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