tikhomirov@112: /*
tikhomirov@710:  * Copyright (c) 2011-2013 TMate Software Ltd
tikhomirov@112:  *  
tikhomirov@112:  * This program is free software; you can redistribute it and/or modify
tikhomirov@112:  * it under the terms of the GNU General Public License as published by
tikhomirov@112:  * the Free Software Foundation; version 2 of the License.
tikhomirov@112:  *
tikhomirov@112:  * This program is distributed in the hope that it will be useful,
tikhomirov@112:  * but WITHOUT ANY WARRANTY; without even the implied warranty of
tikhomirov@112:  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
tikhomirov@112:  * GNU General Public License for more details.
tikhomirov@112:  *
tikhomirov@112:  * For information on how to redistribute this software under
tikhomirov@112:  * the terms of a license other than GNU General Public License
tikhomirov@130:  * contact TMate Software at support@hg4j.com
tikhomirov@112:  */
tikhomirov@112: package org.tmatesoft.hg.internal;
tikhomirov@112: 
tikhomirov@456: import static org.tmatesoft.hg.util.LogFacility.Severity.Error;
tikhomirov@456: 
tikhomirov@112: import java.nio.ByteBuffer;
tikhomirov@114: import java.util.ArrayList;
tikhomirov@354: import java.util.Date;
tikhomirov@711: import java.util.Map;
tikhomirov@112: import java.util.TreeMap;
tikhomirov@112: 
tikhomirov@354: import org.tmatesoft.hg.core.Nodeid;
tikhomirov@154: import org.tmatesoft.hg.repo.HgChangelog.RawChangeset;
tikhomirov@112: import org.tmatesoft.hg.repo.HgRepository;
tikhomirov@423: import org.tmatesoft.hg.repo.HgRuntimeException;
tikhomirov@331: import org.tmatesoft.hg.util.Pair;
tikhomirov@133: import org.tmatesoft.hg.util.Path;
tikhomirov@112: 
tikhomirov@112: /**
tikhomirov@112:  *
tikhomirov@112:  * @author Artem Tikhomirov
tikhomirov@112:  * @author TMate Software Ltd.
tikhomirov@112:  */
tikhomirov@112: public class KeywordFilter implements Filter {
tikhomirov@114: 	private final HgRepository repo;
tikhomirov@112: 	private final boolean isExpanding;
tikhomirov@711: 	private final Map<String,String> keywords;
tikhomirov@114: 	private final Path path;
tikhomirov@154: 	private RawChangeset latestFileCset;
tikhomirov@711: 	private final ByteVector unprocessedBuffer;
tikhomirov@112: 
tikhomirov@112: 	/**
tikhomirov@112: 	 * 
tikhomirov@114: 	 * @param hgRepo 
tikhomirov@114: 	 * @param path 
tikhomirov@112: 	 * @param expand <code>true</code> to expand keywords, <code>false</code> to shrink
tikhomirov@112: 	 */
tikhomirov@711: 	private KeywordFilter(HgRepository hgRepo, Path p, Map<String, String> kw, boolean expand) {
tikhomirov@114: 		repo = hgRepo;
tikhomirov@114: 		path = p;
tikhomirov@112: 		isExpanding = expand;
tikhomirov@711: 		keywords = kw;
tikhomirov@711: 		unprocessedBuffer = expand ? new ByteVector(0, 0) :  new ByteVector(120, 50);
tikhomirov@112: 	}
tikhomirov@112: 
tikhomirov@112: 	/**
tikhomirov@112: 	 * @param src buffer ready to be read
tikhomirov@112: 	 * @return buffer ready to be read and original buffer's position modified to reflect consumed bytes. IOW, if source buffer
tikhomirov@112: 	 * on return has remaining bytes, they are assumed not-read (not processed) and next chunk passed to filter is supposed to 
tikhomirov@112: 	 * start with them  
tikhomirov@112: 	 */
tikhomirov@112: 	public ByteBuffer filter(ByteBuffer src) {
tikhomirov@711: 		// when unprocessedBuffer is empty, we are looking for first $ in the input,
tikhomirov@711: 		// when we've already got anything unprocessed, newline is of interest, too
tikhomirov@711: 		int kwBreak = indexOf(src, '$', src.position(), !unprocessedBuffer.isEmpty());
tikhomirov@711: 		ByteBuffer outBuffer = null;
tikhomirov@711: 		while (kwBreak != -1) {
tikhomirov@711: 			if (unprocessedBuffer.isEmpty()) {
tikhomirov@711: 				// both expand and collapse cases
tikhomirov@711: 				assert src.get(kwBreak) == '$';
tikhomirov@711: 				
tikhomirov@711: 				int end = indexOf(src, '$', kwBreak+1, true);
tikhomirov@711: 				if (end == -1) {
tikhomirov@711: 					for (int i = kwBreak; i < src.limit(); i++) {
tikhomirov@711: 						unprocessedBuffer.add(src.get(i));
tikhomirov@112: 					}
tikhomirov@711: 					src.limit(kwBreak);
tikhomirov@711: 					kwBreak = -1;
tikhomirov@711: 					// src up to kwBreak is left and returned either with outBuffer or alone 
tikhomirov@711: 				} else if (src.get(end) == '$') {
tikhomirov@711: 					StringBuilder sb = new StringBuilder(end - kwBreak);
tikhomirov@711: 					for (int i = kwBreak+1; i < end; i++) {
tikhomirov@711: 						if (src.get(i) == ':' || src.get(i) == ' ') {
tikhomirov@711: 							break;
tikhomirov@112: 						}
tikhomirov@711: 						sb.append((char) src.get(i));
tikhomirov@112: 					}
tikhomirov@711: 					final String keyword = sb.toString();
tikhomirov@711: 					if (knownKeyword(keyword)) {
tikhomirov@711: 						// copy src up to kw, including starting $keyword
tikhomirov@711: 						outBuffer = append(outBuffer, src, kwBreak - src.position() + 1+keyword.length());
tikhomirov@711: 						// replace kwStart..end with new content
tikhomirov@711: 						outBuffer = ensureCapacityFor(outBuffer, (isExpanding ? 200 : 1));
tikhomirov@711: 						if (isExpanding) {
tikhomirov@711: 							outBuffer.put((byte) ':');
tikhomirov@711: 							outBuffer.put((byte) ' ');
tikhomirov@711: 							outBuffer = expandKeywordValue(keyword, outBuffer);
tikhomirov@711: 							outBuffer.put((byte) ' ');
tikhomirov@112: 						}
tikhomirov@711: 						outBuffer.put((byte) '$');
tikhomirov@711: 						// src is consumed up to end
tikhomirov@711: 						src.position(end+1);
tikhomirov@711: 						kwBreak = indexOf(src, '$', end+1, false);
tikhomirov@112: 					} else {
tikhomirov@711: 						// no (or unknown) keyword, try with '$' at src[end]
tikhomirov@711: 						kwBreak = end;
tikhomirov@112: 					}
tikhomirov@112: 				} else {
tikhomirov@711: 					// newline, ignore keyword start
tikhomirov@711: 					kwBreak = indexOf(src, '$', end+1, false);
tikhomirov@711: 				}
tikhomirov@711: 			} else {
tikhomirov@711: 				// we've got smth unprocessed, and we've matched either $ or NL
tikhomirov@711: 				// the only chance to get here is when src is in the very start
tikhomirov@711: 				if (src.get(kwBreak) == '$') {
tikhomirov@711: 					// closed tag
tikhomirov@711: 					for (int i = src.position(); i <= kwBreak; i++) {
tikhomirov@711: 						// consume src: going to handle its [position*()..kwBreak] as part of unprocessedBuffer
tikhomirov@711: 						unprocessedBuffer.add(src.get());   
tikhomirov@112: 					}
tikhomirov@711: 					StringBuilder sb = new StringBuilder(unprocessedBuffer.size());
tikhomirov@711: 					assert unprocessedBuffer.get(0) == '$';
tikhomirov@711: 					for (int i = 1; i < unprocessedBuffer.size(); i++) {
tikhomirov@711: 						char ch = (char) unprocessedBuffer.get(i);
tikhomirov@711: 						if (ch == ':' || ch == ' ') {
tikhomirov@711: 							break;
tikhomirov@711: 						}
tikhomirov@711: 						sb.append(ch);
tikhomirov@711: 					}
tikhomirov@711: 					final String keyword = sb.toString();
tikhomirov@711: 					if (knownKeyword(keyword)) {
tikhomirov@711: 						outBuffer = ensureCapacityFor(outBuffer, keyword.length() + (isExpanding ? 200 : 2));
tikhomirov@711: 						outBuffer.put((byte) '$');
tikhomirov@711: 						outBuffer.put(keyword.getBytes());
tikhomirov@711: 						if (isExpanding) {
tikhomirov@711: 							outBuffer.put((byte) ':');
tikhomirov@711: 							outBuffer.put((byte) ' ');
tikhomirov@711: 							outBuffer = expandKeywordValue(keyword, outBuffer);
tikhomirov@711: 							outBuffer.put((byte) ' ');
tikhomirov@711: 						}
tikhomirov@711: 						outBuffer.put((byte) '$');
tikhomirov@711: 					} else {
tikhomirov@711: 						outBuffer = append(outBuffer, unprocessedBuffer.toByteArray());
tikhomirov@711: 					}
tikhomirov@711: 					// src part is consumed already, do nothing here, look for next possible kw
tikhomirov@711: 					kwBreak = indexOf(src, '$', kwBreak+1, false);
tikhomirov@711: 				} else {
tikhomirov@711: 					// newline => tag without close
tikhomirov@711: 					outBuffer = append(outBuffer, unprocessedBuffer.toByteArray());
tikhomirov@711: 					kwBreak = indexOf(src, '$', kwBreak+1, false);
tikhomirov@112: 				}
tikhomirov@711: 				unprocessedBuffer.clear();
tikhomirov@112: 			}
tikhomirov@711: 		} while (kwBreak != -1);
tikhomirov@711: 		if (outBuffer == null) {
tikhomirov@711: 			return src;
tikhomirov@112: 		}
tikhomirov@711: 		outBuffer = ensureCapacityFor(outBuffer, src.remaining());
tikhomirov@711: 		outBuffer.put(src);
tikhomirov@711: 		outBuffer.flip();
tikhomirov@711: 		return outBuffer;
tikhomirov@711: 	}
tikhomirov@711: 	private boolean knownKeyword(String kw) {
tikhomirov@711: 		return keywords.containsKey(kw);
tikhomirov@711: 	}
tikhomirov@711: 
tikhomirov@711: 	private static ByteBuffer append(ByteBuffer out, byte[] data) {
tikhomirov@711: 		out = ensureCapacityFor(out, data.length);
tikhomirov@711: 		out.put(data);
tikhomirov@711: 		return out;
tikhomirov@711: 	}
tikhomirov@711: 	private static ByteBuffer append(ByteBuffer out, ByteBuffer in, int count) {
tikhomirov@711: 		out = ensureCapacityFor(out, count);
tikhomirov@711: 		while (count-- > 0) {
tikhomirov@711: 			out.put(in.get());
tikhomirov@711: 		}
tikhomirov@711: 		return out;
tikhomirov@711: 	}
tikhomirov@711: 	private static ByteBuffer ensureCapacityFor(ByteBuffer out, int exansion) {
tikhomirov@711: 		if (out == null || out.remaining() < exansion) {
tikhomirov@711: 			ByteBuffer newOut = ByteBuffer.allocate(out == null ? exansion*2 : out.capacity() + exansion);
tikhomirov@711: 			if (out != null) {
tikhomirov@711: 				out.flip();
tikhomirov@711: 				newOut.put(out);
tikhomirov@112: 			}
tikhomirov@711: 			return newOut;
tikhomirov@112: 		}
tikhomirov@711: 		return out;
tikhomirov@112: 	}
tikhomirov@112: 	
tikhomirov@711: 	private ByteBuffer expandKeywordValue(String keyword, ByteBuffer rv) {
tikhomirov@711: 		byte[] toInject;
tikhomirov@112: 		if ("Id".equals(keyword)) {
tikhomirov@711: 			toInject = identityString().getBytes();
tikhomirov@112: 		} else if ("Revision".equals(keyword)) {
tikhomirov@711: 			toInject = revision().getBytes();
tikhomirov@112: 		} else if ("Author".equals(keyword)) {
tikhomirov@711: 			toInject = username().getBytes();
tikhomirov@114: 		} else if ("Date".equals(keyword)) {
tikhomirov@711: 			toInject = date().getBytes();
tikhomirov@114: 		} else {
tikhomirov@114: 			throw new IllegalStateException(String.format("Keyword %s is not yet supported", keyword));
tikhomirov@112: 		}
tikhomirov@711: 		rv = ensureCapacityFor(rv, toInject.length);
tikhomirov@711: 		rv.put(toInject);
tikhomirov@711: 		return rv;
tikhomirov@112: 	}
tikhomirov@112: 	
tikhomirov@112: 	// copies part of the src buffer, [from..to). doesn't modify src position
tikhomirov@113: 	static void copySlice(ByteBuffer src, int from, int to, ByteBuffer dst) {
tikhomirov@112: 		if (to > src.limit()) {
tikhomirov@112: 			throw new IllegalArgumentException("Bad right boundary");
tikhomirov@112: 		}
tikhomirov@112: 		if (dst.remaining() < to - from) {
tikhomirov@112: 			throw new IllegalArgumentException("Not enough room in the destination buffer");
tikhomirov@112: 		}
tikhomirov@112: 		for (int i = from; i < to; i++) {
tikhomirov@112: 			dst.put(src.get(i));
tikhomirov@112: 		}
tikhomirov@112: 	}
tikhomirov@112: 
tikhomirov@112: 	private static int indexOf(ByteBuffer b, char ch, int from, boolean newlineBreaks) {
tikhomirov@113: 		for (int i = from; i < b.limit(); i++) {
tikhomirov@112: 			byte c = b.get(i);
tikhomirov@112: 			if (ch == c) {
tikhomirov@112: 				return i;
tikhomirov@112: 			}
tikhomirov@112: 			if (newlineBreaks && (c == '\n' || c == '\r')) {
tikhomirov@112: 				return i;
tikhomirov@112: 			}
tikhomirov@112: 		}
tikhomirov@112: 		return -1;
tikhomirov@112: 	}
tikhomirov@112: 
tikhomirov@112: 	private String identityString() {
tikhomirov@114: 		return String.format("%s,v %s %s %s", path, revision(), date(), username());
tikhomirov@112: 	}
tikhomirov@112: 
tikhomirov@114: 	private String revision() {
tikhomirov@354: 		try {
tikhomirov@418: 			// TODO post-1.0 Either add cset's nodeid into Changeset class or use own inspector 
tikhomirov@418: 			// when accessing changelog, see below, #getChangeset
tikhomirov@367: 			int csetRev = repo.getFileNode(path).getChangesetRevisionIndex(HgRepository.TIP);
tikhomirov@354: 			return repo.getChangelog().getRevision(csetRev).shortNotation();
tikhomirov@423: 		} catch (HgRuntimeException ex) {
tikhomirov@490: 			repo.getSessionContext().getLog().dump(getClass(), Error, ex, null);
tikhomirov@354: 			return Nodeid.NULL.shortNotation(); // XXX perhaps, might return anything better? Not sure how hg approaches this. 
tikhomirov@354: 		}
tikhomirov@112: 	}
tikhomirov@112: 	
tikhomirov@112: 	private String username() {
tikhomirov@354: 		try {
tikhomirov@354: 			return getChangeset().user();
tikhomirov@423: 		} catch (HgRuntimeException ex) {
tikhomirov@490: 			repo.getSessionContext().getLog().dump(getClass(), Error, ex, null);
tikhomirov@354: 			return "";
tikhomirov@354: 		}
tikhomirov@114: 	}
tikhomirov@114: 	
tikhomirov@114: 	private String date() {
tikhomirov@354: 		Date d;
tikhomirov@354: 		try {
tikhomirov@354: 			d = getChangeset().date();
tikhomirov@423: 		} catch (HgRuntimeException ex) {
tikhomirov@490: 			repo.getSessionContext().getLog().dump(getClass(), Error, ex, null);
tikhomirov@354: 			d = new Date(0l);
tikhomirov@354: 		}
tikhomirov@354: 		return String.format("%tY/%<tm/%<td %<tH:%<tM:%<tS", d);
tikhomirov@114: 	}
tikhomirov@114: 	
tikhomirov@423: 	private RawChangeset getChangeset() throws HgRuntimeException {
tikhomirov@114: 		if (latestFileCset == null) {
tikhomirov@418: 			// TODO post-1.0 Use of TIP is likely incorrect in cases when working copy is not based
tikhomirov@418: 			// on latest revision. Perhaps, a constant like HgRepository.DIRSTATE_PARENT may come handy
tikhomirov@418: 			// Besides, it's reasonable to pass own inspector instead of implicit use of RawCsetCollector
tikhomirov@418: 			// to get changeset nodeid/index right away. Also check ChangelogHelper if may be of any use
tikhomirov@367: 			int csetRev = repo.getFileNode(path).getChangesetRevisionIndex(HgRepository.TIP);
tikhomirov@114: 			latestFileCset = repo.getChangelog().range(csetRev, csetRev).get(0);
tikhomirov@114: 		}
tikhomirov@114: 		return latestFileCset;
tikhomirov@112: 	}
tikhomirov@112: 
tikhomirov@112: 	public static class Factory implements Filter.Factory {
tikhomirov@711: 		private final Map<String,String> keywords;
tikhomirov@114: 		private HgRepository repo;
tikhomirov@114: 		private Path.Matcher matcher;
tikhomirov@711: 		
tikhomirov@711: 		public Factory() {
tikhomirov@711: 			keywords = new TreeMap<String,String>();
tikhomirov@711: 			keywords.put("Id", "Id");
tikhomirov@711: 			keywords.put("Revision", "Revision");
tikhomirov@711: 			keywords.put("Author", "Author");
tikhomirov@711: 			keywords.put("Date", "Date");
tikhomirov@711: 			keywords.put("LastChangedRevision", "LastChangedRevision");
tikhomirov@711: 			keywords.put("LastChangedBy", "LastChangedBy");
tikhomirov@711: 			keywords.put("LastChangedDate", "LastChangedDate");
tikhomirov@711: 			keywords.put("Source", "Source");
tikhomirov@711: 			keywords.put("Header", "Header");
tikhomirov@711: 		}
tikhomirov@112: 
tikhomirov@331: 		public void initialize(HgRepository hgRepo) {
tikhomirov@114: 			repo = hgRepo;
tikhomirov@114: 			ArrayList<String> patterns = new ArrayList<String>();
tikhomirov@331: 			for (Pair<String,String> e : hgRepo.getConfiguration().getSection("keyword")) {
tikhomirov@331: 				if (!"ignore".equalsIgnoreCase(e.second())) {
tikhomirov@331: 					patterns.add(e.first());
tikhomirov@114: 				}
tikhomirov@114: 			}
tikhomirov@114: 			matcher = new PathGlobMatcher(patterns.toArray(new String[patterns.size()]));
tikhomirov@423: 			// TODO post-1.0 read and respect keyword patterns from [keywordmaps]
tikhomirov@114: 		}
tikhomirov@114: 
tikhomirov@114: 		public Filter create(Path path, Options opts) {
tikhomirov@114: 			if (matcher.accept(path)) {
tikhomirov@711: 				return new KeywordFilter(repo, path, keywords, opts.getDirection() == Filter.Direction.FromRepo);
tikhomirov@114: 			}
tikhomirov@114: 			return null;
tikhomirov@112: 		}
tikhomirov@112: 	}
tikhomirov@112: 
tikhomirov@114: //
tikhomirov@114: //	public static void main(String[] args) throws Exception {
tikhomirov@114: //		FileInputStream fis = new FileInputStream(new File("/temp/kwoutput.txt"));
tikhomirov@114: //		FileOutputStream fos = new FileOutputStream(new File("/temp/kwoutput2.txt"));
tikhomirov@114: //		ByteBuffer b = ByteBuffer.allocate(256);
tikhomirov@114: //		KeywordFilter kwFilter = new KeywordFilter(false);
tikhomirov@114: //		while (fis.getChannel().read(b) != -1) {
tikhomirov@114: //			b.flip(); // get ready to be read
tikhomirov@114: //			ByteBuffer f = kwFilter.filter(b);
tikhomirov@114: //			fos.getChannel().write(f); // XXX in fact, f may not be fully consumed
tikhomirov@114: //			if (b.hasRemaining()) {
tikhomirov@114: //				b.compact();
tikhomirov@114: //			} else {
tikhomirov@114: //				b.clear();
tikhomirov@114: //			}
tikhomirov@114: //		}
tikhomirov@114: //		fis.close();
tikhomirov@114: //		fos.flush();
tikhomirov@114: //		fos.close();
tikhomirov@114: //	}
tikhomirov@112: }