tikhomirov@15: /* tikhomirov@74: * Copyright (c) 2010-2011 TMate Software Ltd tikhomirov@74: * tikhomirov@74: * This program is free software; you can redistribute it and/or modify tikhomirov@74: * it under the terms of the GNU General Public License as published by tikhomirov@74: * the Free Software Foundation; version 2 of the License. tikhomirov@74: * tikhomirov@74: * This program is distributed in the hope that it will be useful, tikhomirov@74: * but WITHOUT ANY WARRANTY; without even the implied warranty of tikhomirov@74: * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the tikhomirov@74: * GNU General Public License for more details. tikhomirov@74: * tikhomirov@74: * For information on how to redistribute this software under tikhomirov@74: * the terms of a license other than GNU General Public License tikhomirov@102: * contact TMate Software at support@hg4j.com tikhomirov@15: */ tikhomirov@74: package org.tmatesoft.hg.repo; tikhomirov@15: tikhomirov@15: import java.io.BufferedReader; tikhomirov@15: import java.io.File; tikhomirov@15: import java.io.FileReader; tikhomirov@15: import java.io.IOException; tikhomirov@91: import java.util.ArrayList; tikhomirov@15: import java.util.Collections; tikhomirov@91: import java.util.List; tikhomirov@91: import java.util.regex.Pattern; tikhomirov@15: tikhomirov@15: /** tikhomirov@91: * tikhomirov@74: * @author Artem Tikhomirov tikhomirov@74: * @author TMate Software Ltd. tikhomirov@15: */ tikhomirov@15: public class HgIgnore { tikhomirov@15: tikhomirov@91: private List entries; tikhomirov@15: tikhomirov@91: HgIgnore() { tikhomirov@91: entries = Collections.emptyList(); tikhomirov@15: } tikhomirov@15: tikhomirov@91: /* package-local */void read(File hgignoreFile) throws IOException { tikhomirov@15: if (!hgignoreFile.exists()) { tikhomirov@15: return; tikhomirov@15: } tikhomirov@91: ArrayList result = new ArrayList(entries); // start with existing tikhomirov@91: String syntax = "regex"; // or "glob" tikhomirov@91: BufferedReader fr = new BufferedReader(new FileReader(hgignoreFile)); tikhomirov@91: String line; tikhomirov@91: while ((line = fr.readLine()) != null) { tikhomirov@91: line = line.trim(); tikhomirov@91: if (line.startsWith("syntax:")) { tikhomirov@91: syntax = line.substring("syntax:".length()).trim(); tikhomirov@91: if (!"regex".equals(syntax) && !"glob".equals(syntax)) { tikhomirov@91: throw new IllegalStateException(line); tikhomirov@91: } tikhomirov@91: } else if (line.length() > 0) { tikhomirov@91: // shall I account for local paths in the file (i.e. tikhomirov@91: // back-slashed on windows)? tikhomirov@91: int x; tikhomirov@91: if ((x = line.indexOf('#')) >= 0) { tikhomirov@91: line = line.substring(0, x).trim(); tikhomirov@91: if (line.length() == 0) { tikhomirov@91: continue; tikhomirov@91: } tikhomirov@91: } tikhomirov@91: if ("glob".equals(syntax)) { tikhomirov@91: // hgignore(5) tikhomirov@91: // (http://www.selenic.com/mercurial/hgignore.5.html) says slashes '\' are escape characters, tikhomirov@91: // hence no special treatment of Windows path tikhomirov@91: // however, own attempts make me think '\' on Windows are not treated as escapes tikhomirov@91: line = glob2regex(line); tikhomirov@91: } tikhomirov@91: result.add(Pattern.compile(line)); // case-sensitive tikhomirov@15: } tikhomirov@15: } tikhomirov@91: result.trimToSize(); tikhomirov@91: entries = result; tikhomirov@15: } tikhomirov@15: tikhomirov@91: // note, #isIgnored(), even if queried for directories and returned positive reply, may still get tikhomirov@91: // a file from that ignored folder to get examined. Thus, patterns like "bin" shall match not only a folder, tikhomirov@91: // but any file under that folder as well tikhomirov@91: // Alternatively, file walker may memorize folder is ignored and uses this information for all nested files. However, tikhomirov@91: // this approach would require walker (a) return directories (b) provide nesting information. This may become tikhomirov@91: // troublesome when one walks not over io.File, but Eclipse's IResource or any other custom VFS. tikhomirov@91: // tikhomirov@91: // tikhomirov@91: // might be interesting, although looks like of no direct use in my case tikhomirov@91: // @see http://stackoverflow.com/questions/1247772/is-there-an-equivalent-of-java-util-regex-for-glob-type-patterns tikhomirov@91: private String glob2regex(String line) { tikhomirov@91: assert line.length() > 0; tikhomirov@91: StringBuilder sb = new StringBuilder(line.length() + 10); tikhomirov@91: sb.append('^'); // help avoid matcher.find() to match 'bin' pattern in the middle of the filename tikhomirov@91: int start = 0, end = line.length() - 1; tikhomirov@91: // '*' at the beginning and end of a line are useless for Pattern tikhomirov@114: // XXX although how about **.txt - such globs can be seen in a config, are they valid for HgIgnore? tikhomirov@91: while (start <= end && line.charAt(start) == '*') start++; tikhomirov@91: while (end > start && line.charAt(end) == '*') end--; tikhomirov@91: tikhomirov@91: for (int i = start; i <= end; i++) { tikhomirov@91: char ch = line.charAt(i); tikhomirov@91: if (ch == '.' || ch == '\\') { tikhomirov@91: sb.append('\\'); tikhomirov@91: } else if (ch == '?') { tikhomirov@91: // simple '.' substitution might work out, however, more formally tikhomirov@91: // a char class seems more appropriate to avoid accidentally tikhomirov@91: // matching a subdirectory with ? char (i.e. /a/b?d against /a/bad, /a/bed and /a/b/d) tikhomirov@91: // @see http://pubs.opengroup.org/onlinepubs/009695399/utilities/xcu_chap02.html#tag_02_13_03 tikhomirov@91: // quote: "The slash character in a pathname shall be explicitly matched by using one or more slashes in the pattern; tikhomirov@91: // it shall neither be matched by the asterisk or question-mark special characters nor by a bracket expression" tikhomirov@91: sb.append("[^/]"); tikhomirov@91: continue; tikhomirov@91: } else if (ch == '*') { tikhomirov@91: sb.append("[^/]*?"); tikhomirov@91: continue; tikhomirov@91: } tikhomirov@91: sb.append(ch); tikhomirov@91: } tikhomirov@91: return sb.toString(); tikhomirov@15: } tikhomirov@15: tikhomirov@114: // TODO use Path and PathGlobMatcher tikhomirov@15: public boolean isIgnored(String path) { tikhomirov@91: for (Pattern p : entries) { tikhomirov@91: if (p.matcher(path).find()) { tikhomirov@15: return true; tikhomirov@15: } tikhomirov@15: } tikhomirov@15: return false; tikhomirov@15: } tikhomirov@15: }