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  *
   7  * This library is free software; you can redistribute it and/or
   8  * modify it under the terms of the GNU Lesser General Public
   9  * License as published by the Free Software Foundation; either
  10  * version 2.1 of the License, or (at your option) any later version.
  11  *
  12  * This library is distributed in the hope that it will be useful,
  13  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  14  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
  15  * Lesser General Public License for more details.
  16  *
  17  * You should have received a copy of the GNU Lesser General Public
  18  * License along with this library; if not, write to the Free Software
  19  * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301, USA
  20  */
  21
  22 #define COBJMACROS
  23
  24 #include <stdarg.h>
  25 #include <assert.h>
  26
  27 #include "windef.h"
  28 #include "winbase.h"
  29 #include "winerror.h"
  30 #include "wine/debug.h"
  31 #include "wine/unicode.h"
  32 #include "msi.h"
  33 #include "msiquery.h"
  34 #include "objbase.h"
  35 #include "objidl.h"
  36 #include "msipriv.h"
  37 #include "winnls.h"
  38
  39 #include "query.h"
  40
  41 WINE_DEFAULT_DEBUG_CHANNEL(msidb);
  42
  43 #define HASH_SIZE 0x101
  44 #define LONG_STR_BYTES 3
  45
  46 typedef struct _msistring
  47 {
  48     int hash_next;
  49     UINT persistent_refcount;
  50     UINT nonpersistent_refcount;
  51     LPWSTR str;
  52 } msistring;
  53
  54 struct string_table
  55 {
  56     UINT maxcount;         /* the number of strings */
  57     UINT freeslot;
  58     UINT codepage;
  59     int hash[HASH_SIZE];
  60     msistring *strings; /* an array of strings (in the tree) */
  61 };
  62
  63 static UINT msistring_makehash( const WCHAR *str )
  64 {
  65     UINT hash = 0;
  66
  67     if (str==NULL)
  68         return hash;
  69
  70     while( *str )
  71     {
  72         hash ^= *str++;
  73         hash *= 53;
  74         hash = (hash<<5) | (hash>>27);
  75     }
  76     return hash % HASH_SIZE;
  77 }
  78
  79 static string_table *init_stringtable( int entries, UINT codepage )
  80 {
  81     string_table *st;
  82     int i;
  83
  84     if (codepage != CP_ACP && !IsValidCodePage(codepage))
  85     {
  86         ERR("invalid codepage %d\n", codepage);
  87         return NULL;
  88     }
  89
  90     st = msi_alloc( sizeof (string_table) );
  91     if( !st )
  92         return NULL;
  93     if( entries < 1 )
  94         entries = 1;
  95     st->strings = msi_alloc_zero( sizeof (msistring) * entries );
  96     if( !st->strings )
  97     {
  98         msi_free( st );
  99         return NULL;
 100     }
 101     st->maxcount = entries;
 102     st->freeslot = 1;
 103     st->codepage = codepage;
 104
 105     for( i=0; i<HASH_SIZE; i++ )
 106         st->hash[i] = -1;
 107
 108     return st;
 109 }
 110
 111 VOID msi_destroy_stringtable( string_table *st )
 112 {
 113     UINT i;
 114
 115     for( i=0; i<st->maxcount; i++ )
 116     {
 117         if( st->strings[i].persistent_refcount ||
 118             st->strings[i].nonpersistent_refcount )
 119             msi_free( st->strings[i].str );
 120     }
 121     msi_free( st->strings );
 122     msi_free( st );
 123 }
 124
 125 static int st_find_free_entry( string_table *st )
 126 {
 127     UINT i, sz;
 128     msistring *p;
 129
 130     TRACE("%p\n", st);
 131
 132     if( st->freeslot )
 133     {
 134         for( i = st->freeslot; i < st->maxcount; i++ )
 135             if( !st->strings[i].persistent_refcount &&
 136                 !st->strings[i].nonpersistent_refcount )
 137                 return i;
 138     }
 139     for( i = 1; i < st->maxcount; i++ )
 140         if( !st->strings[i].persistent_refcount &&
 141             !st->strings[i].nonpersistent_refcount )
 142             return i;
 143
 144     /* dynamically resize */
 145     sz = st->maxcount + 1 + st->maxcount/2;
 146     p = msi_realloc_zero( st->strings, sz*sizeof(msistring) );
 147     if( !p )
 148         return -1;
 149     st->strings = p;
 150     st->freeslot = st->maxcount;
 151     st->maxcount = sz;
 152     if( st->strings[st->freeslot].persistent_refcount ||
 153         st->strings[st->freeslot].nonpersistent_refcount )
 154         ERR("oops. expected freeslot to be free...\n");
 155     return st->freeslot;
 156 }
 157
 158 static void set_st_entry( string_table *st, UINT n, LPWSTR str, UINT refcount, enum StringPersistence persistence )
 159 {
 160     UINT hash = msistring_makehash( str );
 161
 162     if (persistence == StringPersistent)
 163     {
 164         st->strings[n].persistent_refcount = refcount;
 165         st->strings[n].nonpersistent_refcount = 0;
 166     }
 167     else
 168     {
 169         st->strings[n].persistent_refcount = 0;
 170         st->strings[n].nonpersistent_refcount = refcount;
 171     }
 172
 173     st->strings[n].str = str;
 174
 175     st->strings[n].hash_next = st->hash[hash];
 176     st->hash[hash] = n;
 177
 178     if( n < st->maxcount )
 179         st->freeslot = n + 1;
 180 }
 181
 182 static int msi_addstring( string_table *st, UINT n, const CHAR *data, int len, UINT refcount, enum StringPersistence persistence )
 183 {
 184     LPWSTR str;
 185     int sz;
 186
 187     if( !data )
 188         return 0;
 189     if( !data[0] )
 190         return 0;
 191     if( n > 0 )
 192     {
 193         if( st->strings[n].persistent_refcount ||
 194             st->strings[n].nonpersistent_refcount )
 195             return -1;
 196     }
 197     else
 198     {
 199         if( ERROR_SUCCESS == msi_string2idA( st, data, &n ) )
 200         {
 201             if (persistence == StringPersistent)
 202                 st->strings[n].persistent_refcount += refcount;
 203             else
 204                 st->strings[n].nonpersistent_refcount += refcount;
 205             return n;
 206         }
 207         n = st_find_free_entry( st );
 208         if( n == -1 )
 209             return -1;
 210     }
 211
 212     if( n < 1 )
 213     {
 214         ERR("invalid index adding %s (%d)\n", debugstr_a( data ), n );
 215         return -1;
 216     }
 217
 218     /* allocate a new string */
 219     if( len < 0 )
 220         len = strlen(data);
 221     sz = MultiByteToWideChar( st->codepage, 0, data, len, NULL, 0 );
 222     str = msi_alloc( (sz+1)*sizeof(WCHAR) );
 223     if( !str )
 224         return -1;
 225     MultiByteToWideChar( st->codepage, 0, data, len, str, sz );
 226     str[sz] = 0;
 227
 228     set_st_entry( st, n, str, refcount, persistence );
 229
 230     return n;
 231 }
 232
 233 int msi_addstringW( string_table *st, UINT n, const WCHAR *data, int len, UINT refcount, enum StringPersistence persistence )
 234 {
 235     LPWSTR str;
 236
 237     /* TRACE("[%2d] = %s\n", string_no, debugstr_an(data,len) ); */
 238
 239     if( !data )
 240         return 0;
 241     if( !data[0] )
 242         return 0;
 243     if( n > 0 )
 244     {
 245         if( st->strings[n].persistent_refcount ||
 246             st->strings[n].nonpersistent_refcount )
 247             return -1;
 248     }
 249     else
 250     {
 251         if( ERROR_SUCCESS == msi_string2idW( st, data, &n ) )
 252         {
 253             if (persistence == StringPersistent)
 254                 st->strings[n].persistent_refcount += refcount;
 255             else
 256                 st->strings[n].nonpersistent_refcount += refcount;
 257             return n;
 258         }
 259         n = st_find_free_entry( st );
 260         if( n == -1 )
 261             return -1;
 262     }
 263
 264     if( n < 1 )
 265     {
 266         ERR("invalid index adding %s (%d)\n", debugstr_w( data ), n );
 267         return -1;
 268     }
 269
 270     /* allocate a new string */
 271     if(len<0)
 272         len = strlenW(data);
 273     TRACE("%s, n = %d len = %d\n", debugstr_w(data), n, len );
 274
 275     str = msi_alloc( (len+1)*sizeof(WCHAR) );
 276     if( !str )
 277         return -1;
 278     memcpy( str, data, len*sizeof(WCHAR) );
 279     str[len] = 0;
 280
 281     set_st_entry( st, n, str, refcount, persistence );
 282
 283     return n;
 284 }
 285
 286 /* find the string identified by an id - return null if there's none */
 287 const WCHAR *msi_string_lookup_id( const string_table *st, UINT id )
 288 {
 289     static const WCHAR zero[] = { 0 };
 290     if( id == 0 )
 291         return zero;
 292
 293     if( id >= st->maxcount )
 294         return NULL;
 295
 296     if( id && !st->strings[id].persistent_refcount && !st->strings[id].nonpersistent_refcount)
 297         return NULL;
 298
 299     return st->strings[id].str;
 300 }
 301
 302 /*
 303  *  msi_id2stringW
 304  *
 305  *  [in] st         - pointer to the string table
 306  *  [in] id  - id of the string to retrieve
 307  *  [out] buffer    - destination of the string
 308  *  [in/out] sz     - number of bytes available in the buffer on input
 309  *                    number of bytes used on output
 310  *
 311  *   The size includes the terminating nul character.  Short buffers
 312  *  will be filled, but not nul terminated.
 313  */
 314 UINT msi_id2stringW( const string_table *st, UINT id, LPWSTR buffer, UINT *sz )
 315 {
 316     UINT len;
 317     const WCHAR *str;
 318
 319     TRACE("Finding string %d of %d\n", id, st->maxcount);
 320
 321     str = msi_string_lookup_id( st, id );
 322     if( !str )
 323         return ERROR_FUNCTION_FAILED;
 324
 325     len = strlenW( str ) + 1;
 326
 327     if( !buffer )
 328     {
 329         *sz = len;
 330         return ERROR_SUCCESS;
 331     }
 332
 333     if( *sz < len )
 334         *sz = len;
 335     memcpy( buffer, str, (*sz)*sizeof(WCHAR) );
 336     *sz = len;
 337
 338     return ERROR_SUCCESS;
 339 }
 340
 341 /*
 342  *  msi_id2stringA
 343  *
 344  *  [in] st         - pointer to the string table
 345  *  [in] id         - id of the string to retrieve
 346  *  [out] buffer    - destination of the UTF8 string
 347  *  [in/out] sz     - number of bytes available in the buffer on input
 348  *                    number of bytes used on output
 349  *
 350  *   The size includes the terminating nul character.  Short buffers
 351  *  will be filled, but not nul terminated.
 352  */
 353 UINT msi_id2stringA( const string_table *st, UINT id, LPSTR buffer, UINT *sz )
 354 {
 355     UINT len;
 356     const WCHAR *str;
 357     int n;
 358
 359     TRACE("Finding string %d of %d\n", id, st->maxcount);
 360
 361     str = msi_string_lookup_id( st, id );
 362     if( !str )
 363         return ERROR_FUNCTION_FAILED;
 364
 365     len = WideCharToMultiByte( st->codepage, 0, str, -1, NULL, 0, NULL, NULL );
 366
 367     if( !buffer )
 368     {
 369         *sz = len;
 370         return ERROR_SUCCESS;
 371     }
 372
 373     if( len > *sz )
 374     {
 375         n = strlenW( str ) + 1;
 376         while( n && (len > *sz) )
 377             len = WideCharToMultiByte( st->codepage, 0,
 378                            str, --n, NULL, 0, NULL, NULL );
 379     }
 380     else
 381         n = -1;
 382
 383     *sz = WideCharToMultiByte( st->codepage, 0, str, n, buffer, len, NULL, NULL );
 384
 385     return ERROR_SUCCESS;
 386 }
 387
 388 /*
 389  *  msi_string2idW
 390  *
 391  *  [in] st         - pointer to the string table
 392  *  [in] str        - string to find in the string table
 393  *  [out] id        - id of the string, if found
 394  */
 395 UINT msi_string2idW( const string_table *st, LPCWSTR str, UINT *id )
 396 {
 397     UINT n, hash = msistring_makehash( str );
 398     msistring *se = st->strings;
 399
 400     for (n = st->hash[hash]; n != -1; n = st->strings[n].hash_next )
 401     {
 402         if ((str == se[n].str) || !lstrcmpW(str, se[n].str))
 403         {
 404             *id = n;
 405             return ERROR_SUCCESS;
 406         }
 407     }
 408
 409     return ERROR_INVALID_PARAMETER;
 410 }
 411
 412 UINT msi_string2idA( const string_table *st, LPCSTR buffer, UINT *id )
 413 {
 414     DWORD sz;
 415     UINT r = ERROR_INVALID_PARAMETER;
 416     LPWSTR str;
 417
 418     TRACE("Finding string %s in string table\n", debugstr_a(buffer) );
 419
 420     if( buffer[0] == 0 )
 421     {
 422         *id = 0;
 423         return ERROR_SUCCESS;
 424     }
 425
 426     sz = MultiByteToWideChar( st->codepage, 0, buffer, -1, NULL, 0 );
 427     if( sz <= 0 )
 428         return r;
 429     str = msi_alloc( sz*sizeof(WCHAR) );
 430     if( !str )
 431         return ERROR_NOT_ENOUGH_MEMORY;
 432     MultiByteToWideChar( st->codepage, 0, buffer, -1, str, sz );
 433
 434     r = msi_string2idW( st, str, id );
 435     msi_free( str );
 436
 437     return r;
 438 }
 439
 440 UINT msi_strcmp( const string_table *st, UINT lval, UINT rval, UINT *res )
 441 {
 442     const WCHAR *l_str, *r_str;
 443
 444     l_str = msi_string_lookup_id( st, lval );
 445     if( !l_str )
 446         return ERROR_INVALID_PARAMETER;
 447
 448     r_str = msi_string_lookup_id( st, rval );
 449     if( !r_str )
 450         return ERROR_INVALID_PARAMETER;
 451
 452     /* does this do the right thing for all UTF-8 strings? */
 453     *res = strcmpW( l_str, r_str );
 454
 455     return ERROR_SUCCESS;
 456 }
 457
 458 static void string_totalsize( const string_table *st, UINT *datasize, UINT *poolsize )
 459 {
 460     UINT i, len, holesize;
 461
 462     if( st->strings[0].str || st->strings[0].persistent_refcount || st->strings[0].nonpersistent_refcount)
 463         ERR("oops. element 0 has a string\n");
 464
 465     *poolsize = 4;
 466     *datasize = 0;
 467     holesize = 0;
 468     for( i=1; i<st->maxcount; i++ )
 469     {
 470         if( !st->strings[i].persistent_refcount )
 471         {
 472             TRACE("[%u] nonpersistent = %s\n", i, debugstr_w(st->strings[i].str));
 473             (*poolsize) += 4;
 474         }
 475         else if( st->strings[i].str )
 476         {
 477             TRACE("[%u] = %s\n", i, debugstr_w(st->strings[i].str));
 478             len = WideCharToMultiByte( st->codepage, 0,
 479                      st->strings[i].str, -1, NULL, 0, NULL, NULL);
 480             if( len )
 481                 len--;
 482             (*datasize) += len;
 483             if (len>0xffff)
 484                 (*poolsize) += 4;
 485             (*poolsize) += holesize + 4;
 486             holesize = 0;
 487         }
 488         else
 489             holesize += 4;
 490     }
 491     TRACE("data %u pool %u codepage %x\n", *datasize, *poolsize, st->codepage );
 492 }
 493
 494 static const WCHAR szStringData[] = {
 495     '_','S','t','r','i','n','g','D','a','t','a',0 };
 496 static const WCHAR szStringPool[] = {
 497     '_','S','t','r','i','n','g','P','o','o','l',0 };
 498
 499 HRESULT msi_init_string_table( IStorage *stg )
 500 {
 501     USHORT zero[2] = { 0, 0 };
 502     UINT ret;
 503
 504     /* create the StringPool stream... add the zero string to it*/
 505     ret = write_stream_data(stg, szStringPool, zero, sizeof zero, TRUE);
 506     if (ret != ERROR_SUCCESS)
 507         return E_FAIL;
 508
 509     /* create the StringData stream... make it zero length */
 510     ret = write_stream_data(stg, szStringData, NULL, 0, TRUE);
 511     if (ret != ERROR_SUCCESS)
 512         return E_FAIL;
 513
 514     return S_OK;
 515 }
 516
 517 string_table *msi_load_string_table( IStorage *stg, UINT *bytes_per_strref )
 518 {
 519     string_table *st = NULL;
 520     CHAR *data = NULL;
 521     USHORT *pool = NULL;
 522     UINT r, datasize = 0, poolsize = 0, codepage;
 523     DWORD i, count, offset, len, n, refs;
 524
 525     r = read_stream_data( stg, szStringPool, TRUE, (BYTE **)&pool, &poolsize );
 526     if( r != ERROR_SUCCESS)
 527         goto end;
 528     r = read_stream_data( stg, szStringData, TRUE, (BYTE **)&data, &datasize );
 529     if( r != ERROR_SUCCESS)
 530         goto end;
 531
 532     if ( (poolsize > 4) && (pool[1] & 0x8000) )
 533         *bytes_per_strref = LONG_STR_BYTES;
 534     else
 535         *bytes_per_strref = sizeof(USHORT);
 536
 537     count = poolsize/4;
 538     if( poolsize > 4 )
 539         codepage = pool[0] | ( (pool[1] & ~0x8000) << 16 );
 540     else
 541         codepage = CP_ACP;
 542     st = init_stringtable( count, codepage );
 543     if (!st)
 544         goto end;
 545
 546     offset = 0;
 547     n = 1;
 548     i = 1;
 549     while( i<count )
 550     {
 551         /* the string reference count is always the second word */
 552         refs = pool[i*2+1];
 553
 554         /* empty entries have two zeros, still have a string id */
 555         if (pool[i*2] == 0 && refs == 0)
 556         {
 557             i++;
 558             n++;
 559             continue;
 560         }
 561
 562         /*
 563          * If a string is over 64k, the previous string entry is made null
 564          * and its the high word of the length is inserted in the null string's
 565          * reference count field.
 566          */
 567         if( pool[i*2] == 0)
 568         {
 569             len = (pool[i*2+3] << 16) + pool[i*2+2];
 570             i += 2;
 571         }
 572         else
 573         {
 574             len = pool[i*2];
 575             i += 1;
 576         }
 577
 578         if ( (offset + len) > datasize )
 579         {
 580             ERR("string table corrupt?\n");
 581             break;
 582         }
 583
 584         r = msi_addstring( st, n, data+offset, len, refs, StringPersistent );
 585         if( r != n )
 586             ERR("Failed to add string %d\n", n );
 587         n++;
 588         offset += len;
 589     }
 590
 591     if ( datasize != offset )
 592         ERR("string table load failed! (%08x != %08x), please report\n", datasize, offset );
 593
 594     TRACE("Loaded %d strings\n", count);
 595
 596 end:
 597     msi_free( pool );
 598     msi_free( data );
 599
 600     return st;
 601 }
 602
 603 UINT msi_save_string_table( const string_table *st, IStorage *storage )
 604 {
 605     UINT i, datasize = 0, poolsize = 0, sz, used, r, codepage, n;
 606     UINT ret = ERROR_FUNCTION_FAILED;
 607     CHAR *data = NULL;
 608     USHORT *pool = NULL;
 609
 610     TRACE("\n");
 611
 612     /* construct the new table in memory first */
 613     string_totalsize( st, &datasize, &poolsize );
 614
 615     TRACE("%u %u %u\n", st->maxcount, datasize, poolsize );
 616
 617     pool = msi_alloc( poolsize );
 618     if( ! pool )
 619     {
 620         WARN("Failed to alloc pool %d bytes\n", poolsize );
 621         goto err;
 622     }
 623     data = msi_alloc( datasize );
 624     if( ! data )
 625     {
 626         WARN("Failed to alloc data %d bytes\n", poolsize );
 627         goto err;
 628     }
 629
 630     used = 0;
 631     codepage = st->codepage;
 632     pool[0]=codepage&0xffff;
 633     pool[1]=(codepage>>16);
 634     n = 1;
 635     for( i=1; i<st->maxcount; i++ )
 636     {
 637         if( !st->strings[i].persistent_refcount )
 638         {
 639             pool[ n*2 ] = 0;
 640             pool[ n*2 + 1] = 0;
 641             n++;
 642             continue;
 643         }
 644
 645         sz = datasize - used;
 646         r = msi_id2stringA( st, i, data+used, &sz );
 647         if( r != ERROR_SUCCESS )
 648         {
 649             ERR("failed to fetch string\n");
 650             sz = 0;
 651         }
 652         if( sz && (sz < (datasize - used ) ) )
 653             sz--;
 654
 655         if (sz)
 656             pool[ n*2 + 1 ] = st->strings[i].persistent_refcount;
 657         else
 658             pool[ n*2 + 1 ] = 0;
 659         if (sz < 0x10000)
 660         {
 661             pool[ n*2 ] = sz;
 662             n++;
 663         }
 664         else
 665         {
 666             pool[ n*2 ] = 0;
 667             pool[ n*2 + 2 ] = sz&0xffff;
 668             pool[ n*2 + 3 ] = (sz>>16);
 669             n += 2;
 670         }
 671         used += sz;
 672         if( used > datasize  )
 673         {
 674             ERR("oops overran %d >= %d\n", used, datasize);
 675             goto err;
 676         }
 677     }
 678
 679     if( used != datasize )
 680     {
 681         ERR("oops used %d != datasize %d\n", used, datasize);
 682         goto err;
 683     }
 684
 685     /* write the streams */
 686     r = write_stream_data( storage, szStringData, data, datasize, TRUE );
 687     TRACE("Wrote StringData r=%08x\n", r);
 688     if( r )
 689         goto err;
 690     r = write_stream_data( storage, szStringPool, pool, poolsize, TRUE );
 691     TRACE("Wrote StringPool r=%08x\n", r);
 692     if( r )
 693         goto err;
 694
 695     ret = ERROR_SUCCESS;
 696
 697 err:
 698     msi_free( data );
 699     msi_free( pool );
 700
 701     return ret;
 702 }