dlls/winhttp/url.c

   1 /*
   2  * Copyright 2008 Hans Leidekker for CodeWeavers
   3  *
   4  * This library is free software; you can redistribute it and/or
   5  * modify it under the terms of the GNU Lesser General Public
   6  * License as published by the Free Software Foundation; either
   7  * version 2.1 of the License, or (at your option) any later version.
   8  *
   9  * This library is distributed in the hope that it will be useful,
  10  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  11  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
  12  * Lesser General Public License for more details.
  13  *
  14  * You should have received a copy of the GNU Lesser General Public
  15  * License along with this library; if not, write to the Free Software
  16  * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301, USA
  17  */
  18
  19 #include <wchar.h>
  20 #include "windef.h"
  21 #include "winbase.h"
  22 #include "ws2tcpip.h"
  23 #include "winreg.h"
  24 #include "winhttp.h"
  25 #include "shlwapi.h"
  26
  27 #include "wine/debug.h"
  28 #include "winhttp_private.h"
  29
  30 WINE_DEFAULT_DEBUG_CHANNEL(winhttp);
  31
  32 struct url_component
  33 {
  34     WCHAR **str;
  35     DWORD  *len;
  36 };
  37
  38 static DWORD set_component( struct url_component *comp, WCHAR *value, DWORD len, DWORD flags, BOOL *overflow )
  39 {
  40     if (*comp->str && !*comp->len) return ERROR_INVALID_PARAMETER;
  41     if (!*comp->len) return ERROR_SUCCESS;
  42     if (!*comp->str)
  43     {
  44         if (len && *comp->len && (flags & (ICU_DECODE|ICU_ESCAPE))) return ERROR_INVALID_PARAMETER;
  45         *comp->str = value;
  46         *comp->len = len;
  47     }
  48     else
  49     {
  50         if (len >= *comp->len)
  51         {
  52             *comp->len = len + 1;
  53             *overflow = TRUE;
  54             return ERROR_SUCCESS;
  55         }
  56         memcpy( *comp->str, value, len * sizeof(WCHAR) );
  57         (*comp->str)[len] = 0;
  58         *comp->len = len;
  59     }
  60     return ERROR_SUCCESS;
  61 }
  62
  63 static WCHAR *decode_url( LPCWSTR url, DWORD *len )
  64 {
  65     const WCHAR *p = url;
  66     WCHAR hex[3], *q, *ret;
  67
  68     if (!(ret = malloc( *len * sizeof(WCHAR) ))) return NULL;
  69     q = ret;
  70     while (*len > 0)
  71     {
  72         if (p[0] == '%' && iswxdigit( p[1] ) && iswxdigit( p[2] ))
  73         {
  74             hex[0] = p[1];
  75             hex[1] = p[2];
  76             hex[2] = 0;
  77             *q++ = wcstol( hex, NULL, 16 );
  78             p += 3;
  79             *len -= 3;
  80         }
  81         else
  82         {
  83             *q++ = *p++;
  84             *len -= 1;
  85         }
  86     }
  87     *len = q - ret;
  88     return ret;
  89 }
  90
  91 static inline BOOL need_escape( WCHAR ch )
  92 {
  93     const WCHAR *p = L" \"#%<>[\\]^`{|}~";
  94
  95     if (ch <= 31 || ch >= 127) return TRUE;
  96     while (*p)
  97     {
  98         if (ch == *p++) return TRUE;
  99     }
 100     return FALSE;
 101 }
 102
 103 static BOOL escape_string( const WCHAR *src, DWORD src_len, WCHAR *dst, DWORD *dst_len )
 104 {
 105     static const WCHAR hex[] = L"0123456789ABCDEF";
 106     WCHAR *p = dst;
 107     DWORD i;
 108
 109     *dst_len = src_len;
 110     for (i = 0; i < src_len; i++)
 111     {
 112         if (src[i] > 0xff) return FALSE;
 113         if (need_escape( src[i] ))
 114         {
 115             if (dst)
 116             {
 117                 p[0] = '%';
 118                 p[1] = hex[(src[i] >> 4) & 0xf];
 119                 p[2] = hex[src[i] & 0xf];
 120                 p += 3;
 121             }
 122             *dst_len += 2;
 123         }
 124         else if (dst) *p++ = src[i];
 125     }
 126
 127     if (dst) dst[*dst_len] = 0;
 128     return TRUE;
 129 }
 130
 131 static DWORD escape_url( const WCHAR *url, DWORD *len, WCHAR **ret )
 132 {
 133     const WCHAR *p;
 134     DWORD len_base, len_path;
 135
 136     if ((p = wcsrchr( url, '/' )))
 137     {
 138         len_base = p - url;
 139         if (!escape_string( p, *len - len_base, NULL, &len_path )) return ERROR_INVALID_PARAMETER;
 140     }
 141     else
 142     {
 143         len_base = *len;
 144         len_path = 0;
 145     }
 146
 147     if (!(*ret = malloc( (len_base + len_path + 1) * sizeof(WCHAR) ))) return ERROR_OUTOFMEMORY;
 148     memcpy( *ret, url, len_base * sizeof(WCHAR) );
 149
 150     if (p) escape_string( p, *len - (p - url), *ret + len_base, &len_path );
 151     (*ret)[len_base + len_path] = 0;
 152
 153     *len = len_base + len_path;
 154     return ERROR_SUCCESS;
 155 }
 156
 157 static DWORD parse_port( const WCHAR *str, DWORD len, INTERNET_PORT *ret )
 158 {
 159     const WCHAR *p = str;
 160     DWORD port = 0;
 161     while (len && '0' <= *p && *p <= '9')
 162     {
 163         if ((port = port * 10 + *p - '0') > 65535) return ERROR_WINHTTP_INVALID_URL;
 164         p++; len--;
 165     }
 166     *ret = port;
 167     return ERROR_SUCCESS;
 168 }
 169
 170 /***********************************************************************
 171  *          WinHttpCrackUrl (winhttp.@)
 172  */
 173 BOOL WINAPI WinHttpCrackUrl( const WCHAR *url, DWORD len, DWORD flags, URL_COMPONENTSW *uc )
 174 {
 175     WCHAR *p, *q, *r, *url_transformed = NULL;
 176     INTERNET_SCHEME scheme_number = 0;
 177     struct url_component scheme, username, password, hostname, path, extra;
 178     BOOL overflow = FALSE;
 179     DWORD err;
 180
 181     TRACE( "%s, %lu, %#lx, %p\n", debugstr_wn(url, len), len, flags, uc );
 182
 183     if (!url || !uc || uc->dwStructSize != sizeof(*uc))
 184     {
 185         SetLastError( ERROR_INVALID_PARAMETER );
 186         return FALSE;
 187     }
 188     if (!len) len = lstrlenW( url );
 189
 190     if (flags & ICU_ESCAPE)
 191     {
 192         if ((err = escape_url( url, &len, &url_transformed )))
 193         {
 194             SetLastError( err );
 195             return FALSE;
 196         }
 197         url = url_transformed;
 198     }
 199     else if (flags & ICU_DECODE)
 200     {
 201         if (!(url_transformed = decode_url( url, &len )))
 202         {
 203             SetLastError( ERROR_OUTOFMEMORY );
 204             return FALSE;
 205         }
 206         url = url_transformed;
 207     }
 208     if (!(p = wcschr( url, ':' )))
 209     {
 210         SetLastError( ERROR_WINHTTP_UNRECOGNIZED_SCHEME );
 211         free( url_transformed );
 212         return FALSE;
 213     }
 214     if (p - url == 4 && !wcsnicmp( url, L"http", 4 )) scheme_number = INTERNET_SCHEME_HTTP;
 215     else if (p - url == 5 && !wcsnicmp( url, L"https", 5 )) scheme_number = INTERNET_SCHEME_HTTPS;
 216     else
 217     {
 218         err = ERROR_WINHTTP_UNRECOGNIZED_SCHEME;
 219         goto exit;
 220     }
 221
 222     scheme.str = &uc->lpszScheme;
 223     scheme.len = &uc->dwSchemeLength;
 224
 225     if ((err = set_component( &scheme, (WCHAR *)url, p - url, flags, &overflow ))) goto exit;
 226
 227     p++; /* skip ':' */
 228     if (!p[0] || p[0] != '/' || p[1] != '/')
 229     {
 230         err = ERROR_WINHTTP_INVALID_URL;
 231         goto exit;
 232     }
 233     p += 2;
 234     if (!p[0])
 235     {
 236         err = ERROR_WINHTTP_INVALID_URL;
 237         goto exit;
 238     }
 239
 240     username.str = &uc->lpszUserName;
 241     username.len = &uc->dwUserNameLength;
 242
 243     password.str = &uc->lpszPassword;
 244     password.len = &uc->dwPasswordLength;
 245
 246     if ((q = wmemchr( p, '@', len - (p - url) )) && !(wmemchr( p, '/', q - p )))
 247     {
 248
 249         if ((r = wmemchr( p, ':', q - p )))
 250         {
 251             if ((err = set_component( &username, p, r - p, flags, &overflow ))) goto exit;
 252             r++;
 253             if ((err = set_component( &password, r, q - r, flags, &overflow ))) goto exit;
 254         }
 255         else
 256         {
 257             if ((err = set_component( &username, p, q - p, flags, &overflow ))) goto exit;
 258             if ((err = set_component( &password, NULL, 0, flags, &overflow ))) goto exit;
 259         }
 260         p = q + 1;
 261     }
 262     else
 263     {
 264         if ((err = set_component( &username, NULL, 0, flags, &overflow ))) goto exit;
 265         if ((err = set_component( &password, NULL, 0, flags, &overflow ))) goto exit;
 266     }
 267
 268     hostname.str = &uc->lpszHostName;
 269     hostname.len = &uc->dwHostNameLength;
 270
 271     path.str = &uc->lpszUrlPath;
 272     path.len = &uc->dwUrlPathLength;
 273
 274     extra.str = &uc->lpszExtraInfo;
 275     extra.len = &uc->dwExtraInfoLength;
 276
 277     if ((q = wmemchr( p, '/', len - (p - url) )))
 278     {
 279         if ((r = wmemchr( p, ':', q - p )))
 280         {
 281             if ((err = set_component( &hostname, p, r - p, flags, &overflow ))) goto exit;
 282             r++;
 283             if ((err = parse_port( r, q - r, &uc->nPort ))) goto exit;
 284         }
 285         else
 286         {
 287             if ((err = set_component( &hostname, p, q - p, flags, &overflow ))) goto exit;
 288             if (scheme_number == INTERNET_SCHEME_HTTP) uc->nPort = INTERNET_DEFAULT_HTTP_PORT;
 289             if (scheme_number == INTERNET_SCHEME_HTTPS) uc->nPort = INTERNET_DEFAULT_HTTPS_PORT;
 290         }
 291
 292         if ((r = wmemchr( q, '?', len - (q - url) )))
 293         {
 294             if (*extra.len)
 295             {
 296                 if ((err = set_component( &path, q, r - q, flags, &overflow ))) goto exit;
 297                 if ((err = set_component( &extra, r, len - (r - url), flags, &overflow ))) goto exit;
 298             }
 299             else if ((err = set_component( &path, q, len - (q - url), flags, &overflow ))) goto exit;
 300         }
 301         else
 302         {
 303             if ((err = set_component( &path, q, len - (q - url), flags, &overflow ))) goto exit;
 304             if ((err = set_component( &extra, (WCHAR *)url + len, 0, flags, &overflow ))) goto exit;
 305         }
 306     }
 307     else
 308     {
 309         if ((r = wmemchr( p, ':', len - (p - url) )))
 310         {
 311             if ((err = set_component( &hostname, p, r - p, flags, &overflow ))) goto exit;
 312             r++;
 313             if ((err = parse_port( r, len - (r - url), &uc->nPort ))) goto exit;
 314         }
 315         else
 316         {
 317             if ((err = set_component( &hostname, p, len - (p - url), flags, &overflow ))) goto exit;
 318             if (scheme_number == INTERNET_SCHEME_HTTP) uc->nPort = INTERNET_DEFAULT_HTTP_PORT;
 319             if (scheme_number == INTERNET_SCHEME_HTTPS) uc->nPort = INTERNET_DEFAULT_HTTPS_PORT;
 320         }
 321         if ((err = set_component( &path, (WCHAR *)url + len, 0, flags, &overflow ))) goto exit;
 322         if ((err = set_component( &extra, (WCHAR *)url + len, 0, flags, &overflow ))) goto exit;
 323     }
 324
 325     TRACE("scheme(%s) host(%s) port(%d) path(%s) extra(%s)\n", debugstr_wn(*scheme.str, *scheme.len),
 326           debugstr_wn(*hostname.str, *hostname.len ), uc->nPort, debugstr_wn(*path.str, *path.len),
 327           debugstr_wn(*extra.str, *extra.len));
 328
 329 exit:
 330     if (!err)
 331     {
 332         if (overflow) err = ERROR_INSUFFICIENT_BUFFER;
 333         uc->nScheme = scheme_number;
 334     }
 335     free( url_transformed );
 336     SetLastError( err );
 337     return !err;
 338 }
 339
 340 static INTERNET_SCHEME get_scheme( const WCHAR *scheme, DWORD len )
 341 {
 342     if (!wcsncmp( scheme, L"http", len )) return INTERNET_SCHEME_HTTP;
 343     if (!wcsncmp( scheme, L"https", len )) return INTERNET_SCHEME_HTTPS;
 344     return 0;
 345 }
 346
 347 static const WCHAR *get_scheme_string( INTERNET_SCHEME scheme )
 348 {
 349     if (scheme == INTERNET_SCHEME_HTTP) return L"http";
 350     if (scheme == INTERNET_SCHEME_HTTPS) return L"https";
 351     return NULL;
 352 }
 353
 354 static BOOL uses_default_port( INTERNET_SCHEME scheme, INTERNET_PORT port )
 355 {
 356     if ((scheme == INTERNET_SCHEME_HTTP) && (port == INTERNET_DEFAULT_HTTP_PORT)) return TRUE;
 357     if ((scheme == INTERNET_SCHEME_HTTPS) && (port == INTERNET_DEFAULT_HTTPS_PORT)) return TRUE;
 358     return FALSE;
 359 }
 360
 361 static DWORD get_comp_length( DWORD len, DWORD flags, WCHAR *comp )
 362 {
 363     DWORD ret;
 364     unsigned int i;
 365
 366     ret = len ? len : lstrlenW( comp );
 367     if (!(flags & ICU_ESCAPE)) return ret;
 368     for (i = 0; i < len; i++) if (need_escape( comp[i] )) ret += 2;
 369     return ret;
 370 }
 371
 372 static BOOL get_url_length( URL_COMPONENTS *uc, DWORD flags, DWORD *len )
 373 {
 374     INTERNET_SCHEME scheme;
 375
 376     *len = 0;
 377     if (uc->lpszScheme)
 378     {
 379         DWORD scheme_len = get_comp_length( uc->dwSchemeLength, 0, uc->lpszScheme );
 380         *len += scheme_len;
 381         scheme = get_scheme( uc->lpszScheme, scheme_len );
 382     }
 383     else
 384     {
 385         scheme = uc->nScheme;
 386         if (!scheme) scheme = INTERNET_SCHEME_HTTP;
 387         *len += lstrlenW( get_scheme_string( scheme ) );
 388     }
 389     *len += 3; /* "://" */
 390
 391     if (uc->lpszUserName)
 392     {
 393         *len += get_comp_length( uc->dwUserNameLength, 0, uc->lpszUserName );
 394         *len += 1; /* "@" */
 395     }
 396     else
 397     {
 398         if (uc->lpszPassword)
 399         {
 400             SetLastError( ERROR_INVALID_PARAMETER );
 401             return FALSE;
 402         }
 403     }
 404     if (uc->lpszPassword)
 405     {
 406         *len += 1; /* ":" */
 407         *len += get_comp_length( uc->dwPasswordLength, 0, uc->lpszPassword );
 408     }
 409     if (uc->lpszHostName)
 410     {
 411         *len += get_comp_length( uc->dwHostNameLength, 0, uc->lpszHostName );
 412
 413         if (!uses_default_port( scheme, uc->nPort ))
 414         {
 415             WCHAR port[sizeof("65535")];
 416
 417             *len += swprintf( port, ARRAY_SIZE(port), L"%u", uc->nPort );
 418             *len += 1; /* ":" */
 419         }
 420         if (uc->lpszUrlPath && *uc->lpszUrlPath != '/') *len += 1; /* '/' */
 421     }
 422     if (uc->lpszUrlPath) *len += get_comp_length( uc->dwUrlPathLength, flags, uc->lpszUrlPath );
 423     if (uc->lpszExtraInfo) *len += get_comp_length( uc->dwExtraInfoLength, flags, uc->lpszExtraInfo );
 424     return TRUE;
 425 }
 426
 427 /***********************************************************************
 428  *          WinHttpCreateUrl (winhttp.@)
 429  */
 430 BOOL WINAPI WinHttpCreateUrl( URL_COMPONENTS *uc, DWORD flags, WCHAR *url, DWORD *required )
 431 {
 432     DWORD len, len_escaped;
 433     INTERNET_SCHEME scheme;
 434
 435     TRACE( "%p, %#lx, %p, %p\n", uc, flags, url, required );
 436
 437     if (!uc || uc->dwStructSize != sizeof(URL_COMPONENTS) || !required)
 438     {
 439         SetLastError( ERROR_INVALID_PARAMETER );
 440         return FALSE;
 441     }
 442
 443     if (!get_url_length( uc, flags, &len )) return FALSE;
 444
 445     if (*required < len)
 446     {
 447         *required = len + 1;
 448         SetLastError( ERROR_INSUFFICIENT_BUFFER );
 449         return FALSE;
 450     }
 451     if (!url)
 452     {
 453         SetLastError( ERROR_INVALID_PARAMETER );
 454         return FALSE;
 455     }
 456
 457     url[0] = 0;
 458     *required = len;
 459     if (uc->lpszScheme)
 460     {
 461         len = get_comp_length( uc->dwSchemeLength, 0, uc->lpszScheme );
 462         memcpy( url, uc->lpszScheme, len * sizeof(WCHAR) );
 463         url += len;
 464
 465         scheme = get_scheme( uc->lpszScheme, len );
 466     }
 467     else
 468     {
 469         const WCHAR *schemeW;
 470         scheme = uc->nScheme;
 471
 472         if (!scheme) scheme = INTERNET_SCHEME_HTTP;
 473
 474         schemeW = get_scheme_string( scheme );
 475         len = lstrlenW( schemeW );
 476         memcpy( url, schemeW, len * sizeof(WCHAR) );
 477         url += len;
 478     }
 479
 480     *url++ = ':';
 481     *url++ = '/';
 482     *url++ = '/';
 483
 484     if (uc->lpszUserName)
 485     {
 486         len = get_comp_length( uc->dwUserNameLength, 0, uc->lpszUserName );
 487         memcpy( url, uc->lpszUserName, len * sizeof(WCHAR) );
 488         url += len;
 489
 490         if (uc->lpszPassword)
 491         {
 492             *url++ = ':';
 493             len = get_comp_length( uc->dwPasswordLength, 0, uc->lpszPassword );
 494             memcpy( url, uc->lpszPassword, len * sizeof(WCHAR) );
 495             url += len;
 496         }
 497         *url++ = '@';
 498     }
 499     if (uc->lpszHostName)
 500     {
 501         len = get_comp_length( uc->dwHostNameLength, 0, uc->lpszHostName );
 502         memcpy( url, uc->lpszHostName, len * sizeof(WCHAR) );
 503         url += len;
 504
 505         if (!uses_default_port( scheme, uc->nPort ))
 506         {
 507             *url++ = ':';
 508             url += swprintf( url, sizeof("65535"), L"%u", uc->nPort );
 509         }
 510
 511         /* add slash between hostname and path if necessary */
 512         if (uc->lpszUrlPath && *uc->lpszUrlPath != '/')
 513         {
 514             *url++ = '/';
 515         }
 516     }
 517     if (uc->lpszUrlPath)
 518     {
 519         len = get_comp_length( uc->dwUrlPathLength, 0, uc->lpszUrlPath );
 520         if (flags & ICU_ESCAPE)
 521         {
 522             if (!escape_string( uc->lpszUrlPath, len, url, &len_escaped ))
 523             {
 524                 SetLastError( ERROR_INVALID_PARAMETER );
 525                 return FALSE;
 526             }
 527             url += len_escaped;
 528         }
 529         else
 530         {
 531             memcpy( url, uc->lpszUrlPath, len * sizeof(WCHAR) );
 532             url += len;
 533         }
 534     }
 535     if (uc->lpszExtraInfo)
 536     {
 537         len = get_comp_length( uc->dwExtraInfoLength, 0, uc->lpszExtraInfo );
 538         if (flags & ICU_ESCAPE)
 539         {
 540             if (!escape_string( uc->lpszExtraInfo, len, url, &len_escaped ))
 541             {
 542                 SetLastError( ERROR_INVALID_PARAMETER );
 543                 return FALSE;
 544             }
 545             url += len_escaped;
 546         }
 547         else
 548         {
 549             memcpy( url, uc->lpszExtraInfo, len * sizeof(WCHAR) );
 550             url += len;
 551         }
 552     }
 553     *url = 0;
 554     SetLastError( ERROR_SUCCESS );
 555     return TRUE;
 556 }