cups/http-support.c

   1 /*
   2  * "$Id: http-support.c 6304 2007-02-22 22:06:23Z mike $"
   3  *
   4  *   HTTP support routines for the Common UNIX Printing System (CUPS) scheduler.
   5  *
   6  *   Copyright 1997-2007 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, (int)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
 599     if (outptr < outend)
 600     {
 601       if (inlen > 1)
 602         *outptr ++ = base64[(((in[0] & 255) << 4) | ((in[1] & 255) >> 4)) & 63];
 603       else
 604         *outptr ++ = base64[((in[0] & 255) << 4) & 63];
 605     }
 606
 607     in ++;
 608     inlen --;
 609     if (inlen <= 0)
 610     {
 611       if (outptr < outend)
 612         *outptr ++ = '=';
 613       if (outptr < outend)
 614         *outptr ++ = '=';
 615       break;
 616     }
 617
 618     if (outptr < outend)
 619     {
 620       if (inlen > 1)
 621         *outptr ++ = base64[(((in[0] & 255) << 2) | ((in[1] & 255) >> 6)) & 63];
 622       else
 623         *outptr ++ = base64[((in[0] & 255) << 2) & 63];
 624     }
 625
 626     in ++;
 627     inlen --;
 628     if (inlen <= 0)
 629     {
 630       if (outptr < outend)
 631         *outptr ++ = '=';
 632       break;
 633     }
 634
 635     if (outptr < outend)
 636       *outptr ++ = base64[in[0] & 63];
 637   }
 638
 639   *outptr = '\0';
 640
 641  /*
 642   * Return the encoded string...
 643   */
 644
 645   return (out);
 646 }
 647
 648
 649 /*
 650  * 'httpGetDateString()' - Get a formatted date/time string from a time value.
 651  *
 652  * @deprecated@
 653  */
 654
 655 const char *                            /* O - Date/time string */
 656 httpGetDateString(time_t t)             /* I - UNIX time */
 657 {
 658   _cups_globals_t *cg = _cupsGlobals(); /* Pointer to library globals */
 659
 660
 661   return (httpGetDateString2(t, cg->http_date, sizeof(cg->http_date)));
 662 }
 663
 664
 665 /*
 666  * 'httpGetDateString2()' - Get a formatted date/time string from a time value.
 667  *
 668  * @since CUPS 1.2@
 669  */
 670
 671 const char *                            /* O - Date/time string */
 672 httpGetDateString2(time_t t,            /* I - UNIX time */
 673                    char   *s,           /* I - String buffer */
 674                    int    slen)         /* I - Size of string buffer */
 675 {
 676   struct tm     *tdate;                 /* UNIX date/time data */
 677
 678
 679   tdate = gmtime(&t);
 680   snprintf(s, slen, "%s, %02d %s %d %02d:%02d:%02d GMT",
 681            http_days[tdate->tm_wday], tdate->tm_mday,
 682            http_months[tdate->tm_mon], tdate->tm_year + 1900,
 683            tdate->tm_hour, tdate->tm_min, tdate->tm_sec);
 684
 685   return (s);
 686 }
 687
 688
 689 /*
 690  * 'httpGetDateTime()' - Get a time value from a formatted date/time string.
 691  */
 692
 693 time_t                                  /* O - UNIX time */
 694 httpGetDateTime(const char *s)          /* I - Date/time string */
 695 {
 696   int           i;                      /* Looping var */
 697   char          mon[16];                /* Abbreviated month name */
 698   int           day, year;              /* Day of month and year */
 699   int           hour, min, sec;         /* Time */
 700   int           days;                   /* Number of days since 1970 */
 701   static const int normal_days[] =      /* Days to a month, normal years */
 702                 { 0, 31, 59, 90, 120, 151, 181, 212, 243, 273, 304, 334, 365 };
 703   static const int leap_days[] =        /* Days to a month, leap years */
 704                 { 0, 31, 60, 91, 121, 152, 182, 213, 244, 274, 305, 335, 366 };
 705
 706
 707   DEBUG_printf(("httpGetDateTime(s=\"%s\")\n", s));
 708
 709  /*
 710   * Extract the date and time from the formatted string...
 711   */
 712
 713   if (sscanf(s, "%*s%d%15s%d%d:%d:%d", &day, mon, &year, &hour, &min, &sec) < 6)
 714     return (0);
 715
 716   DEBUG_printf(("    day=%d, mon=\"%s\", year=%d, hour=%d, min=%d, sec=%d\n",
 717                 day, mon, year, hour, min, sec));
 718
 719  /*
 720   * Convert the month name to a number from 0 to 11.
 721   */
 722
 723   for (i = 0; i < 12; i ++)
 724     if (!strcasecmp(mon, http_months[i]))
 725       break;
 726
 727   if (i >= 12)
 728     return (0);
 729
 730   DEBUG_printf(("    i=%d\n", i));
 731
 732  /*
 733   * Now convert the date and time to a UNIX time value in seconds since
 734   * 1970.  We can't use mktime() since the timezone may not be UTC but
 735   * the date/time string *is* UTC.
 736   */
 737
 738   if ((year & 3) == 0 && ((year % 100) != 0 || (year % 400) == 0))
 739     days = leap_days[i] + day - 1;
 740   else
 741     days = normal_days[i] + day - 1;
 742
 743   DEBUG_printf(("    days=%d\n", days));
 744
 745   days += (year - 1970) * 365 +         /* 365 days per year (normally) */
 746           ((year - 1) / 4 - 492) -      /* + leap days */
 747           ((year - 1) / 100 - 19) +     /* - 100 year days */
 748           ((year - 1) / 400 - 4);       /* + 400 year days */
 749
 750   DEBUG_printf(("    days=%d\n", days));
 751
 752   return (days * 86400 + hour * 3600 + min * 60 + sec);
 753 }
 754
 755
 756 /*
 757  * 'httpSeparate()' - Separate a Universal Resource Identifier into its
 758  *                    components.
 759  *
 760  * This function is deprecated; use the httpSeparateURI() function instead.
 761  *
 762  * @deprecated@
 763  */
 764
 765 void
 766 httpSeparate(const char *uri,           /* I - Universal Resource Identifier */
 767              char       *scheme,        /* O - Scheme [32] (http, https, etc.) */
 768              char       *username,      /* O - Username [1024] */
 769              char       *host,          /* O - Hostname [1024] */
 770              int        *port,          /* O - Port number to use */
 771              char       *resource)      /* O - Resource/filename [1024] */
 772 {
 773   httpSeparateURI(HTTP_URI_CODING_ALL, uri, scheme, 32, username,
 774                   HTTP_MAX_URI, host, HTTP_MAX_URI, port, resource,
 775                   HTTP_MAX_URI);
 776 }
 777
 778
 779 /*
 780  * 'httpSeparate2()' - Separate a Universal Resource Identifier into its
 781  *                     components.
 782  *
 783  * This function is deprecated; use the httpSeparateURI() function instead.
 784  *
 785  * @since CUPS 1.1.21@
 786  * @deprecated@
 787  */
 788
 789 void
 790 httpSeparate2(const char *uri,          /* I - Universal Resource Identifier */
 791               char       *scheme,       /* O - Scheme (http, https, etc.) */
 792               int        schemelen,     /* I - Size of scheme buffer */
 793               char       *username,     /* O - Username */
 794               int        usernamelen,   /* I - Size of username buffer */
 795               char       *host,         /* O - Hostname */
 796               int        hostlen,       /* I - Size of hostname buffer */
 797               int        *port,         /* O - Port number to use */
 798               char       *resource,     /* O - Resource/filename */
 799               int        resourcelen)   /* I - Size of resource buffer */
 800 {
 801   httpSeparateURI(HTTP_URI_CODING_ALL, uri, scheme, schemelen, username,
 802                   usernamelen, host, hostlen, port, resource, resourcelen);
 803 }
 804
 805
 806 /*
 807  * 'httpSeparateURI()' - Separate a Universal Resource Identifier into its
 808  *                       components.
 809  *
 810  * @since CUPS 1.2@
 811  */
 812
 813 http_uri_status_t                       /* O - Result of separation */
 814 httpSeparateURI(
 815     http_uri_coding_t decoding,         /* I - Decoding flags */
 816     const char        *uri,             /* I - Universal Resource Identifier */
 817     char              *scheme,          /* O - Scheme (http, https, etc.) */
 818     int               schemelen,        /* I - Size of scheme buffer */
 819     char              *username,        /* O - Username */
 820     int               usernamelen,      /* I - Size of username buffer */
 821     char              *host,            /* O - Hostname */
 822     int               hostlen,          /* I - Size of hostname buffer */
 823     int               *port,            /* O - Port number to use */
 824     char              *resource,        /* O - Resource/filename */
 825     int               resourcelen)      /* I - Size of resource buffer */
 826 {
 827   char                  *ptr,           /* Pointer into string... */
 828                         *end;           /* End of string */
 829   const char            *sep;           /* Separator character */
 830   http_uri_status_t     status;         /* Result of separation */
 831
 832
 833  /*
 834   * Initialize everything to blank...
 835   */
 836
 837   if (scheme && schemelen > 0)
 838     *scheme = '\0';
 839
 840   if (username && usernamelen > 0)
 841     *username = '\0';
 842
 843   if (host && hostlen > 0)
 844     *host = '\0';
 845
 846   if (port)
 847     *port = 0;
 848
 849   if (resource && resourcelen > 0)
 850     *resource = '\0';
 851
 852  /*
 853   * Range check input...
 854   */
 855
 856   if (!uri || !port || !scheme || schemelen <= 0 || !username ||
 857       usernamelen <= 0 || !host || hostlen <= 0 || !resource ||
 858       resourcelen <= 0)
 859     return (HTTP_URI_BAD_ARGUMENTS);
 860
 861   if (!*uri)
 862     return (HTTP_URI_BAD_URI);
 863
 864  /*
 865   * Grab the scheme portion of the URI...
 866   */
 867
 868   status = HTTP_URI_OK;
 869
 870   if (!strncmp(uri, "//", 2))
 871   {
 872    /*
 873     * Workaround for HP IPP client bug...
 874     */
 875
 876     strlcpy(scheme, "ipp", schemelen);
 877     status = HTTP_URI_MISSING_SCHEME;
 878   }
 879   else if (*uri == '/')
 880   {
 881    /*
 882     * Filename...
 883     */
 884
 885     strlcpy(scheme, "file", schemelen);
 886     status = HTTP_URI_MISSING_SCHEME;
 887   }
 888   else
 889   {
 890    /*
 891     * Standard URI with scheme...
 892     */
 893
 894     for (ptr = scheme, end = scheme + schemelen - 1;
 895          *uri && *uri != ':' && ptr < end;)
 896       if (isalnum(*uri & 255) || *uri == '-' || *uri == '+' || *uri == '.')
 897         *ptr++ = *uri++;
 898       else
 899         break;
 900
 901     *ptr = '\0';
 902
 903     if (*uri != ':')
 904     {
 905       *scheme = '\0';
 906       return (HTTP_URI_BAD_SCHEME);
 907     }
 908
 909     uri ++;
 910   }
 911
 912  /*
 913   * Set the default port number...
 914   */
 915
 916   if (!strcmp(scheme, "http"))
 917     *port = 80;
 918   else if (!strcmp(scheme, "https"))
 919     *port = 443;
 920   else if (!strcmp(scheme, "ipp"))
 921     *port = 631;
 922   else if (!strcasecmp(scheme, "lpd"))
 923     *port = 515;
 924   else if (!strcmp(scheme, "socket"))   /* Not yet registered with IANA... */
 925     *port = 9100;
 926   else if (strcmp(scheme, "file") && strcmp(scheme, "mailto"))
 927     status = HTTP_URI_UNKNOWN_SCHEME;
 928
 929  /*
 930   * Now see if we have a hostname...
 931   */
 932
 933   if (!strncmp(uri, "//", 2))
 934   {
 935    /*
 936     * Yes, extract it...
 937     */
 938
 939     uri += 2;
 940
 941    /*
 942     * Grab the username, if any...
 943     */
 944
 945     if ((sep = strpbrk(uri, "@/")) != NULL && *sep == '@')
 946     {
 947      /*
 948       * Get a username:password combo...
 949       */
 950
 951       uri = http_copy_decode(username, uri, usernamelen, "@",
 952                              decoding & HTTP_URI_CODING_USERNAME);
 953
 954       if (!uri)
 955       {
 956         *username = '\0';
 957         return (HTTP_URI_BAD_USERNAME);
 958       }
 959
 960       uri ++;
 961     }
 962
 963    /*
 964     * Then the hostname/IP address...
 965     */
 966
 967     if (*uri == '[')
 968     {
 969      /*
 970       * Grab IPv6 address...
 971       */
 972
 973       uri ++;
 974       if (!strncmp(uri, "v1.", 3))
 975         uri += 3;                       /* Skip IPvN leader... */
 976
 977       uri = http_copy_decode(host, uri, hostlen, "]",
 978                              decoding & HTTP_URI_CODING_HOSTNAME);
 979
 980       if (!uri)
 981       {
 982         *host = '\0';
 983         return (HTTP_URI_BAD_HOSTNAME);
 984       }
 985
 986      /*
 987       * Validate value...
 988       */
 989
 990       if (*uri != ']')
 991       {
 992         *host = '\0';
 993         return (HTTP_URI_BAD_HOSTNAME);
 994       }
 995
 996       uri ++;
 997
 998       for (ptr = host; *ptr; ptr ++)
 999         if (*ptr == '+')
1000         {
1001          /*
1002           * Convert zone separator to % and stop here...
1003           */
1004
1005           *ptr = '%';
1006           break;
1007         }
1008         else if (*ptr != ':' && *ptr != '.' && !isxdigit(*ptr & 255))
1009         {
1010           *host = '\0';
1011           return (HTTP_URI_BAD_HOSTNAME);
1012         }
1013     }
1014     else
1015     {
1016      /*
1017       * Validate the hostname or IPv4 address first...
1018       */
1019
1020       for (ptr = (char *)uri; *ptr; ptr ++)
1021         if (strchr(":?/", *ptr))
1022           break;
1023         else if (!strchr("abcdefghijklmnopqrstuvwxyz"
1024                          "ABCDEFGHIJKLMNOPQRSTUVWXYZ"
1025                          "0123456789"
1026                          "-._~"
1027                          "%"
1028                          "!$&'()*+,;=", *ptr))
1029         {
1030           *host = '\0';
1031           return (HTTP_URI_BAD_HOSTNAME);
1032         }
1033
1034      /*
1035       * Then copy the hostname or IPv4 address to the buffer...
1036       */
1037
1038       uri = http_copy_decode(host, uri, hostlen, ":?/",
1039                              decoding & HTTP_URI_CODING_HOSTNAME);
1040
1041       if (!uri)
1042       {
1043         *host = '\0';
1044         return (HTTP_URI_BAD_HOSTNAME);
1045       }
1046     }
1047
1048    /*
1049     * Validate hostname for file scheme - only empty and localhost are
1050     * acceptable.
1051     */
1052
1053     if (!strcmp(scheme, "file") && strcmp(host, "localhost") && host[0])
1054     {
1055       *host = '\0';
1056       return (HTTP_URI_BAD_HOSTNAME);
1057     }
1058
1059    /*
1060     * See if we have a port number...
1061     */
1062
1063     if (*uri == ':')
1064     {
1065      /*
1066       * Yes, collect the port number...
1067       */
1068
1069       *port = strtol(uri + 1, (char **)&uri, 10);
1070
1071       if (*uri != '/' && *uri)
1072       {
1073         *port = 0;
1074         return (HTTP_URI_BAD_PORT);
1075       }
1076     }
1077   }
1078
1079  /*
1080   * The remaining portion is the resource string...
1081   */
1082
1083   if (*uri == '?' || !*uri)
1084   {
1085    /*
1086     * Hostname but no path...
1087     */
1088
1089     status    = HTTP_URI_MISSING_RESOURCE;
1090     *resource = '/';
1091
1092    /*
1093     * Copy any query string...
1094     */
1095
1096     if (*uri == '?')
1097       uri = http_copy_decode(resource + 1, uri, resourcelen - 1, NULL,
1098                              decoding & HTTP_URI_CODING_QUERY);
1099     else
1100       resource[1] = '\0';
1101   }
1102   else
1103   {
1104     uri = http_copy_decode(resource, uri, resourcelen, "?",
1105                            decoding & HTTP_URI_CODING_RESOURCE);
1106
1107     if (uri && *uri == '?')
1108     {
1109      /*
1110       * Concatenate any query string...
1111       */
1112
1113       char *resptr = resource + strlen(resource);
1114
1115       uri = http_copy_decode(resptr, uri, resourcelen - (int)(resptr - resource),
1116                              NULL, decoding & HTTP_URI_CODING_QUERY);
1117     }
1118   }
1119
1120   if (!uri)
1121   {
1122     *resource = '\0';
1123     return (HTTP_URI_BAD_RESOURCE);
1124   }
1125
1126  /*
1127   * Return the URI separation status...
1128   */
1129
1130   return (status);
1131 }
1132
1133
1134 /*
1135  * 'httpStatus()' - Return a short string describing a HTTP status code.
1136  */
1137
1138 const char *                            /* O - String or NULL */
1139 httpStatus(http_status_t status)        /* I - HTTP status code */
1140 {
1141   switch (status)
1142   {
1143     case HTTP_CONTINUE :
1144         return ("Continue");
1145     case HTTP_SWITCHING_PROTOCOLS :
1146         return ("Switching Protocols");
1147     case HTTP_OK :
1148         return ("OK");
1149     case HTTP_CREATED :
1150         return ("Created");
1151     case HTTP_ACCEPTED :
1152         return ("Accepted");
1153     case HTTP_NO_CONTENT :
1154         return ("No Content");
1155     case HTTP_MOVED_PERMANENTLY :
1156         return ("Moved Permanently");
1157     case HTTP_SEE_OTHER :
1158         return ("See Other");
1159     case HTTP_NOT_MODIFIED :
1160         return ("Not Modified");
1161     case HTTP_BAD_REQUEST :
1162         return ("Bad Request");
1163     case HTTP_UNAUTHORIZED :
1164         return ("Unauthorized");
1165     case HTTP_FORBIDDEN :
1166         return ("Forbidden");
1167     case HTTP_NOT_FOUND :
1168         return ("Not Found");
1169     case HTTP_REQUEST_TOO_LARGE :
1170         return ("Request Entity Too Large");
1171     case HTTP_URI_TOO_LONG :
1172         return ("URI Too Long");
1173     case HTTP_UPGRADE_REQUIRED :
1174         return ("Upgrade Required");
1175     case HTTP_NOT_IMPLEMENTED :
1176         return ("Not Implemented");
1177     case HTTP_NOT_SUPPORTED :
1178         return ("Not Supported");
1179     case HTTP_EXPECTATION_FAILED :
1180         return ("Expectation Failed");
1181
1182     default :
1183         return ("Unknown");
1184   }
1185 }
1186
1187
1188 #ifndef HAVE_HSTRERROR
1189 /*
1190  * '_cups_hstrerror()' - hstrerror() emulation function for Solaris and others...
1191  */
1192
1193 const char *                            /* O - Error string */
1194 _cups_hstrerror(int error)              /* I - Error number */
1195 {
1196   static const char * const errors[] =  /* Error strings */
1197                 {
1198                   "OK",
1199                   "Host not found.",
1200                   "Try again.",
1201                   "Unrecoverable lookup error.",
1202                   "No data associated with name."
1203                 };
1204
1205
1206   if (error < 0 || error > 4)
1207     return ("Unknown hostname lookup error.");
1208   else
1209     return (errors[error]);
1210 }
1211 #endif /* !HAVE_HSTRERROR */
1212
1213
1214 /*
1215  * 'http_copy_decode()' - Copy and decode a URI.
1216  */
1217
1218 static const char *                     /* O - New source pointer or NULL on error */
1219 http_copy_decode(char       *dst,       /* O - Destination buffer */
1220                  const char *src,       /* I - Source pointer */
1221                  int        dstsize,    /* I - Destination size */
1222                  const char *term,      /* I - Terminating characters */
1223                  int        decode)     /* I - Decode %-encoded values */
1224 {
1225   char  *ptr,                           /* Pointer into buffer */
1226         *end;                           /* End of buffer */
1227   int   quoted;                         /* Quoted character */
1228
1229
1230  /*
1231   * Copy the src to the destination until we hit a terminating character
1232   * or the end of the string.
1233   */
1234
1235   for (ptr = dst, end = dst + dstsize - 1;
1236        *src && (!term || !strchr(term, *src));
1237        src ++)
1238     if (ptr < end)
1239     {
1240       if (*src == '%' && decode)
1241       {
1242         if (isxdigit(src[1] & 255) && isxdigit(src[2] & 255))
1243         {
1244          /*
1245           * Grab a hex-encoded character...
1246           */
1247
1248           src ++;
1249           if (isalpha(*src))
1250             quoted = (tolower(*src) - 'a' + 10) << 4;
1251           else
1252             quoted = (*src - '0') << 4;
1253
1254           src ++;
1255           if (isalpha(*src))
1256             quoted |= tolower(*src) - 'a' + 10;
1257           else
1258             quoted |= *src - '0';
1259
1260           *ptr++ = quoted;
1261         }
1262         else
1263         {
1264          /*
1265           * Bad hex-encoded character...
1266           */
1267
1268           *ptr = '\0';
1269           return (NULL);
1270         }
1271       }
1272       else
1273         *ptr++ = *src;
1274     }
1275
1276   *ptr = '\0';
1277
1278   return (src);
1279 }
1280
1281
1282 /*
1283  * 'http_copy_encode()' - Copy and encode a URI.
1284  */
1285
1286 static char *                           /* O - End of current URI */
1287 http_copy_encode(char       *dst,       /* O - Destination buffer */
1288                  const char *src,       /* I - Source pointer */
1289                  char       *dstend,    /* I - End of destination buffer */
1290                  const char *reserved,  /* I - Extra reserved characters */
1291                  const char *term,      /* I - Terminating characters */
1292                  int        encode)     /* I - %-encode reserved chars? */
1293 {
1294   static const char *hex = "0123456789ABCDEF";
1295
1296
1297   while (*src && dst < dstend)
1298   {
1299     if (term && *src == *term)
1300       return (dst);
1301
1302     if (encode && (*src == '%' || *src <= ' ' || *src & 128 ||
1303                    (reserved && strchr(reserved, *src))))
1304     {
1305      /*
1306       * Hex encode reserved characters...
1307       */
1308
1309       if ((dst + 2) >= dstend)
1310         break;
1311
1312       *dst++ = '%';
1313       *dst++ = hex[(*src >> 4) & 15];
1314       *dst++ = hex[*src & 15];
1315
1316       src ++;
1317     }
1318     else
1319       *dst++ = *src++;
1320   }
1321
1322   if (*src)
1323     return (NULL);
1324   else
1325     return (dst);
1326 }
1327
1328
1329 /*
1330  * End of "$Id: http-support.c 6304 2007-02-22 22:06:23Z mike $".
1331  */