--- a/src/org/tmatesoft/hg/repo/HgChangelog.java	Fri Jul 12 16:29:06 2013 +0200
+++ b/src/org/tmatesoft/hg/repo/HgChangelog.java	Fri Jul 12 20:14:24 2013 +0200
@@ -20,10 +20,8 @@
 import java.util.ArrayList;
 import java.util.Arrays;
 import java.util.Calendar;
-import java.util.Collections;
 import java.util.Date;
 import java.util.Formatter;
-import java.util.HashMap;
 import java.util.List;
 import java.util.Locale;
 import java.util.Map;
@@ -32,12 +30,10 @@
 import org.tmatesoft.hg.core.Nodeid;
 import org.tmatesoft.hg.core.SessionContext;
 import org.tmatesoft.hg.internal.Callback;
+import org.tmatesoft.hg.internal.ChangesetParser;
 import org.tmatesoft.hg.internal.DataAccess;
-import org.tmatesoft.hg.internal.EncodingHelper;
-import org.tmatesoft.hg.internal.Internals;
 import org.tmatesoft.hg.internal.Lifecycle;
 import org.tmatesoft.hg.internal.LifecycleBridge;
-import org.tmatesoft.hg.internal.Pool;
 import org.tmatesoft.hg.internal.RevlogStream;
 import org.tmatesoft.hg.util.Adaptable;
 import org.tmatesoft.hg.util.CancelSupport;
@@ -150,14 +146,14 @@
 	 */
 	public static final class RawChangeset implements Cloneable /* for those that would like to keep a copy */{
 		// would be nice to get it immutable, but then we can't reuse instances
-		private/* final */Nodeid manifest;
-		private String user;
-		private String comment;
-		private String[] files; // shall not be modified (#clone() does shallow copy)
-		private Date time;
-		private int timezone;
+		/* final */Nodeid manifest;
+		String user;
+		String comment;
+		String[] files; // shall not be modified (#clone() does shallow copy)
+		Date time;
+		int timezone;
 		// http://mercurial.selenic.com/wiki/PruningDeadBranches - Closing changesets can be identified by close=1 in the changeset's extra field.
-		private Map<String, String> extras;
+		Map<String, String> extras;
 
 		private RawChangeset() {
 		}
@@ -241,170 +237,33 @@
 		}
 	}
 	
-	/**
-	 * @see mercurial/changelog.py:read()
-	 * 
-	 *      <pre>
-	 *         format used:
-	 *         nodeid\n        : manifest node in ascii
-	 *         user\n          : user, no \n or \r allowed
-	 *         time tz extra\n : date (time is int or float, timezone is int)
-	 *                         : extra is metadatas, encoded and separated by '\0'
-	 *                         : older versions ignore it
-	 *         files\n\n       : files modified by the cset, no \n or \r allowed
-	 *         (.*)            : comment (free text, ideally utf-8)
-	 * 
-	 *         changelog v0 doesn't use extra
-	 * </pre>
-	 */
-	/*package-local*/static final class ChangesetParser {
-		private final EncodingHelper encHelper;
-		// it's likely user names get repeated again and again throughout repository. 
-		private final Pool<String> usersPool;
-		private final Pool<String> filesPool;
-		private final boolean reuseChangesetInstance;
-		private RawChangeset target;
-		
-		public ChangesetParser(SessionContext.Source sessionContex, boolean shallReuseCsetInstance) {
-			encHelper = Internals.buildFileNameEncodingHelper(sessionContex);
-			usersPool = new Pool<String>();
-			filesPool = new Pool<String>();
-			reuseChangesetInstance = shallReuseCsetInstance;
+	/*package-local*/static final class RawCsetFactory implements ChangesetParser.CsetFactory {
+		private RawChangeset cset;
+
+		public RawCsetFactory(boolean shallReuseCsetInstance) {
 			if (shallReuseCsetInstance) {
-				target = new RawChangeset();
+				cset = new RawChangeset();
 			}
 		}
-		
-		public void dispose() {
-			usersPool.clear();
-			filesPool.clear();
-		}
-
-		public RawChangeset parse(DataAccess da) throws IOException, HgInvalidDataFormatException {
-			byte[] data = da.byteArray();
-			if (!reuseChangesetInstance) {
-				target = new RawChangeset();
-			}
-			init(data, 0, data.length);
-			return target;
-		}
 
-		private void init(byte[] data, int offset, int length) throws HgInvalidDataFormatException {
-			final int bufferEndIndex = offset + length;
-			final byte lineBreak = (byte) '\n';
-			int breakIndex1 = indexOf(data, lineBreak, offset, bufferEndIndex);
-			if (breakIndex1 == -1) {
-				throw new HgInvalidDataFormatException("Bad Changeset data");
-			}
-			Nodeid _nodeid = Nodeid.fromAscii(data, 0, breakIndex1);
-			int breakIndex2 = indexOf(data, lineBreak, breakIndex1 + 1, bufferEndIndex);
-			if (breakIndex2 == -1) {
-				throw new HgInvalidDataFormatException("Bad Changeset data");
-			}
-			String _user;
-			_user = encHelper.userFromChangeset(data, breakIndex1 + 1, breakIndex2 - breakIndex1 - 1);
-			_user = usersPool.unify(_user);
-
-			int breakIndex3 = indexOf(data, lineBreak, breakIndex2 + 1, bufferEndIndex);
-			if (breakIndex3 == -1) {
-				throw new HgInvalidDataFormatException("Bad Changeset data");
-			}
-			String _timeString = new String(data, breakIndex2 + 1, breakIndex3 - breakIndex2 - 1);
-			int space1 = _timeString.indexOf(' ');
-			if (space1 == -1) {
-				throw new HgInvalidDataFormatException(String.format("Bad Changeset data: %s in [%d..%d]", "time string", breakIndex2+1, breakIndex3));
-			}
-			int space2 = _timeString.indexOf(' ', space1 + 1);
-			if (space2 == -1) {
-				space2 = _timeString.length();
+		public RawChangeset create(Nodeid nodeidManifest, String user, Date time, int timezone, List<String> files, String comment, Map<String, String> extrasMap) {
+			RawChangeset target;
+			if (cset != null) {
+				target = cset;
+			} else {
+				target = new RawChangeset();
 			}
-			long unixTime = Long.parseLong(_timeString.substring(0, space1));
-			int _timezone = Integer.parseInt(_timeString.substring(space1 + 1, space2));
-			// unixTime is local time, and timezone records difference of the local time to UTC.
-			Date _time = new Date(unixTime * 1000);
-			String _extras = space2 < _timeString.length() ? _timeString.substring(space2 + 1) : null;
-			Map<String, String> _extrasMap = parseExtras(_extras);
-			//
-			int lastStart = breakIndex3 + 1;
-			int breakIndex4 = indexOf(data, lineBreak, lastStart, bufferEndIndex);
-			ArrayList<String> _files = null;
-			if (breakIndex4 > lastStart) {
-				// if breakIndex4 == lastStart, we already found \n\n and hence there are no files (e.g. merge revision)
-				_files = new ArrayList<String>(5);
-				while (breakIndex4 != -1 && breakIndex4 + 1 < bufferEndIndex) {
-					String fname = encHelper.fileFromChangeset(data, lastStart, breakIndex4 - lastStart);
-					_files.add(filesPool.unify(fname));
-					lastStart = breakIndex4 + 1;
-					if (data[breakIndex4 + 1] == lineBreak) {
-						// found \n\n
-						break;
-					} else {
-						breakIndex4 = indexOf(data, lineBreak, lastStart, bufferEndIndex);
-					}
-				}
-				if (breakIndex4 == -1 || breakIndex4 >= bufferEndIndex) {
-					throw new HgInvalidDataFormatException("Bad Changeset data");
-				}
-			} else {
-				breakIndex4--;
-			}
-			String _comment = encHelper.commentFromChangeset(data, breakIndex4 + 2, bufferEndIndex - breakIndex4 - 2);
-			// change this instance at once, don't leave it partially changes in case of error
-			target.manifest = _nodeid;
-			target.user = _user;
-			target.time = _time;
-			target.timezone = _timezone;
-			target.files = _files == null ? new String[0] : _files.toArray(new String[_files.size()]);
-			target.comment = _comment;
-			target.extras = _extrasMap;
-		}
-
-		private Map<String, String> parseExtras(String _extras) {
-			final String extras_branch_key = "branch";
-			_extras = _extras == null ? null : _extras.trim();
-			if (_extras == null || _extras.length() == 0) {
-				return Collections.singletonMap(extras_branch_key, HgRepository.DEFAULT_BRANCH_NAME);
-			}
-			Map<String, String> _extrasMap = new HashMap<String, String>();
-			int lastIndex = 0;
-			do {
-				String pair;
-				int sp = _extras.indexOf('\0', lastIndex);
-				if (sp == -1) {
-					sp = _extras.length();
-				}
-				if (sp > lastIndex) {
-					pair = _extras.substring(lastIndex, sp);
-					pair = decode(pair);
-					int eq = pair.indexOf(':');
-					_extrasMap.put(pair.substring(0, eq), pair.substring(eq + 1));
-					lastIndex = sp + 1;
-				}
-			} while (lastIndex < _extras.length());
-			if (!_extrasMap.containsKey(extras_branch_key)) {
-				_extrasMap.put(extras_branch_key, HgRepository.DEFAULT_BRANCH_NAME);
-			}
-			return Collections.unmodifiableMap(_extrasMap);
-		}
-
-		private static int indexOf(byte[] src, byte what, int startOffset, int endIndex) {
-			for (int i = startOffset; i < endIndex; i++) {
-				if (src[i] == what) {
-					return i;
-				}
-			}
-			return -1;
-		}
-		
-		private static String decode(String s) {
-			if (s != null && s.indexOf('\\') != -1) {
-				// TestAuxUtilities#testChangelogExtrasDecode
-				return s.replace("\\\\", "\\").replace("\\n", "\n").replace("\\r", "\r").replace("\\0", "\00");
-			}
-			return s;
+			target.manifest = nodeidManifest;
+			target.user = user;
+			target.time = time;
+			target.timezone = timezone;
+			target.files = files == null ? new String[0] : files.toArray(new String[files.size()]);
+			target.comment = comment;
+			target.extras = extrasMap;
+			return target;
 		}
 	}
-
+	
 	private static class RawCsetCollector implements Inspector {
 		final ArrayList<RawChangeset> result;
 		
@@ -430,7 +289,7 @@
 		public RawCsetParser(SessionContext.Source sessionContext, HgChangelog.Inspector delegate) {
 			assert delegate != null;
 			inspector = delegate;
-			csetBuilder = new ChangesetParser(sessionContext, true);
+			csetBuilder = new ChangesetParser(sessionContext, new RawCsetFactory(true));
 			inspectorLifecycle = Adaptable.Factory.getAdapter(delegate, Lifecycle.class, null);
 			if (inspectorLifecycle == null) {
 				ProgressSupport ph = Adaptable.Factory.getAdapter(delegate, ProgressSupport.class, null);