dlls/msi/string.c

   1 /*
   2  * String Table Functions
   3  *
   4  * Copyright 2002-2004, Mike McCormack for CodeWeavers
   5  * Copyright 2007 Robert Shearman for CodeWeavers
   6  * Copyright 2010 Hans Leidekker for CodeWeavers
   7  *
   8  * This library is free software; you can redistribute it and/or
   9  * modify it under the terms of the GNU Lesser General Public
  10  * License as published by the Free Software Foundation; either
  11  * version 2.1 of the License, or (at your option) any later version.
  12  *
  13  * This library is distributed in the hope that it will be useful,
  14  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  15  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
  16  * Lesser General Public License for more details.
  17  *
  18  * You should have received a copy of the GNU Lesser General Public
  19  * License along with this library; if not, write to the Free Software
  20  * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301, USA
  21  */
  22
  23 #define COBJMACROS
  24
  25 #include <stdarg.h>
  26 #include <assert.h>
  27
  28 #include "windef.h"
  29 #include "winbase.h"
  30 #include "winerror.h"
  31 #include "wine/debug.h"
  32 #include "wine/unicode.h"
  33 #include "msi.h"
  34 #include "msiquery.h"
  35 #include "objbase.h"
  36 #include "objidl.h"
  37 #include "msipriv.h"
  38 #include "winnls.h"
  39
  40 #include "query.h"
  41
  42 WINE_DEFAULT_DEBUG_CHANNEL(msidb);
  43
  44 struct msistring
  45 {
  46     USHORT persistent_refcount;
  47     USHORT nonpersistent_refcount;
  48     WCHAR *data;
  49     int    len;
  50 };
  51
  52 struct string_table
  53 {
  54     UINT maxcount;         /* the number of strings */
  55     UINT freeslot;
  56     UINT codepage;
  57     UINT sortcount;
  58     struct msistring *strings; /* an array of strings */
  59     UINT *sorted;              /* index */
  60 };
  61
  62 static BOOL validate_codepage( UINT codepage )
  63 {
  64     if (codepage != CP_ACP && !IsValidCodePage( codepage ))
  65     {
  66         WARN("invalid codepage %u\n", codepage);
  67         return FALSE;
  68     }
  69     return TRUE;
  70 }
  71
  72 static string_table *init_stringtable( int entries, UINT codepage )
  73 {
  74     string_table *st;
  75
  76     if (!validate_codepage( codepage ))
  77         return NULL;
  78
  79     st = msi_alloc( sizeof (string_table) );
  80     if( !st )
  81         return NULL;
  82     if( entries < 1 )
  83         entries = 1;
  84
  85     st->strings = msi_alloc_zero( sizeof(struct msistring) * entries );
  86     if( !st->strings )
  87     {
  88         msi_free( st );
  89         return NULL;
  90     }
  91
  92     st->sorted = msi_alloc( sizeof (UINT) * entries );
  93     if( !st->sorted )
  94     {
  95         msi_free( st->strings );
  96         msi_free( st );
  97         return NULL;
  98     }
  99
 100     st->maxcount = entries;
 101     st->freeslot = 1;
 102     st->codepage = codepage;
 103     st->sortcount = 0;
 104
 105     return st;
 106 }
 107
 108 VOID msi_destroy_stringtable( string_table *st )
 109 {
 110     UINT i;
 111
 112     for( i=0; i<st->maxcount; i++ )
 113     {
 114         if( st->strings[i].persistent_refcount ||
 115             st->strings[i].nonpersistent_refcount )
 116             msi_free( st->strings[i].data );
 117     }
 118     msi_free( st->strings );
 119     msi_free( st->sorted );
 120     msi_free( st );
 121 }
 122
 123 static int st_find_free_entry( string_table *st )
 124 {
 125     UINT i, sz, *s;
 126     struct msistring *p;
 127
 128     TRACE("%p\n", st);
 129
 130     if( st->freeslot )
 131     {
 132         for( i = st->freeslot; i < st->maxcount; i++ )
 133             if( !st->strings[i].persistent_refcount &&
 134                 !st->strings[i].nonpersistent_refcount )
 135                 return i;
 136     }
 137     for( i = 1; i < st->maxcount; i++ )
 138         if( !st->strings[i].persistent_refcount &&
 139             !st->strings[i].nonpersistent_refcount )
 140             return i;
 141
 142     /* dynamically resize */
 143     sz = st->maxcount + 1 + st->maxcount/2;
 144     p = msi_realloc_zero( st->strings, sz * sizeof(struct msistring) );
 145     if( !p )
 146         return -1;
 147
 148     s = msi_realloc( st->sorted, sz*sizeof(UINT) );
 149     if( !s )
 150     {
 151         msi_free( p );
 152         return -1;
 153     }
 154
 155     st->strings = p;
 156     st->sorted = s;
 157
 158     st->freeslot = st->maxcount;
 159     st->maxcount = sz;
 160     if( st->strings[st->freeslot].persistent_refcount ||
 161         st->strings[st->freeslot].nonpersistent_refcount )
 162         ERR("oops. expected freeslot to be free...\n");
 163     return st->freeslot;
 164 }
 165
 166 static inline int cmp_string( const WCHAR *str1, int len1, const WCHAR *str2, int len2 )
 167 {
 168     if (len1 < len2) return -1;
 169     else if (len1 > len2) return 1;
 170     while (len1)
 171     {
 172         if (*str1 == *str2) { str1++; str2++; }
 173         else return *str1 - *str2;
 174         len1--;
 175     }
 176     return 0;
 177 }
 178
 179 static int find_insert_index( const string_table *st, UINT string_id )
 180 {
 181     int i, c, low = 0, high = st->sortcount - 1;
 182
 183     while (low <= high)
 184     {
 185         i = (low + high) / 2;
 186         c = cmp_string( st->strings[string_id].data, st->strings[string_id].len,
 187                         st->strings[st->sorted[i]].data, st->strings[st->sorted[i]].len );
 188         if (c < 0)
 189             high = i - 1;
 190         else if (c > 0)
 191             low = i + 1;
 192         else
 193             return -1; /* already exists */
 194     }
 195     return high + 1;
 196 }
 197
 198 static void insert_string_sorted( string_table *st, UINT string_id )
 199 {
 200     int i;
 201
 202     i = find_insert_index( st, string_id );
 203     if (i == -1)
 204         return;
 205
 206     memmove( &st->sorted[i] + 1, &st->sorted[i], (st->sortcount - i) * sizeof(UINT) );
 207     st->sorted[i] = string_id;
 208     st->sortcount++;
 209 }
 210
 211 static void set_st_entry( string_table *st, UINT n, WCHAR *str, int len, USHORT refcount,
 212                           enum StringPersistence persistence )
 213 {
 214     if (persistence == StringPersistent)
 215     {
 216         st->strings[n].persistent_refcount = refcount;
 217         st->strings[n].nonpersistent_refcount = 0;
 218     }
 219     else
 220     {
 221         st->strings[n].persistent_refcount = 0;
 222         st->strings[n].nonpersistent_refcount = refcount;
 223     }
 224
 225     st->strings[n].data = str;
 226     st->strings[n].len  = len;
 227
 228     insert_string_sorted( st, n );
 229
 230     if( n < st->maxcount )
 231         st->freeslot = n + 1;
 232 }
 233
 234 static UINT string2id( const string_table *st, const char *buffer, UINT *id )
 235 {
 236     int sz;
 237     UINT r = ERROR_INVALID_PARAMETER;
 238     LPWSTR str;
 239
 240     TRACE("Finding string %s in string table\n", debugstr_a(buffer) );
 241
 242     if( buffer[0] == 0 )
 243     {
 244         *id = 0;
 245         return ERROR_SUCCESS;
 246     }
 247
 248     if (!(sz = MultiByteToWideChar( st->codepage, 0, buffer, -1, NULL, 0 )))
 249         return r;
 250     str = msi_alloc( sz*sizeof(WCHAR) );
 251     if( !str )
 252         return ERROR_NOT_ENOUGH_MEMORY;
 253     MultiByteToWideChar( st->codepage, 0, buffer, -1, str, sz );
 254
 255     r = msi_string2id( st, str, sz - 1, id );
 256     msi_free( str );
 257     return r;
 258 }
 259
 260 static int add_string( string_table *st, UINT n, const char *data, UINT len, USHORT refcount, enum StringPersistence persistence )
 261 {
 262     LPWSTR str;
 263     int sz;
 264
 265     if( !data || !len )
 266         return 0;
 267     if( n > 0 )
 268     {
 269         if( st->strings[n].persistent_refcount ||
 270             st->strings[n].nonpersistent_refcount )
 271             return -1;
 272     }
 273     else
 274     {
 275         if (string2id( st, data, &n ) == ERROR_SUCCESS)
 276         {
 277             if (persistence == StringPersistent)
 278                 st->strings[n].persistent_refcount += refcount;
 279             else
 280                 st->strings[n].nonpersistent_refcount += refcount;
 281             return n;
 282         }
 283         n = st_find_free_entry( st );
 284         if( n == -1 )
 285             return -1;
 286     }
 287
 288     if( n < 1 )
 289     {
 290         ERR("invalid index adding %s (%d)\n", debugstr_a( data ), n );
 291         return -1;
 292     }
 293
 294     /* allocate a new string */
 295     sz = MultiByteToWideChar( st->codepage, 0, data, len, NULL, 0 );
 296     str = msi_alloc( (sz+1)*sizeof(WCHAR) );
 297     if( !str )
 298         return -1;
 299     MultiByteToWideChar( st->codepage, 0, data, len, str, sz );
 300     str[sz] = 0;
 301
 302     set_st_entry( st, n, str, sz, refcount, persistence );
 303     return n;
 304 }
 305
 306 int msi_add_string( string_table *st, const WCHAR *data, int len, enum StringPersistence persistence )
 307 {
 308     UINT n;
 309     LPWSTR str;
 310
 311     if( !data )
 312         return 0;
 313
 314     if (len < 0) len = strlenW( data );
 315
 316     if( !data[0] && !len )
 317         return 0;
 318
 319     if (msi_string2id( st, data, len, &n) == ERROR_SUCCESS )
 320     {
 321         if (persistence == StringPersistent)
 322             st->strings[n].persistent_refcount++;
 323         else
 324             st->strings[n].nonpersistent_refcount++;
 325         return n;
 326     }
 327
 328     n = st_find_free_entry( st );
 329     if( n == -1 )
 330         return -1;
 331
 332     /* allocate a new string */
 333     TRACE( "%s, n = %d len = %d\n", debugstr_wn(data, len), n, len );
 334
 335     str = msi_alloc( (len+1)*sizeof(WCHAR) );
 336     if( !str )
 337         return -1;
 338     memcpy( str, data, len*sizeof(WCHAR) );
 339     str[len] = 0;
 340
 341     set_st_entry( st, n, str, len, 1, persistence );
 342     return n;
 343 }
 344
 345 /* find the string identified by an id - return null if there's none */
 346 const WCHAR *msi_string_lookup( const string_table *st, UINT id, int *len )
 347 {
 348     if( id == 0 )
 349     {
 350         if (len) *len = 0;
 351         return szEmpty;
 352     }
 353     if( id >= st->maxcount )
 354         return NULL;
 355
 356     if( id && !st->strings[id].persistent_refcount && !st->strings[id].nonpersistent_refcount)
 357         return NULL;
 358
 359     if (len) *len = st->strings[id].len;
 360
 361     return st->strings[id].data;
 362 }
 363
 364 /*
 365  *  id2string
 366  *
 367  *  [in] st         - pointer to the string table
 368  *  [in] id         - id of the string to retrieve
 369  *  [out] buffer    - destination of the UTF8 string
 370  *  [in/out] sz     - number of bytes available in the buffer on input
 371  *                    number of bytes used on output
 372  *
 373  *  Returned string is not nul terminated.
 374  */
 375 static UINT id2string( const string_table *st, UINT id, char *buffer, UINT *sz )
 376 {
 377     int len, lenW;
 378     const WCHAR *str;
 379
 380     TRACE("Finding string %d of %d\n", id, st->maxcount);
 381
 382     str = msi_string_lookup( st, id, &lenW );
 383     if( !str )
 384         return ERROR_FUNCTION_FAILED;
 385
 386     len = WideCharToMultiByte( st->codepage, 0, str, lenW, NULL, 0, NULL, NULL );
 387     if( *sz < len )
 388     {
 389         *sz = len;
 390         return ERROR_MORE_DATA;
 391     }
 392     *sz = WideCharToMultiByte( st->codepage, 0, str, lenW, buffer, *sz, NULL, NULL );
 393     return ERROR_SUCCESS;
 394 }
 395
 396 /*
 397  *  msi_string2id
 398  *
 399  *  [in] st         - pointer to the string table
 400  *  [in] str        - string to find in the string table
 401  *  [out] id        - id of the string, if found
 402  */
 403 UINT msi_string2id( const string_table *st, const WCHAR *str, int len, UINT *id )
 404 {
 405     int i, c, low = 0, high = st->sortcount - 1;
 406
 407     if (len < 0) len = strlenW( str );
 408
 409     while (low <= high)
 410     {
 411         i = (low + high) / 2;
 412         c = cmp_string( str, len, st->strings[st->sorted[i]].data, st->strings[st->sorted[i]].len );
 413
 414         if (c < 0)
 415             high = i - 1;
 416         else if (c > 0)
 417             low = i + 1;
 418         else
 419         {
 420             *id = st->sorted[i];
 421             return ERROR_SUCCESS;
 422         }
 423     }
 424     return ERROR_INVALID_PARAMETER;
 425 }
 426
 427 static void string_totalsize( const string_table *st, UINT *datasize, UINT *poolsize )
 428 {
 429     UINT i, len, holesize;
 430
 431     if( st->strings[0].data || st->strings[0].persistent_refcount || st->strings[0].nonpersistent_refcount)
 432         ERR("oops. element 0 has a string\n");
 433
 434     *poolsize = 4;
 435     *datasize = 0;
 436     holesize = 0;
 437     for( i=1; i<st->maxcount; i++ )
 438     {
 439         if( !st->strings[i].persistent_refcount )
 440         {
 441             TRACE("[%u] nonpersistent = %s\n", i, debugstr_wn(st->strings[i].data, st->strings[i].len));
 442             (*poolsize) += 4;
 443         }
 444         else if( st->strings[i].data )
 445         {
 446             TRACE("[%u] = %s\n", i, debugstr_wn(st->strings[i].data, st->strings[i].len));
 447             len = WideCharToMultiByte( st->codepage, 0, st->strings[i].data, st->strings[i].len + 1,
 448                                        NULL, 0, NULL, NULL);
 449             if( len )
 450                 len--;
 451             (*datasize) += len;
 452             if (len>0xffff)
 453                 (*poolsize) += 4;
 454             (*poolsize) += holesize + 4;
 455             holesize = 0;
 456         }
 457         else
 458             holesize += 4;
 459     }
 460     TRACE("data %u pool %u codepage %x\n", *datasize, *poolsize, st->codepage );
 461 }
 462
 463 HRESULT msi_init_string_table( IStorage *stg )
 464 {
 465     USHORT zero[2] = { 0, 0 };
 466     UINT ret;
 467
 468     /* create the StringPool stream... add the zero string to it*/
 469     ret = write_stream_data(stg, szStringPool, zero, sizeof zero, TRUE);
 470     if (ret != ERROR_SUCCESS)
 471         return E_FAIL;
 472
 473     /* create the StringData stream... make it zero length */
 474     ret = write_stream_data(stg, szStringData, NULL, 0, TRUE);
 475     if (ret != ERROR_SUCCESS)
 476         return E_FAIL;
 477
 478     return S_OK;
 479 }
 480
 481 string_table *msi_load_string_table( IStorage *stg, UINT *bytes_per_strref )
 482 {
 483     string_table *st = NULL;
 484     CHAR *data = NULL;
 485     USHORT *pool = NULL;
 486     UINT r, datasize = 0, poolsize = 0, codepage;
 487     DWORD i, count, offset, len, n, refs;
 488
 489     r = read_stream_data( stg, szStringPool, TRUE, (BYTE **)&pool, &poolsize );
 490     if( r != ERROR_SUCCESS)
 491         goto end;
 492     r = read_stream_data( stg, szStringData, TRUE, (BYTE **)&data, &datasize );
 493     if( r != ERROR_SUCCESS)
 494         goto end;
 495
 496     if ( (poolsize > 4) && (pool[1] & 0x8000) )
 497         *bytes_per_strref = LONG_STR_BYTES;
 498     else
 499         *bytes_per_strref = sizeof(USHORT);
 500
 501     count = poolsize/4;
 502     if( poolsize > 4 )
 503         codepage = pool[0] | ( (pool[1] & ~0x8000) << 16 );
 504     else
 505         codepage = CP_ACP;
 506     st = init_stringtable( count, codepage );
 507     if (!st)
 508         goto end;
 509
 510     offset = 0;
 511     n = 1;
 512     i = 1;
 513     while( i<count )
 514     {
 515         /* the string reference count is always the second word */
 516         refs = pool[i*2+1];
 517
 518         /* empty entries have two zeros, still have a string id */
 519         if (pool[i*2] == 0 && refs == 0)
 520         {
 521             i++;
 522             n++;
 523             continue;
 524         }
 525
 526         /*
 527          * If a string is over 64k, the previous string entry is made null
 528          * and its the high word of the length is inserted in the null string's
 529          * reference count field.
 530          */
 531         if( pool[i*2] == 0)
 532         {
 533             len = (pool[i*2+3] << 16) + pool[i*2+2];
 534             i += 2;
 535         }
 536         else
 537         {
 538             len = pool[i*2];
 539             i += 1;
 540         }
 541
 542         if ( (offset + len) > datasize )
 543         {
 544             ERR("string table corrupt?\n");
 545             break;
 546         }
 547
 548         r = add_string( st, n, data+offset, len, refs, StringPersistent );
 549         if( r != n )
 550             ERR("Failed to add string %d\n", n );
 551         n++;
 552         offset += len;
 553     }
 554
 555     if ( datasize != offset )
 556         ERR("string table load failed! (%08x != %08x), please report\n", datasize, offset );
 557
 558     TRACE("Loaded %d strings\n", count);
 559
 560 end:
 561     msi_free( pool );
 562     msi_free( data );
 563
 564     return st;
 565 }
 566
 567 UINT msi_save_string_table( const string_table *st, IStorage *storage, UINT *bytes_per_strref )
 568 {
 569     UINT i, datasize = 0, poolsize = 0, sz, used, r, codepage, n;
 570     UINT ret = ERROR_FUNCTION_FAILED;
 571     CHAR *data = NULL;
 572     USHORT *pool = NULL;
 573
 574     TRACE("\n");
 575
 576     /* construct the new table in memory first */
 577     string_totalsize( st, &datasize, &poolsize );
 578
 579     TRACE("%u %u %u\n", st->maxcount, datasize, poolsize );
 580
 581     pool = msi_alloc( poolsize );
 582     if( ! pool )
 583     {
 584         WARN("Failed to alloc pool %d bytes\n", poolsize );
 585         goto err;
 586     }
 587     data = msi_alloc( datasize );
 588     if( ! data )
 589     {
 590         WARN("Failed to alloc data %d bytes\n", datasize );
 591         goto err;
 592     }
 593
 594     used = 0;
 595     codepage = st->codepage;
 596     pool[0] = codepage & 0xffff;
 597     pool[1] = codepage >> 16;
 598     if (st->maxcount > 0xffff)
 599     {
 600         pool[1] |= 0x8000;
 601         *bytes_per_strref = LONG_STR_BYTES;
 602     }
 603     else
 604         *bytes_per_strref = sizeof(USHORT);
 605
 606     n = 1;
 607     for( i=1; i<st->maxcount; i++ )
 608     {
 609         if( !st->strings[i].persistent_refcount )
 610         {
 611             pool[ n*2 ] = 0;
 612             pool[ n*2 + 1] = 0;
 613             n++;
 614             continue;
 615         }
 616
 617         sz = datasize - used;
 618         r = id2string( st, i, data+used, &sz );
 619         if( r != ERROR_SUCCESS )
 620         {
 621             ERR("failed to fetch string\n");
 622             sz = 0;
 623         }
 624
 625         if (sz)
 626             pool[ n*2 + 1 ] = st->strings[i].persistent_refcount;
 627         else
 628             pool[ n*2 + 1 ] = 0;
 629         if (sz < 0x10000)
 630         {
 631             pool[ n*2 ] = sz;
 632             n++;
 633         }
 634         else
 635         {
 636             pool[ n*2 ] = 0;
 637             pool[ n*2 + 2 ] = sz&0xffff;
 638             pool[ n*2 + 3 ] = (sz>>16);
 639             n += 2;
 640         }
 641         used += sz;
 642         if( used > datasize  )
 643         {
 644             ERR("oops overran %d >= %d\n", used, datasize);
 645             goto err;
 646         }
 647     }
 648
 649     if( used != datasize )
 650     {
 651         ERR("oops used %d != datasize %d\n", used, datasize);
 652         goto err;
 653     }
 654
 655     /* write the streams */
 656     r = write_stream_data( storage, szStringData, data, datasize, TRUE );
 657     TRACE("Wrote StringData r=%08x\n", r);
 658     if( r )
 659         goto err;
 660     r = write_stream_data( storage, szStringPool, pool, poolsize, TRUE );
 661     TRACE("Wrote StringPool r=%08x\n", r);
 662     if( r )
 663         goto err;
 664
 665     ret = ERROR_SUCCESS;
 666
 667 err:
 668     msi_free( data );
 669     msi_free( pool );
 670
 671     return ret;
 672 }
 673
 674 UINT msi_get_string_table_codepage( const string_table *st )
 675 {
 676     return st->codepage;
 677 }
 678
 679 UINT msi_set_string_table_codepage( string_table *st, UINT codepage )
 680 {
 681     if (validate_codepage( codepage ))
 682     {
 683         st->codepage = codepage;
 684         return ERROR_SUCCESS;
 685     }
 686     return ERROR_FUNCTION_FAILED;
 687 }