modules/access/zip/zipaccess.c

   1 /*****************************************************************************
   2  * zipaccess.c: Module (access) to extract different archives, based on zlib
   3  *****************************************************************************
   4  * Copyright (C) 2009 the VideoLAN team
   5  * $Id$
   6  *
   7  * Authors: Jean-Philippe André <jpeg@videolan.org>
   8  *
   9  * This program is free software; you can redistribute it and/or modify
  10  * it under the terms of the GNU General Public License as published by
  11  * the Free Software Foundation; either version 2 of the License, or
  12  * (at your option) any later version.
  13  *
  14  * This program is distributed in the hope that it will be useful,
  15  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  16  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  17  * GNU General Public License for more details.
  18  *
  19  * You should have received a copy of the GNU General Public License
  20  * along with this program; if not, write to the Free Software
  21  * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston MA 02110-1301, USA.
  22  *****************************************************************************/
  23
  24 /** @todo:
  25  * - implement crypto (using url zip://user:password@path-to-archive!/file)
  26  * - read files in zip with long name (use unz_file_info.size_filename)
  27  * - multi-volume archive support ?
  28  */
  29
  30 #ifdef HAVE_CONFIG_H
  31 # include "config.h"
  32 #endif
  33
  34 #include "zip.h"
  35 #include <vlc_access.h>
  36
  37 /** **************************************************************************
  38  * This is our own access_sys_t for zip files
  39  *****************************************************************************/
  40 struct access_sys_t
  41 {
  42     /* zlib / unzip members */
  43     unzFile            zipFile;
  44     zlib_filefunc_def *fileFunctions;
  45
  46     /* file in zip information */
  47     char              *psz_fileInzip;
  48 };
  49
  50 static int AccessControl( access_t *p_access, int i_query, va_list args );
  51 static ssize_t AccessRead( access_t *, uint8_t *, size_t );
  52 static int AccessSeek( access_t *, uint64_t );
  53 static int OpenFileInZip( access_t *p_access, uint64_t i_pos );
  54 static char *unescapeXml( const char *psz_text );
  55
  56 /** **************************************************************************
  57  * \brief Unescape valid XML string
  58  * The exact reverse of escapeToXml (zipstream.c)
  59  *****************************************************************************/
  60 static char *unescapeXml( const char *psz_text )
  61 {
  62     char *psz_ret = malloc( strlen( psz_text ) + 1 );
  63     if( !psz_ret ) return NULL;
  64
  65     char *psz_tmp = psz_ret;
  66     for( char *psz_iter = (char*) psz_text; *psz_iter; ++psz_iter, ++psz_tmp )
  67     {
  68         if( *psz_iter == '?' )
  69         {
  70             int i_value;
  71             if( !sscanf( ++psz_iter, "%02x", &i_value ) )
  72             {
  73                 /* Invalid number: URL incorrectly encoded */
  74                 free( psz_ret );
  75                 return NULL;
  76             }
  77             *psz_tmp = (char) i_value;
  78             psz_iter++;
  79         }
  80         else if( isAllowedChar( *psz_iter ) )
  81         {
  82             *psz_tmp = *psz_iter;
  83         }
  84         else
  85         {
  86             /* Invalid character encoding for the URL */
  87             free( psz_ret );
  88             return NULL;
  89         }
  90     }
  91     *psz_tmp = '\0';
  92
  93     return psz_ret;
  94 }
  95
  96 /** **************************************************************************
  97  * \brief Open access
  98  *****************************************************************************/
  99 int AccessOpen( vlc_object_t *p_this )
 100 {
 101     access_t     *p_access = (access_t*)p_this;
 102     access_sys_t *p_sys;
 103     int i_ret              = VLC_EGENERIC;
 104     unzFile file           = 0;
 105
 106     char *psz_pathToZip = NULL, *psz_path = NULL, *psz_sep = NULL;
 107
 108     if( !strstr( p_access->psz_location, ZIP_SEP ) )
 109     {
 110         msg_Dbg( p_access, "location does not contain separator " ZIP_SEP );
 111         return VLC_EGENERIC;
 112     }
 113
 114     p_access->p_sys = p_sys = (access_sys_t*)
 115             calloc( 1, sizeof( access_sys_t ) );
 116     if( !p_sys )
 117         return VLC_ENOMEM;
 118
 119     /* Split the MRL */
 120     psz_path = strdup( p_access->psz_location );
 121     psz_sep = strstr( psz_path, ZIP_SEP );
 122
 123     *psz_sep = '\0';
 124     psz_pathToZip = unescapeXml( psz_path );
 125     if( !psz_pathToZip )
 126     {
 127         /* Maybe this was not an encoded string */
 128         msg_Dbg( p_access, "not an encoded URL  Trying file '%s'",
 129                  psz_path );
 130         psz_pathToZip = strdup( psz_path );
 131     }
 132     p_sys->psz_fileInzip = unescapeXml( psz_sep + ZIP_SEP_LEN );
 133     if( !p_sys->psz_fileInzip )
 134     {
 135         p_sys->psz_fileInzip = strdup( psz_sep + ZIP_SEP_LEN );
 136     }
 137
 138     /* Define IO functions */
 139     zlib_filefunc_def *p_func = (zlib_filefunc_def*)
 140                                     calloc( 1, sizeof( zlib_filefunc_def ) );
 141     p_func->zopen_file   = ZipIO_Open;
 142     p_func->zread_file   = ZipIO_Read;
 143     p_func->zwrite_file  = ZipIO_Write; // see comment
 144     p_func->ztell_file   = ZipIO_Tell;
 145     p_func->zseek_file   = ZipIO_Seek;
 146     p_func->zclose_file  = ZipIO_Close;
 147     p_func->zerror_file  = ZipIO_Error;
 148     p_func->opaque       = p_access;
 149
 150     /* Open zip archive */
 151     file = p_access->p_sys->zipFile = unzOpen2( psz_pathToZip, p_func );
 152     if( !file )
 153     {
 154         msg_Err( p_access, "not a valid zip archive: '%s'", psz_pathToZip );
 155         goto exit;
 156     }
 157
 158     /* Open file in zip */
 159     OpenFileInZip( p_access, 0 );
 160
 161     /* Set callback */
 162     ACCESS_SET_CALLBACKS( AccessRead, NULL, AccessControl, AccessSeek );
 163
 164     /* Get some infos about current file. Maybe we could want some more ? */
 165     unz_file_info z_info;
 166     unzGetCurrentFileInfo( file, &z_info, NULL, 0, NULL, 0, NULL, 0 );
 167
 168     /* Set access information: size is needed for AccessSeek */
 169     p_access->info.i_size = z_info.uncompressed_size;
 170     p_access->info.i_pos  = 0;
 171     p_access->info.b_eof  = false;
 172
 173     i_ret = VLC_SUCCESS;
 174
 175 exit:
 176     if( i_ret != VLC_SUCCESS )
 177     {
 178         if( file )
 179         {
 180             unzCloseCurrentFile( file );
 181             unzClose( file );
 182         }
 183         free( p_sys->psz_fileInzip );
 184         free( p_sys->fileFunctions );
 185         free( p_sys );
 186     }
 187
 188     free( psz_pathToZip );
 189     free( psz_path );
 190     return i_ret;
 191 }
 192
 193 /** **************************************************************************
 194  * \brief Close access: free structures
 195  *****************************************************************************/
 196 void AccessClose( vlc_object_t *p_this )
 197 {
 198     access_t     *p_access = (access_t*)p_this;
 199     access_sys_t *p_sys = p_access->p_sys;
 200     if( p_sys )
 201     {
 202         unzFile file = p_sys->zipFile;
 203         if( file )
 204         {
 205             unzCloseCurrentFile( file );
 206             unzClose( file );
 207         }
 208         free( p_sys->psz_fileInzip );
 209         free( p_sys->fileFunctions );
 210         free( p_sys );
 211     }
 212 }
 213
 214 /** **************************************************************************
 215  * \brief Control access
 216  *****************************************************************************/
 217 static int AccessControl( access_t *p_access, int i_query, va_list args )
 218 {
 219     bool         *pb_bool;
 220     int64_t      *pi_64;
 221
 222     switch( i_query )
 223     {
 224         /* */
 225         case ACCESS_CAN_SEEK:
 226         case ACCESS_CAN_PAUSE:
 227         case ACCESS_CAN_CONTROL_PACE:
 228             pb_bool = (bool*)va_arg( args, bool* );
 229             *pb_bool = true;
 230             break;
 231
 232         case ACCESS_CAN_FASTSEEK:
 233             pb_bool = (bool*)va_arg( args, bool* );
 234             *pb_bool = false;
 235             break;
 236
 237         case ACCESS_GET_PTS_DELAY:
 238             pi_64 = (int64_t*)va_arg( args, int64_t * );
 239             *pi_64 = DEFAULT_PTS_DELAY;
 240             break;
 241
 242         /* */
 243         case ACCESS_SET_PAUSE_STATE:
 244             /* Nothing to do */
 245             break;
 246
 247         case ACCESS_GET_TITLE_INFO:
 248         case ACCESS_SET_TITLE:
 249         case ACCESS_SET_SEEKPOINT:
 250         case ACCESS_SET_PRIVATE_ID_STATE:
 251         case ACCESS_GET_META:
 252         case ACCESS_GET_PRIVATE_ID_STATE:
 253         case ACCESS_GET_CONTENT_TYPE:
 254             return VLC_EGENERIC;
 255
 256         default:
 257             msg_Warn( p_access, "unimplemented query %d in control", i_query );
 258             return VLC_EGENERIC;
 259
 260     }
 261     return VLC_SUCCESS;
 262 }
 263
 264 /** **************************************************************************
 265  * \brief Read access
 266  * Reads current opened file in zip. This does not open the file in zip.
 267  * Return -1 if no data yet, 0 if no more data, else real data read
 268  *****************************************************************************/
 269 static ssize_t AccessRead( access_t *p_access, uint8_t *p_buffer, size_t sz )
 270 {
 271     access_sys_t *p_sys = p_access->p_sys;
 272     assert( p_sys );
 273     unzFile file = p_sys->zipFile;
 274     if( !file )
 275     {
 276         msg_Err( p_access, "archive not opened !" );
 277         return VLC_EGENERIC;
 278     }
 279
 280     int i_read = 0;
 281     i_read = unzReadCurrentFile( file, p_buffer, sz );
 282
 283     p_access->info.i_pos = unztell( file );
 284     return ( i_read >= 0 ? i_read : VLC_EGENERIC );
 285 }
 286
 287 /** **************************************************************************
 288  * \brief Seek inside zip file
 289  *****************************************************************************/
 290 static int AccessSeek( access_t *p_access, uint64_t seek_len )
 291 {
 292     access_sys_t *p_sys = p_access->p_sys;
 293     assert( p_sys );
 294     unzFile file = p_sys->zipFile;
 295     if( !file )
 296     {
 297         msg_Err( p_access, "archive not opened !" );
 298         return VLC_EGENERIC;
 299     }
 300
 301     /* Reopen file in zip if needed */
 302     if( p_access->info.i_pos != 0 )
 303     {
 304         OpenFileInZip( p_access, p_access->info.i_pos + seek_len );
 305     }
 306
 307     /* Read seek_len data and drop it */
 308     unsigned i_seek = 0;
 309     int i_read = 1;
 310     char *p_buffer = ( char* ) calloc( 1, ZIP_BUFFER_LEN );
 311     while( ( i_seek < seek_len ) && ( i_read > 0 ) )
 312     {
 313         i_read = ( seek_len - i_seek < ZIP_BUFFER_LEN )
 314                ? ( seek_len - i_seek ) : ZIP_BUFFER_LEN;
 315         i_read = unzReadCurrentFile( file, p_buffer, i_read );
 316         if( i_read < 0 )
 317         {
 318             msg_Warn( p_access, "could not seek in file" );
 319             free( p_buffer );
 320             return VLC_EGENERIC;
 321         }
 322         else
 323         {
 324             i_seek += i_read;
 325         }
 326     }
 327     free( p_buffer );
 328
 329     p_access->info.i_pos = unztell( file );
 330     return VLC_SUCCESS;
 331 }
 332
 333 /** **************************************************************************
 334  * \brief Open file in zip
 335  *****************************************************************************/
 336 static int OpenFileInZip( access_t *p_access, uint64_t i_pos )
 337 {
 338     access_sys_t *p_sys = p_access->p_sys;
 339     unzFile file = p_sys->zipFile;
 340     if( !p_sys->psz_fileInzip )
 341     {
 342         return VLC_EGENERIC;
 343     }
 344
 345     p_access->info.i_pos = 0;
 346
 347     unzCloseCurrentFile( file ); /* returns UNZ_PARAMERROR if file not opened */
 348     if( unzLocateFile( file, p_sys->psz_fileInzip, 0 ) != UNZ_OK )
 349     {
 350         msg_Err( p_access, "could not [re]locate file in zip: '%s'",
 351                  p_sys->psz_fileInzip );
 352         return VLC_EGENERIC;
 353     }
 354     if( unzOpenCurrentFile( file ) != UNZ_OK )
 355     {
 356         msg_Err( p_access, "could not [re]open file in zip: '%s'",
 357                  p_sys->psz_fileInzip );
 358         return VLC_EGENERIC;
 359     }
 360     if( i_pos > 0 )
 361         return AccessSeek( p_access, i_pos );
 362     else
 363         return VLC_SUCCESS;
 364 }
 365
 366 /** **************************************************************************
 367  * \brief I/O functions for the ioapi: open (read only)
 368  *****************************************************************************/
 369 static void* ZCALLBACK ZipIO_Open( void* opaque, const char* file, int mode )
 370 {
 371     assert(opaque != NULL);
 372     assert(mode == (ZLIB_FILEFUNC_MODE_READ | ZLIB_FILEFUNC_MODE_EXISTING));
 373
 374     access_t *p_access = (access_t*) opaque;
 375
 376     return stream_UrlNew( p_access, file );
 377 }
 378
 379 /** **************************************************************************
 380  * \brief I/O functions for the ioapi: read
 381  *****************************************************************************/
 382 static uLong ZCALLBACK ZipIO_Read( void* opaque, void* stream,
 383                                    void* buf, uLong size )
 384 {
 385     (void)opaque;
 386     //access_t *p_access = (access_t*) opaque;
 387     //msg_Dbg(p_access, "read %d", size);
 388     return stream_Read( (stream_t*) stream, buf, size );
 389 }
 390
 391 /** **************************************************************************
 392  * \brief I/O functions for the ioapi: write (assert insteadof segfault)
 393  *****************************************************************************/
 394 static uLong ZCALLBACK ZipIO_Write( void* opaque, void* stream,
 395                                     const void* buf, uLong size )
 396 {
 397     (void)opaque; (void)stream; (void)buf; (void)size;
 398     int zip_access_cannot_write_this_should_not_happen = 0;
 399     assert(zip_access_cannot_write_this_should_not_happen);
 400     return 0;
 401 }
 402
 403 /** **************************************************************************
 404  * \brief I/O functions for the ioapi: tell
 405  *****************************************************************************/
 406 static long ZCALLBACK ZipIO_Tell( void* opaque, void* stream )
 407 {
 408     (void)opaque;
 409     int64_t i64_tell = stream_Tell( (stream_t*) stream );
 410     //access_t *p_access = (access_t*) opaque;
 411     //msg_Dbg(p_access, "tell %" PRIu64, i64_tell);
 412     return (long)i64_tell;
 413 }
 414
 415 /** **************************************************************************
 416  * \brief I/O functions for the ioapi: seek
 417  *****************************************************************************/
 418 static long ZCALLBACK ZipIO_Seek( void* opaque, void* stream,
 419                                   uLong offset, int origin )
 420 {
 421     (void)opaque;
 422     //access_t *p_access = (access_t*) opaque;
 423     int64_t pos = offset;
 424     switch( origin )
 425     {
 426         case SEEK_CUR:
 427             pos += stream_Tell( (stream_t*) stream );
 428             break;
 429         case SEEK_SET:
 430             break;
 431         case SEEK_END:
 432             pos += stream_Size( (stream_t*) stream );
 433             break;
 434         default:
 435             return -1;
 436     }
 437     if( pos < 0 )
 438         return -1;
 439     //msg_Dbg( p_access, "seek (%d,%d): %" PRIu64, offset, origin, pos );
 440     stream_Seek( (stream_t*) stream, pos );
 441     /* Note: in unzip.c, unzlocal_SearchCentralDir seeks to the end of
 442              the stream, which is doable but returns an error in VLC.
 443              That's why we always assume this was OK. FIXME */
 444     return 0;
 445 }
 446
 447 /** **************************************************************************
 448  * \brief I/O functions for the ioapi: close
 449  *****************************************************************************/
 450 static int ZCALLBACK ZipIO_Close( void* opaque, void* stream )
 451 {
 452     (void)opaque;
 453     stream_Delete( (stream_t*) stream );
 454     return 0;
 455 }
 456
 457 /** **************************************************************************
 458  * \brief I/O functions for the ioapi: test error (man 3 ferror)
 459  *****************************************************************************/
 460 static int ZCALLBACK ZipIO_Error( void* opaque, void* stream )
 461 {
 462     (void)opaque;
 463     (void)stream;
 464     //msg_Dbg( p_access, "error" );
 465     return 0;
 466 }