/* * Copyright (C) 1997 Pharos IP Pty Ltd * $Id: GdbmFile.java,v 3.12 1997-10-28 17:23:56+10 mbp Exp $ * * This program is free software; you can redistribute it and/or modify * it under the terms of the GNU General Public License as published by * the Free Software Foundation; either version 2 of the License, or * (at your option) any later version. * * This program is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the * GNU General Public License for more details. * * You should have received a copy of the GNU General Public License * along with this program; if not, write to the Free Software * Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA. */ package au.com.pharos.gdbm; import java.util.Enumeration; import java.util.NoSuchElementException; import au.com.pharos.packing.Packing; import au.com.pharos.packing.RawPacking; /** Java interface to a GDBM database table. * *
This database is a simple on-disk hash table. * *
Both the hash keys and values * are binary strings of any length. They are converted to and * from Java objects using customizable packing strategies. * *
The implementation of this class consists of two levels: a * Java-level interface, and a set of private native functions * implemented in C. As much functionality as possible is implemented * in Java: the C functions generally just marshal the information for * presentation to the native GDBM library. * *
The native library gdbmjava
must be available
* for dynamic loading in a system-dependant manner.
*
*
See the GDBM documentation file, and the * JavaGDBM home page * for more information. * * @see au.com.pharos.packing.Packing * @see au.com.pharos.gdbm.GdbmTest * @see au.com.pharos.gdbm.GdbmDictionary * * @author Martin Pool * @version $Revision: 3.12 $ **/ public class GdbmFile implements Closeable { /** The GDBM handle for the database file, or 0 if the database has * been closed. Java doesn't understand it, but it stores it and * passes it to the C routines. */ private transient volatile long dbf; // Remember the parameters used to open the database private int openFlags; private String dbFilename; // ----------------------------------------------------------------- // Constructors // Values must match those from gdbm.h ! /** Indicates that the caller will just read the database. Many * readers may share a database. */ public final static int READER = 0; /** The caller wants read/write access to an existing database and * requires exclusive access. */ public final static int WRITER = 1; /** The caller wants exclusive read/write access, and the database * should be created if it does not already exist. */ public final static int WRCREAT = 2; /** The caller wants exclusive read/write access, and the database * should be replaced if it already exists. */ public final static int NEWDB = 3; /** Flag indicating GDBM should write to the database without disc * synchronization. This allows faster writes, but may produce an * inconsistent database in the event of abnormal termination of * the writer. */ public final static int FAST = 16; // TODO: Allow FAST mode to be toggled /** Creates a new GdbmFile object representing an disk database. The * database is opened or created on disk. * *
Both key and value strategies default to * RawPacking, meaning that the database will use raw byte arrays * for both key and value. * *
TODO: Allow the caller to specify the block size and/or * cache size. * * @param fileName the disk filename in which the data will be stored. * * @param flags any of READER, WRITER, WRCREAT, and NEWDB, optionally * ORed with FAST * * @exception GdbmException if the database cannot be opened or * created. * * @see GdbmFile#READER * @see GdbmFile#WRITER * @see GdbmFile#WRCREAT * @see GdbmFile#NEWDB * @see GdbmFile#FAST */ public GdbmFile(String fileName, int flags) throws GdbmException { openFlags = flags; dbFilename = fileName; keyPacking = new RawPacking(); valuePacking = new RawPacking(); dbf = gdbm_open(fileName, flags); } /** Close the database file if it is still open. * *
Note that the disk file is locked against access from * other processes while it is open. To prevent contention, it * may be useful to explicitly close the file rather than waiting * for the garbage-collector. * * @see GdbmFile#isOpen() **/ public synchronized void close() throws GdbmException { if (dbf != 0) gdbm_close(dbf); dbf = 0; // No longer connected } /** Close the database when the GdbmFile is garbage-collected. * * @see GdbmFile#close() */ public void finalize() throws GdbmException { close(); } /* The Gdbm file is unlocked when the process terminates, but * nevertheless it would be good to close it in finalization, in * case the OS process continues after the JVM terminates. * * XXX: Unfortunately, anybody else can turn this off again: it * would be nice if there was a way to avoid this. */ static { System.runFinalizersOnExit(true); } // ----------------------------------------------------------------- // Packing strategies private Packing keyPacking, valuePacking; /** Set the object to be used as the packing strategy for * converting key objects to and from the byte arrays stored in * the database. * *
Depending on the class of Java object used as the key of your
* database, it may be appropriate to use a packing strategy from
* the
* au.com.pharos.packing
* package, or to define a new subclass of one of those strategies.
*
* @see au.com.pharos.packing.Packing
* @see GdbmFile#setValuePacking(au.com.pharos.packing.Packing)
**/
public void setKeyPacking(Packing newPacking) {
keyPacking = newPacking;
}
/** Set the object to be used as the packing strategy for
* converting value objects to and from the byte arrays stored
* in the database.
*
*
Depending on the class of Java object used as the key of your
* database, it may be appropriate to use a packing strategy from
* the
* au.com.pharos.packing
* package, or to define a new subclass of one of those strategies.
*
* @see au.com.pharos.packing.Packing
* @see GdbmFile#setKeyPacking(au.com.pharos.packing.Packing )
**/
public void setValuePacking(Packing newPacking) {
valuePacking = newPacking;
}
// -----------------------------------------------------------------
/** Returns a string indicating the version of the underlying GDBM
* library.
*
* @return the version of the native GDBM library.
**/
public static String getLibraryVersion()
{
return gdbm_getversion();
}
/** Return a string indicating the version of the JavaGDBM library
* wrapper.
*
*
The most current release is available from the * JavaGDBM home page. * * @return the release number of the JavaGDBM library **/ public static String getWrapperVersion() { return gdbm_wrapperVersion(); } /** Indicate whether the database is writable. * *
Databases are opened in either read-write or read-only mode, * and remain in that mode until they are closed. * * @return true if the database may be written; otherwise false. * * @see GdbmFile#GdbmFile(java.lang.String, int) **/ public boolean isWritable() { return (openFlags & 0x03) != READER; } /** Indicate whether the database is open or not. * *
A database is open from the point of creation until close() * is called, if ever, after which it is closed. * * @return false if the database has been closed; otherwise true. * * @see GdbmFile#close() **/ public boolean isOpen() { return dbf != 0; } /** Compact the database file. * *
If you have had a lot of deletions and would like to shrink the * space used by the GDBM file, this function will reorganize the * database. GDBM will not shorten the length of a GDBM file * (deleted file space will be reused) except by using this * reorganization, though it will reuse the vacant space. * *
The database must be writable. */ public void reorganize() throws GdbmException { gdbm_reorganize(dbf); } /** Flush changes to disk. * *
This function is only required when the database is * opened with the FAST flag set. By default, changes are * flushed to disk after every update. * *
The database must be writable. * * @see GdbmFile#FAST **/ public void sync() throws GdbmException { gdbm_sync(dbf); } /** Retrieve the value corresponding to a particular key. * * @param key the key of the record to be retrieved * * @return the value of the record with the specified key; or * null if the key is not present in the database. */ public Object fetch(Object key) throws GdbmException { byte[] keyBytes = keyPacking.toBytes(key); return valuePacking.fromBytes(gdbm_fetch(dbf, keyBytes)); } /** Indicate whether the specified key is in the database, * without returning the value. * * @param key the key of the record to be retrieved * * @return true if a record with the specified key is present; * otherwise false. */ public boolean exists(Object key) throws GdbmException { byte[] keyBytes = keyPacking.toBytes(key); return gdbm_exists(dbf, keyBytes); } /** Store a value in the database, replacing any existing value * with the same key. * * @param key key under which to store the value * * @param value value to be stored * * @exception GdbmException if the object is a reader; or * if an IO error occurs **/ public void store(Object key, Object value) throws GdbmException { byte[] keyBytes = keyPacking.toBytes(key); byte[] valueBytes = valuePacking.toBytes(value); gdbm_store(dbf, keyBytes, valueBytes, true); } /** Store a value in the database, unless a record with the same * key already exists. * * @param key key under which to store the value. * * @param value value to be stored. * * @exception GdbmException if a record with the specified key * already exists; or if the object is a reader; or * if an IO error occurs */ public void storeNoReplace(Object key, Object value) throws GdbmException { byte[] keyBytes = keyPacking.toBytes(key); byte[] valueBytes = valuePacking.toBytes(value); gdbm_store(dbf, keyBytes, valueBytes, false); } /** Remove a record from the database. * * @param key key of the record to be removed * * @exception GdbmException if there is no record with the * specified key; or if the object is a reader; or if an IO error * occurs. **/ public void delete(Object key) throws GdbmException { byte[] keyBytes = keyPacking.toBytes(key); gdbm_delete(dbf, keyBytes); } // ----------------------------------------------------------------- // Iterator support // TODO: Use a flag/timestamp approach to throw an exception if // somebody modifies the database while they're trying to iterate // over it? Approach suggested by Doug Lea's Collections API. -- // mbp /** Return an enumeration which will return all of the keys for the * database of the file in (apparently) random order. * *
Caution: If the database is modified while * an enumeration is in progress, * then changes in the hash table may cause the enumeration to * miss some records. * * @see java.util.Enumeration **/ public Enumeration keys() throws GdbmException { // Return an inner class which will do the iteration in the // context of this GdbmFile object. return new KeyEnumeration(); } // Synchronization is performed in the GdbmFile's methods class KeyEnumeration implements Enumeration { private byte[] currKey, nextKey; KeyEnumeration() throws GdbmException { currKey = null; nextKey = getFirstKeyRaw(); } public boolean hasMoreElements() { return nextKey != null; } public Object nextElement() throws NoSuchElementException { try { currKey = nextKey; if (currKey == null) throw new NoSuchElementException(); else nextKey = getNextKeyRaw(currKey); return keyPacking.fromBytes(currKey); } catch ( GdbmException e ) { // Can't propagate this through java.util.Enumeration, dammit throw new NoSuchElementException(); } } } /** Start a visit to all keys in the hashtable, using raw data values. * * @return the first key in the hash table as a byte array; or null if * the database is empty. */ byte[] getFirstKeyRaw() throws GdbmException { return gdbm_firstkey(dbf); } /** Return the first record in the hashtable. * *
Note that the database * is not ordered, so the key returned is simply the first in the * hashtable and effectively randomly selected. * * @return the first key in the hash table; or null if * the database is empty. * * @see GdbmFile#getNextKey(java.lang.Object) * @see GdbmFile#keys() **/ Object getFirstKey() throws GdbmException { return keyPacking.fromBytes(getFirstKeyRaw()); } /** Find and read the next key in the hashtable. * * @return the next key; or null if keyBytes is the last key * in the hashtable. **/ byte[] getNextKeyRaw(byte[] keyBytes) throws GdbmException { return gdbm_nextkey(dbf, keyBytes); } /** Check whether the database is empty. * * @return true if the database contains no records; otherwise * false. **/ public boolean isEmpty() throws GdbmException { return getFirstKeyRaw() == null; } /** Count the records in the database. * *
This is implemented by iterating over the database, and so is * a fairly expensive operation. * *
This method locks the database to make sure the count is * accurate. * * @return the number of records in the database * @see GdbmFile#isEmpty() **/ public synchronized int size() throws GdbmException { int s = 0; byte[] key = getFirstKeyRaw(); while (key != null) { s++; key = getNextKeyRaw(key); } return s; } static { String libraryFile = System.getProperty ("au.com.pharos.gdbm.libraryFile"); if (libraryFile != null) { System.load(libraryFile); } else { System.loadLibrary("gdbmjava"); } } private synchronized native long gdbm_open(String fileName, int flags) throws GdbmException; private synchronized native void gdbm_close(long dbf); private synchronized native void gdbm_store(long dbf, byte[] key, byte[] content, boolean replace); private synchronized native byte[] gdbm_fetch(long dbf, byte[] key); private synchronized native boolean gdbm_exists(long dbf, byte[] key); private synchronized native void gdbm_delete(long dbf, byte[] key); private synchronized native byte[] gdbm_firstkey(long dbf) throws GdbmException; private synchronized native byte[] gdbm_nextkey(long dbf, byte[] key) throws GdbmException; private synchronized static native String gdbm_getversion(); private synchronized static native String gdbm_wrapperVersion(); private synchronized native void gdbm_reorganize(long dbf); private synchronized native void gdbm_sync(long dbf); }