#include "test.h" #include "md5.h" #include "mongo.h" #include "gridfs.h" #include #include #include #include #include #ifndef _WIN32 #include #endif #define LARGE 3*1024*1024 #define UPPER 2000*1024 #define MEDIUM 1024*512 #define LOWER 1024*128 #define DELTA 1024*128 #define READ_WRITE_BUF_SIZE 10 * 1024 #define GRIDFILE_COMPRESS 2 #ifdef _MSC_VER #define gridfs_test_unlink _unlink #else #define gridfs_test_unlink unlink #endif #define GFS_INIT \ gridfs_init( conn, "test", "fs", gfs ) void fill_buffer_randomly( char *data, int64_t length ) { int64_t i; int random; char *letters = "abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ"; int nletters = (int)strlen( letters ); char *cur; for ( i = 0, cur = data; i < length; i++, cur++ ) { random = rand() % nletters; *cur = letters[random]; } } static void digest2hex( mongo_md5_byte_t digest[16], char hex_digest[33] ) { static const char hex[16] = {'0','1','2','3','4','5','6','7','8','9','a','b','c','d','e','f'}; int i; for ( i=0; i<16; i++ ) { hex_digest[2*i] = hex[( digest[i] & 0xf0 ) >> 4]; hex_digest[2*i + 1] = hex[ digest[i] & 0x0f ]; } hex_digest[32] = '\0'; } void test_gridfile( gridfs *gfs, char *data_before, int64_t length, char *filename, char *content_type ) { gridfile gfile[1]; FILE *stream; mongo_md5_state_t pms[1]; mongo_md5_byte_t digest[16]; char hex_digest[33]; int64_t i = length; int n; char *data_after = (char*)bson_malloc( LARGE ); int truncBytes; char* lowerName; ASSERT(gridfs_find_filename( gfs, filename, gfile ) == MONGO_OK); ASSERT( gridfile_exists( gfile ) ); stream = fopen( "output", "w+" ); gridfile_write_file( gfile, stream ); fseek( stream, 0, SEEK_SET ); ASSERT( fread( data_after, (size_t)length, sizeof( char ), stream ) ); fclose( stream ); ASSERT( memcmp( data_before, data_after, (size_t)length ) == 0 ); gridfile_read_buffer( gfile, data_after, length ); ASSERT( memcmp( data_before, data_after, (size_t)length ) == 0 ); lowerName = (char*) bson_malloc( (int)strlen( filename ) + 1); strcpy( lowerName, filename ); _strlwr( lowerName ); ASSERT( strcmp( gridfile_get_filename( gfile ), lowerName ) == 0 ); bson_free( lowerName ); ASSERT( gridfile_get_contentlength( gfile ) == (size_t)length ); ASSERT( gridfile_get_chunksize( gfile ) == DEFAULT_CHUNK_SIZE ); ASSERT( strcmp( gridfile_get_contenttype( gfile ), content_type ) == 0 ) ; ASSERT( memcmp( data_before, data_after, (size_t)length ) == 0 ); if( !( gfile->flags & GRIDFILE_COMPRESS ) ) { mongo_md5_init( pms ); n = 0; while( i > INT_MAX ) { mongo_md5_append( pms, ( const mongo_md5_byte_t * )data_before + ( n * INT_MAX ), INT_MAX ); i -= INT_MAX; n += 1; } if( i > 0 ) mongo_md5_append( pms, ( const mongo_md5_byte_t * )data_before + ( n * INT_MAX ), (int)i ); mongo_md5_finish( pms, digest ); digest2hex( digest, hex_digest ); ASSERT( strcmp( gridfile_get_md5( gfile ), hex_digest ) == 0 ); } truncBytes = (int) (length > DEFAULT_CHUNK_SIZE * 4 ? length - DEFAULT_CHUNK_SIZE * 2 - 13 : 23); gridfile_writer_init( gfile, gfs, filename, content_type, GRIDFILE_DEFAULT); ASSERT( gridfile_truncate(gfile, (size_t)(length - truncBytes)) == (size_t)(length - truncBytes)); gridfile_writer_done( gfile ); gridfile_seek(gfile, 0); ASSERT( gridfile_get_contentlength( gfile ) == (size_t)(length - truncBytes) ); ASSERT( gridfile_read_buffer( gfile, data_after, length ) == (size_t)(length - truncBytes)); ASSERT( memcmp( data_before, data_after, (size_t)(length - truncBytes) ) == 0 ); gridfile_writer_init( gfile, gfs, filename, content_type, GRIDFILE_DEFAULT); gridfile_truncate(gfile, 0); gridfile_writer_done( gfile ); ASSERT( gridfile_get_contentlength( gfile ) == 0 ); ASSERT( gridfile_read_buffer( gfile, data_after, length ) == 0 ); gridfile_destroy( gfile ); ASSERT( gridfs_remove_filename( gfs, filename ) == MONGO_OK ); free( data_after ); gridfs_test_unlink( "output" ); } void test_basic( void ) { mongo conn[1]; gridfs gfs[1]; char *data_before = (char*)bson_malloc( UPPER ); int64_t i; FILE *fd; srand((unsigned int) time( NULL ) ); INIT_SOCKETS_FOR_WINDOWS; CONN_CLIENT_TEST; GFS_INIT; fill_buffer_randomly( data_before, UPPER ); for ( i = LOWER; i <= UPPER; i += DELTA ) { /* Input from buffer */ gridfs_store_buffer( gfs, data_before, i, "input-buffer", "text/html", GRIDFILE_COMPRESS ); test_gridfile( gfs, data_before, i, "input-buffer", "text/html" ); /* Input from file */ fd = fopen( "input-file", "w" ); fwrite( data_before, sizeof( char ), (size_t)i, fd ); fclose( fd ); gridfs_store_file( gfs, "input-file", "input-file", "text/html", GRIDFILE_DEFAULT ); test_gridfile( gfs, data_before, i, "input-file", "text/html" ); gfs->caseInsensitive = 1; gridfs_store_file( gfs, "input-file", "input-file", "text/html", GRIDFILE_DEFAULT ); test_gridfile( gfs, data_before, i, "inPut-file", "text/html" ); } gridfs_destroy( gfs ); mongo_disconnect( conn ); mongo_destroy( conn ); free( data_before ); /* Clean up files. */ gridfs_test_unlink( "input-file" ); gridfs_test_unlink( "output" ); } void test_delete( void ) { mongo conn[1]; gridfs gfs[1]; gridfile gfile[1]; char *data = (char*)bson_malloc( 1024 ); const char *testFile = "test-delete"; INIT_SOCKETS_FOR_WINDOWS; CONN_CLIENT_TEST; GFS_INIT; memset( data, '\0', 1024 ); ASSERT( gridfs_store_buffer( gfs, data, 1024, testFile, "text/html", GRIDFILE_DEFAULT ) == MONGO_OK ); ASSERT( gridfs_find_filename( gfs, testFile, gfile ) == MONGO_OK ); gridfile_destroy( gfile ); ASSERT( gridfs_remove_filename( gfs, testFile ) == MONGO_OK ); ASSERT( gridfs_find_filename( gfs, testFile, gfile ) == MONGO_ERROR ); ASSERT( gridfs_find_filename( gfs, "bogus-file-does-not-exist", gfile ) == MONGO_ERROR ); ASSERT( gridfs_remove_filename( gfs, "bogus-file-does-not-exist" ) == MONGO_ERROR ); gridfs_destroy( gfs ); mongo_disconnect( conn ); mongo_destroy( conn ); bson_free( data ); } void test_streaming( void ) { mongo conn[1]; gridfs gfs[1]; gridfile gfile[1]; char *medium = (char*)bson_malloc( 2*MEDIUM ); char *small = (char*)bson_malloc( LOWER ); char *buf = (char*)bson_malloc( LARGE ); int n; if( buf == NULL || small == NULL ) { printf( "Failed to allocate" ); exit( 1 ); } srand( (unsigned int)time( NULL ) ); INIT_SOCKETS_FOR_WINDOWS; CONN_CLIENT_TEST; fill_buffer_randomly( medium, ( int64_t )2 * MEDIUM ); fill_buffer_randomly( small, ( int64_t )LOWER ); fill_buffer_randomly( buf, ( int64_t )LARGE ); GFS_INIT; gridfile_init( gfs, NULL, gfile ); gridfile_writer_init( gfile, gfs, "medium", "text/html", GRIDFILE_DEFAULT ); ASSERT( gridfile_write_buffer( gfile, medium, MEDIUM ) == MEDIUM); ASSERT( gridfile_write_buffer( gfile, medium + MEDIUM, MEDIUM ) == MEDIUM); gridfile_writer_done( gfile ); test_gridfile( gfs, medium, 2 * MEDIUM, "medium", "text/html" ); gridfs_destroy( gfs ); GFS_INIT; gridfs_store_buffer( gfs, small, LOWER, "small", "text/html", GRIDFILE_DEFAULT ); test_gridfile( gfs, small, LOWER, "small", "text/html" ); gridfs_destroy( gfs ); GFS_INIT; gridfs_remove_filename( gfs, "large" ); gridfile_writer_init( gfile, gfs, "large", "text/html", GRIDFILE_DEFAULT ); for( n=0; n < ( LARGE / 1024 ); n++ ) { ASSERT( gridfile_write_buffer( gfile, buf + ( n * 1024 ), 1024 ) == 1024 ); } gridfile_writer_done( gfile ); test_gridfile( gfs, buf, LARGE, "large", "text/html" ); gridfs_destroy( gfs ); gridfile_destroy( gfile ); mongo_destroy( conn ); free( buf ); free( small ); free( medium ); } void test_random_write() { mongo conn[1]; gridfs gfs[1]; gridfile* gfile; char *data_before = (char*)bson_malloc( UPPER ); char *random_data = (char*)bson_malloc( UPPER ); char *buf = (char*) bson_malloc( UPPER ); int64_t i; FILE *fd; srand((unsigned int) time( NULL ) ); INIT_SOCKETS_FOR_WINDOWS; CONN_CLIENT_TEST; GFS_INIT; fill_buffer_randomly( data_before, UPPER ); fill_buffer_randomly( random_data, UPPER ); for ( i = LOWER; i <= UPPER; i += DELTA ) { int64_t j = i / 2 - 3; int n, bytes_to_write_first; /* Input from buffer */ gridfs_store_buffer( gfs, data_before, i, "input-buffer", "text/html", GRIDFILE_DEFAULT ); if ( i > DEFAULT_CHUNK_SIZE * 4 ) { n = DEFAULT_CHUNK_SIZE * 3 + 6; memcpy(&data_before[j], random_data, n); // Let's overwrite the buffer with bytes crossing multiple chunks bytes_to_write_first = 10; } else { n = 6; memcpy(random_data, "123456", n); strncpy(&data_before[j], random_data, n); // Let's overwrite the buffer with a few some bytes bytes_to_write_first = 0; } gfile = gridfile_create(); ASSERT(gridfs_find_filename(gfs, "input-buffer", gfile) == 0); gridfile_writer_init(gfile, gfs, "input-buffer", "text/html", GRIDFILE_DEFAULT ); gridfile_seek(gfile, j); // Seek into the same buffer position within the GridFS file if ( bytes_to_write_first ) { ASSERT( gridfile_write_buffer(gfile, random_data, bytes_to_write_first) == bytes_to_write_first ); // Let's write 10 bytes first, and later the rest } ASSERT( gridfile_write_buffer(gfile, &random_data[bytes_to_write_first], n - bytes_to_write_first) == n - bytes_to_write_first ); // Try to write to the existing GridFS file on the position given by j gridfile_seek(gfile, j); gridfile_read_buffer( gfile, buf, n ); ASSERT(memcmp( buf, &data_before[j], n) == 0); gridfile_writer_done(gfile); ASSERT(gfile->pos == j + n); gridfile_destroy(gfile); gridfile_dealloc(gfile); test_gridfile( gfs, data_before, j + n > i ? j + n : i, "input-buffer", "text/html" ); /* Input from file */ fd = fopen( "input-file", "w" ); fwrite( data_before, sizeof( char ), (size_t) (j + n > i ? j + n : i), fd ); fclose( fd ); gridfs_store_file( gfs, "input-file", "input-file", "text/html", GRIDFILE_DEFAULT ); test_gridfile( gfs, data_before, j + n > i ? j + n : i, "input-file", "text/html" ); } gridfs_destroy( gfs ); mongo_disconnect( conn ); mongo_destroy( conn ); free( data_before ); free( random_data ); free( buf ); /* Clean up files. */ gridfs_test_unlink( "input-file" ); gridfs_test_unlink( "output" ); } void test_random_write2( void ) { mongo conn[1]; gridfs gfs[1]; gridfile gfile[1]; bson meta; char *buf = (char*)bson_malloc( LARGE ); char *zeroedbuf = (char*)bson_malloc( LARGE ); int n; if( buf == NULL ) { printf( "Failed to allocate" ); exit( 1 ); } srand( 123 ); // Init with a predictable value INIT_SOCKETS_FOR_WINDOWS; CONN_CLIENT_TEST; fill_buffer_randomly( buf, ( int64_t )LARGE ); memset( zeroedbuf, 0, LARGE ); bson_init_empty( &meta ); GFS_INIT; /* This portion of the test we will write zeroes by using new API gridfile_set_size function gridfile_expand tested implicitally by using gridfile_set_size making file larger */ gridfile_init( gfs, &meta, gfile ); gridfile_writer_init( gfile, gfs, "random_access", "text/html", 0 ); gridfile_set_size( gfile, LARGE ); // New API, this zero fills the file gridfile_writer_done( gfile ); test_gridfile( gfs, zeroedbuf, LARGE, "random_access", "text/html" ); // Test zero filled file /* This portion of the test we will write zeroes by using new API gridfile_set_size then we will truncate the file */ gridfile_destroy( gfile ); gridfile_init( gfs, &meta, gfile ); gridfile_writer_init( gfile, gfs, "random_access", "text/html", 0 ); gridfile_set_size( gfile, LARGE ); // New API, this zero fills the file with LARGE bytes gridfile_truncate( gfile, LARGE / 2 ); // Let's truncate the file now gridfile_writer_done( gfile ); test_gridfile( gfs, zeroedbuf, LARGE / 2, "random_access", "text/html" ); // Test zero filled file truncated by half /* Let's re-create the file, now let's randomly write real data */ gridfile_destroy( gfile ); gridfile_init( gfs, &meta, gfile ); gridfile_writer_init( gfile, gfs, "random_access", "text/html", 0 ); gridfile_set_size( gfile, LARGE ); // We need to reserve LARGE bytes on file before writing backwards /* Here we will write backwards the file with our random numbers. We will use a 3072 size buffer to cross over buffers given the fact 256K is not a multiple of 3072. Let's stress the buffering logic a little bit */ for( n = LARGE / 3072 - 1; n >= 0; n-- ) { gridfile_seek( gfile, 3072 * n ); ASSERT( gridfile_write_buffer( gfile, buf + ( n * 3072 ), 3072 ) == 3072 ); } gridfile_writer_done( gfile ); test_gridfile( gfs, buf, LARGE, "random_access", "text/html" ); gridfs_destroy( gfs ); gridfile_destroy( gfile ); mongo_destroy( conn ); free( buf ); free( zeroedbuf ); } void test_large( void ) { mongo conn[1]; gridfs gfs[1]; gridfile gfile[1]; FILE *fd; size_t i, n; char *buffer = (char*)bson_malloc( LARGE ); char *read_buf = (char*)bson_malloc( LARGE ); int64_t filesize = ( int64_t )1024 * ( int64_t )LARGE; mongo_write_concern wc; bson lastError; bson lastErrorCmd; srand( (unsigned int) time( NULL ) ); INIT_SOCKETS_FOR_WINDOWS; CONN_CLIENT_TEST; mongo_write_concern_init(&wc); wc.j = 1; mongo_write_concern_finish(&wc); mongo_set_write_concern(conn, &wc); GFS_INIT; fd = fopen( "bigfile", "r" ); if( fd ) { fclose( fd ); } else { /* Create a very large file */ fill_buffer_randomly( buffer, ( int64_t )LARGE ); fd = fopen( "bigfile", "w" ); for( i=0; i<1024; i++ ) { fwrite( buffer, 1, LARGE, fd ); } fclose( fd ); } /* Now read the file into GridFS */ gridfs_remove_filename( gfs, "bigfile" ); gridfs_store_file( gfs, "bigfile", "bigfile", "text/html", GRIDFILE_NOMD5 | GRIDFILE_COMPRESS); gridfs_find_filename( gfs, "bigfile", gfile ); ASSERT( strcmp( gridfile_get_filename( gfile ), "bigfile" ) == 0 ); ASSERT( gridfile_get_contentlength( gfile ) == filesize ); fd = fopen( "bigfile", "r" ); while( ( n = fread( buffer, 1, MEDIUM, fd ) ) != 0 ) { ASSERT( gridfile_read_buffer( gfile, read_buf, MEDIUM ) == n ); ASSERT( memcmp( buffer, read_buf, n ) == 0 ); } fclose( fd ); gridfile_destroy( gfile ); /* Read the file using the streaming interface */ gridfs_remove_filename( gfs, "bigfile" ); gridfs_remove_filename( gfs, "bigfile-stream" ); gridfile_writer_init( gfile, gfs, "bigfile-stream", "text/html", GRIDFILE_NOMD5 | GRIDFILE_COMPRESS ); mongo_write_concern_destroy( &wc ); mongo_write_concern_init(&wc); wc.j = 0; /* Let's reset write concern j field to zero, we will manually call getLastError with j = 1 */ mongo_write_concern_finish(&wc); mongo_set_write_concern(conn, &wc); fd = fopen( "bigfile", "r" ); i = 0; while( ( n = fread( buffer, 1, READ_WRITE_BUF_SIZE, fd ) ) != 0 ) { ASSERT( gridfile_write_buffer( gfile, buffer, n ) == n ); if(i++ % 10 == 0) { bson_init( &lastErrorCmd ); bson_append_int( &lastErrorCmd, "getLastError", 1); bson_append_int( &lastErrorCmd, "j", 1); bson_finish( &lastErrorCmd ); bson_init( &lastError ); mongo_run_command( conn, "test", &lastErrorCmd, &lastError ); bson_destroy( &lastError ); bson_destroy( &lastErrorCmd ); } } mongo_write_concern_destroy( &wc ); mongo_write_concern_init(&wc); wc.j = 1; /* Let's reset write concern j field to 1 */ mongo_write_concern_finish(&wc); mongo_set_write_concern(conn, &wc); fclose( fd ); gridfile_writer_done( gfile ); gridfs_find_filename( gfs, "bigfile-stream", gfile ); ASSERT( strcmp( gridfile_get_filename( gfile ), "bigfile-stream" ) == 0 ); ASSERT( gridfile_get_contentlength( gfile ) == filesize ); gridfs_remove_filename( gfs, "bigfile-stream" ); gridfs_destroy( gfs ); mongo_disconnect( conn ); mongo_destroy( conn ); bson_free( buffer ); bson_free( read_buf ); mongo_write_concern_destroy( &wc ); } int main( void ) { /* See https://jira.mongodb.org/browse/CDRIVER-126 * on why we exclude this test from running on WIN32 */ test_basic(); test_delete(); test_streaming(); test_random_write(); test_random_write2(); /* Normally not necessary to run test_large(), as it * deals with very large (5GB) files and is therefore slow. */ /*test_large();*/ return 0; }