day - 1;
 731
 732   DEBUG_printf(("    days=%d\n", days));
 733
 734   days += (year - 1970) * 365 +         /* 365 days per year (normally) */
 735           ((year - 1) / 4 - 492) -      /* + leap days */
 736           ((year - 1) / 100 - 19) +     /* - 100 year days */
 737           ((year - 1) / 400 - 4);       /* + 400 year days */
 738
 739   DEBUG_printf(("    days=%d\n", days));
 740
 741   return (days * 86400 + hour * 3600 + min * 60 + sec);
 742 }
 743
 744
 745 /*
 746  * 'httpSeparate()' - Separate a Universal Resource Identifier into its
 747  *                    components.
 748  *
 749  * This function is deprecated; use the httpSeparateURI() function instead.
 750  *
 751  * @deprecated@
 752  */
 753
 754 void
 755 httpSeparate(const char *uri,           /* I - Universal Resource Identifier */
 756              char       *scheme,        /* O - Scheme [32] (http, https, etc.) */
 757              char       *username,      /* O - Username [1024] */
 758              char       *host,          /* O - Hostname [1024] */
 759              int        *port,          /* O - Port number to use */
 760              char       *resource)      /* O - Resource/filename [1024] */
 761 {
 762   httpSeparateURI(HTTP_URI_CODING_ALL, uri, scheme, 32, username,
 763                   HTTP_MAX_URI, host, HTTP_MAX_URI, port, resource,
 764                   HTTP_MAX_URI);
 765 }
 766
 767
 768 /*
 769  * 'httpSeparate2()' - Separate a Universal Resource Identifier into its
 770  *                     components.
 771  *
 772  * This function is deprecated; use the httpSeparateURI() function instead.
 773  *
 774  * @since CUPS 1.1.21@
 775  * @deprecated@
 776  */
 777
 778 void
 779 httpSeparate2(const char *uri,          /* I - Universal Resource Identifier */
 780               char       *scheme,       /* O - Scheme (http, https, etc.) */
 781               int        schemelen,     /* I - Size of scheme buffer */
 782               char       *username,     /* O - Username */
 783               int        usernamelen,   /* I - Size of username buffer */
 784               char       *host,         /* O - Hostname */
 785               int        hostlen,       /* I - Size of hostname buffer */
 786               int        *port,         /* O - Port number to use */
 787               char       *resource,     /* O - Resource/filename */
 788               int        resourcelen)   /* I - Size of resource buffer */
 789 {
 790   httpSeparateURI(HTTP_URI_CODING_ALL, uri, scheme, schemelen, username,
 791                   usernamelen, host, hostlen, port, resource, resourcelen);
 792 }
 793
 794
 795 /*
 796  * 'httpSeparateURI()' - Separate a Universal Resource Identifier into its
 797  *                       components.
 798  *
 799  * @since CUPS 1.2@
 800  */
 801
 802 http_uri_status_t                       /* O - Result of separation */
 803 httpSeparateURI(
 804     http_uri_coding_t decoding,         /* I - Decoding flags */
 805     const char        *uri,             /* I - Universal Resource Identifier */
 806     char              *scheme,          /* O - Scheme (http, https, etc.) */
 807     int               schemelen,        /* I - Size of scheme buffer */
 808     char              *username,        /* O - Username */
 809     int               usernamelen,      /* I - Size of username buffer */
 810     char              *host,            /* O - Hostname */
 811     int               hostlen,          /* I - Size of hostname buffer */
 812     int               *port,            /* O - Port number to use */
 813     char              *resource,        /* O - Resource/filename */
 814     int               resourcelen)      /* I - Size of resource buffer */
 815 {
 816   char                  *ptr,           /* Pointer into string... */
 817                         *end;           /* End of string */
 818   const char            *sep;           /* Separator character */
 819   http_uri_status_t     status;         /* Result of separation */
 820
 821
 822  /*
 823   * Initialize everything to blank...
 824   */
 825
 826   if (scheme && schemelen > 0)
 827     *scheme = '\0';
 828
 829   if (username && usernamelen > 0)
 830     *username = '\0';
 831
 832   if (host && hostlen > 0)
 833     *host = '\0';
 834
 835   if (port)
 836     *port = 0;
 837
 838   if (resource && resourcelen > 0)
 839     *resource = '\0';
 840
 841  /*
 842   * Range check input...
 843   */
 844
 845   if (!uri || !port || !scheme || schemelen <= 0 || !username ||
 846       usernamelen <= 0 || !host || hostlen <= 0 || !resource ||
 847       resourcelen <= 0)
 848     return (HTTP_URI_BAD_ARGUMENTS);
 849
 850   if (!*uri)
 851     return (HTTP_URI_BAD_URI);
 852
 853  /*
 854   * Grab the scheme portion of the URI...
 855   */
 856
 857   status = HTTP_URI_OK;
 858
 859   if (!strncmp(uri, "//", 2))
 860   {
 861    /*
 862     * Workaround for HP IPP client bug...
 863     */
 864
 865     strlcpy(scheme, "ipp", schemelen);
 866     status = HTTP_URI_MISSING_SCHEME;
 867   }
 868   else if (*uri == '/')
 869   {
 870    /*
 871     * Filename...
 872     */
 873
 874     strlcpy(scheme, "file", schemelen);
 875     status = HTTP_URI_MISSING_SCHEME;
 876   }
 877   else
 878   {
 879    /*
 880     * Standard URI with scheme...
 881     */
 882
 883     for (ptr = scheme, end = scheme + schemelen - 1;
 884          *uri && *uri != ':' && ptr < end;)
 885       if (isalnum(*uri & 255) || *uri == '-' || *uri == '+' || *uri == '.')
 886         *ptr++ = *uri++;
 887       else
 888         break;
 889
 890     *ptr = '\0';
 891
 892     if (*uri != ':')
 893     {
 894       *scheme = '\0';
 895       return (HTTP_URI_BAD_SCHEME);
 896     }
 897
 898     uri ++;
 899   }
 900
 901  /*
 902   * Set the default port number...
 903   */
 904
 905   if (!strcmp(scheme, "http"))
 906     *port = 80;
 907   else if (!strcmp(scheme, "https"))
 908     *port = 443;
 909   else if (!strcmp(scheme, "ipp"))
 910     *port = 631;
 911   else if (!strcasecmp(scheme, "lpd"))
 912     *port = 515;
 913   else if (!strcmp(scheme, "socket"))   /* Not yet registered with IANA... */
 914     *port = 9100;
 915   else if (strcmp(scheme, "file") && strcmp(scheme, "mailto"))
 916     status = HTTP_URI_UNKNOWN_SCHEME;
 917
 918  /*
 919   * Now see if we have a hostname...
 920   */
 921
 922   if (!strncmp(uri, "//", 2))
 923   {
 924    /*
 925     * Yes, extract it...
 926     */
 927
 928     uri += 2;
 929
 930    /*
 931     * Grab the username, if any...
 932     */
 933
 934     if ((sep = strpbrk(uri, "@/")) != NULL && *sep == '@')
 935     {
 936      /*
 937       * Get a username:password combo...
 938       */
 939
 940       uri = http_copy_decode(username, uri, usernamelen, "@",
 941                              decoding & HTTP_URI_CODING_USERNAME);
 942
 943       if (!uri)
 944       {
 945         *username = '\0';
 946         return (HTTP_URI_BAD_USERNAME);
 947       }
 948
 949       uri ++;
 950     }
 951
 952    /*
 953     * Then the hostname/IP address...
 954     */
 955
 956     if (*uri == '[')
 957     {
 958      /*
 959       * Grab IPv6 address...
 960       */
 961
 962       uri ++;
 963       if (!strncmp(uri, "v1.", 3))
 964         uri += 3;                       /* Skip IPvN leader... */
 965
 966       uri = http_copy_decode(host, uri, hostlen, "]",
 967                              decoding & HTTP_URI_CODING_HOSTNAME);
 968
 969       if (!uri)
 970       {
 971         *host = '\0';
 972         return (HTTP_URI_BAD_HOSTNAME);
 973       }
 974
 975      /*
 976       * Validate value...
 977       */
 978
 979       if (*uri != ']')
 980       {
 981         *host = '\0';
 982         return (HTTP_URI_BAD_HOSTNAME);
 983       }
 984
 985       uri ++;
 986
 987       for (ptr = host; *ptr; ptr ++)
 988         if (*ptr == '+')
 989         {
 990          /*
 991           * Convert zone separator to % and stop here...
 992           */
 993
 994           *ptr = '%';
 995           break;
 996         }
 997         else if (*ptr != ':' && *ptr != '.' && !isxdigit(*ptr & 255))
 998         {
 999           *host = '\0';
1000           return (HTTP_URI_BAD_HOSTNAME);
1001         }
1002     }
1003     else
1004     {
1005      /*
1006       * Validate the hostname or IPv4 address first...
1007       */
1008
1009       for (ptr = (char *)uri; *ptr; ptr ++)
1010         if (strchr(":?/", *ptr))
1011           break;
1012         else if (!strchr("abcdefghijklmnopqrstuvwxyz"
1013                          "ABCDEFGHIJKLMNOPQRSTUVWXYZ"
1014                          "0123456789"
1015                          "-._~"
1016                          "%"
1017                          "!$&'()*+,;=", *ptr))
1018         {
1019           *host = '\0';
1020           return (HTTP_URI_BAD_HOSTNAME);
1021         }
1022
1023      /*
1024       * Then copy the hostname or IPv4 address to the buffer...
1025       */
1026
1027       uri = http_copy_decode(host, uri, hostlen, ":?/",
1028                              decoding & HTTP_URI_CODING_HOSTNAME);
1029
1030       if (!uri)
1031       {
1032         *host = '\0';
1033         return (HTTP_URI_BAD_HOSTNAME);
1034       }
1035     }
1036
1037    /*
1038     * Validate hostname for file scheme - only empty and localhost are
1039     * acceptable.
1040     */
1041
1042     if (!strcmp(scheme, "file") && strcmp(host, "localhost") && host[0])
1043     {
1044       *host = '\0';
1045       return (HTTP_URI_BAD_HOSTNAME);
1046     }
1047
1048    /*
1049     * See if we have a port number...
1050     */
1051
1052     if (*uri == ':')
1053     {
1054      /*
1055       * Yes, collect the port number...
1056       */
1057
1058       *port = strtol(uri + 1, (char **)&uri, 10);
1059
1060       if (*uri != '/' && *uri)
1061       {
1062         *port = 0;
1063         return (HTTP_URI_BAD_PORT);
1064       }
1065     }
1066   }
1067
1068  /*
1069   * The remaining portion is the resource string...
1070   */
1071
1072   if (*uri == '?' || !*uri)
1073   {
1074    /*
1075     * Hostname but no path...
1076     */
1077
1078     status    = HTTP_URI_MISSING_RESOURCE;
1079     *resource = '/';
1080
1081    /*
1082     * Copy any query string...
1083     */
1084
1085     if (*uri == '?')
1086       uri = http_copy_decode(resource + 1, uri, resourcelen - 1, NULL,
1087                              decoding & HTTP_URI_CODING_QUERY);
1088     else
1089       resource[1] = '\0';
1090   }
1091   else
1092   {
1093     uri = http_copy_decode(resource, uri, resourcelen, "?",
1094                            decoding & HTTP_URI_CODING_RESOURCE);
1095
1096     if (uri && *uri == '?')
1097     {
1098      /*
1099       * Concatenate any query string...
1100       */
1101
1102       char *resptr = resource + strlen(resource);
1103
1104       uri = http_copy_decode(resptr, uri, resourcelen - (resptr - resource),
1105                              NULL, decoding & HTTP_URI_CODING_QUERY);
1106     }
1107   }
1108
1109   if (!uri)
1110   {
1111     *resource = '\0';
1112     return (HTTP_URI_BAD_RESOURCE);
1113   }
1114
1115  /*
1116   * Return the URI separation status...
1117   */
1118
1119   return (status);
1120 }
1121
1122
1123 /*
1124  * 'httpStatus()' - Return a short string describing a HTTP status code.
1125  */
1126
1127 const char *                            /* O - String or NULL */
1128 httpStatus(http_status_t status)        /* I - HTTP status code */
1129 {
1130   switch (status)
1131   {
1132     case HTTP_CONTINUE :
1133         return ("Continue");
1134     case HTTP_SWITCHING_PROTOCOLS :
1135         return ("Switching Protocols");
1136     case HTTP_OK :
1137         return ("OK");
1138     case HTTP_CREATED :
1139         return ("Created");
1140     case HTTP_ACCEPTED :
1141         return ("Accepted");
1142     case HTTP_NO_CONTENT :
1143         return ("No Content");
1144     case HTTP_MOVED_PERMANENTLY :
1145         return ("Moved Permanently");
1146     case HTTP_SEE_OTHER :
1147         return ("See Other");
1148     case HTTP_NOT_MODIFIED :
1149         return ("Not Modified");
1150     case HTTP_BAD_REQUEST :
1151         return ("Bad Request");
1152     case HTTP_UNAUTHORIZED :
1153         return ("Unauthorized");
1154     case HTTP_FORBIDDEN :
1155         return ("Forbidden");
1156     case HTTP_NOT_FOUND :
1157         return ("Not Found");
1158     case HTTP_REQUEST_TOO_LARGE :
1159         return ("Request Entity Too Large");
1160     case HTTP_URI_TOO_LONG :
1161         return ("URI Too Long");
1162     case HTTP_UPGRADE_REQUIRED :
1163         return ("Upgrade Required");
1164     case HTTP_NOT_IMPLEMENTED :
1165         return ("Not Implemented");
1166     case HTTP_NOT_SUPPORTED :
1167         return ("Not Supported");
1168     case HTTP_EXPECTATION_FAILED :
1169         return ("Expectation Failed");
1170
1171     default :
1172         return ("Unknown");
1173   }
1174 }
1175
1176
1177 #ifndef HAVE_HSTRERROR
1178 /*
1179  * '_cups_hstrerror()' - hstrerror() emulation function for Solaris and others...
1180  */
1181
1182 const char *                            /* O - Error string */
1183 _cups_hstrerror(int error)              /* I - Error number */
1184 {
1185   static const char * const errors[] =  /* Error strings */
1186                 {
1187                   "OK",
1188                   "Host not found.",
1189                   "Try again.",
1190                   "Unrecoverable lookup error.",
1191                   "No data associated with name."
1192                 };
1193
1194
1195   if (error < 0 || error > 4)
1196     return ("Unknown hostname lookup error.");
1197   else
1198     return (errors[error]);
1199 }
1200 #endif /* !HAVE_HSTRERROR */
1201
1202
1203 /*
1204  * 'http_copy_decode()' - Copy and decode a URI.
1205  */
1206
1207 static const char *                     /* O - New source pointer or NULL on error */
1208 http_copy_decode(char       *dst,       /* O - Destination buffer */
1209                  const char *src,       /* I - Source pointer */
1210                  int        dstsize,    /* I - Destination size */
1211                  const char *term,      /* I - Terminating characters */
1212                  int        decode)     /* I - Decode %-encoded values */
1213 {
1214   char  *ptr,                           /* Pointer into buffer */
1215         *end;                           /* End of buffer */
1216   int   quoted;                         /* Quoted character */
1217
1218
1219  /*
1220   * Copy the src to the destination until we hit a terminating character
1221   * or the end of the string.
1222   */
1223
1224   for (ptr = dst, end = dst + dstsize - 1;
1225        *src && (!term || !strchr(term, *src));
1226        src ++)
1227     if (ptr < end)
1228     {
1229       if (*src == '%' && decode)
1230       {
1231         if (isxdigit(src[1] & 255) && isxdigit(src[2] & 255))
1232         {
1233          /*
1234           * Grab a hex-encoded character...
1235           */
1236
1237           src ++;
1238           if (isalpha(*src))
1239             quoted = (tolower(*src) - 'a' + 10) << 4;
1240           else
1241             quoted = (*src - '0') << 4;
1242
1243           src ++;
1244           if (isalpha(*src))
1245             quoted |= tolower(*src) - 'a' + 10;
1246           else
1247             quoted |= *src - '0';
1248
1249           *ptr++ = quoted;
1250         }
1251         else
1252         {
1253          /*
1254           * Bad hex-encoded character...
1255           */
1256
1257           *ptr = '\0';
1258           return (NULL);
1259         }
1260       }
1261       else
1262         *ptr++ = *src;
1263     }
1264
1265   *ptr = '\0';
1266
1267   return (src);
1268 }
1269
1270
1271 /*
1272  * 'http_copy_encode()' - Copy and encode a URI.
1273  */
1274
1275 static char *                           /* O - End of current URI */
1276 http_copy_encode(char       *dst,       /* O - Destination buffer */
1277                  const char *src,       /* I - Source pointer */
1278                  char       *dstend,    /* I - End of destination buffer */
1279                  const char *reserved,  /* I - Extra reserved characters */
1280                  const char *term,      /* I - Terminating characters */
1281                  int        encode)     /* I - %-encode reserved chars? */
1282 {
1283   static const char *hex = "0123456789ABCDEF";
1284
1285
1286   while (*src && dst < dstend)
1287   {
1288     if (term && *src == *term)
1289       return (dst);
1290
1291     if (encode && (*src == '%' || *src <= ' ' || *src & 128 ||
1292                    (reserved && strchr(reserved, *src))))
1293     {
1294      /*
1295       * Hex encode reserved characters...
1296       */
1297
1298       if ((dst + 2) >= dstend)
1299         break;
1300
1301       *dst++ = '%';
1302       *dst++ = hex[(*src >> 4) & 15];
1303       *dst++ = hex[*src & 15];
1304
1305       src ++;
1306     }
1307     else
1308       *dst++ = *src++;
1309   }
1310
1311   if (*src)
1312     return (NULL);
1313   else
1314     return (dst);
1315 }
1316
1317
1318 /*
1319  * End of "$Id: http-support.c 177 2006-06-21 00:20:03Z jlovell $".
1320  */