import java.util.*; import java.util.zip.*; import java.util.List; import java.util.regex.*; import java.util.concurrent.*; import java.util.concurrent.atomic.*; import javax.swing.*; import javax.swing.event.*; import javax.swing.text.*; import javax.swing.table.*; import java.io.*; import java.net.*; import java.lang.reflect.*; import java.lang.ref.*; import java.lang.management.*; import java.security.*; import java.security.spec.*; import java.awt.*; import java.awt.event.*; import java.awt.image.*; import javax.imageio.*; import java.math.*; public class main { static List urls = litlist( "http://tinybrain.de:8080/tb-int/get-transpiled.php?raw=1&withlibs=1&id=1000825&utf8=1", "http://tinybrain.de:8080/tb-int/get-transpiled.php?raw=1&withlibs=1&id=1000825&utf8=1"); public static void main(String[] args) throws Exception { /*loadPage_allowGzip =*/ loadPage_debug = true; print(l(loadPage("tinybrain.de:8080/tb/all-computers.php"))); List data = new ArrayList(); for (String url : urls) { String s; { long __time_startTime = now(); try { data.add(s = loadPage(url)); } finally { __time_startTime = now()-__time_startTime; saveTiming(__time_startTime); } } print("Chars loaded: " + l(s)); loadPage_allowGzip = false; } print(unidiff(data.get(0), data.get(1))); } static String unidiff(String a, String b) { int contextSize = 1; return fromLines(BlockDiffer.generateUniDiff(toLines(a), toLines(b), contextSize)); } static volatile StringBuffer local_log = new StringBuffer(); // not redirected static volatile StringBuffer print_log = local_log; // might be redirected, e.g. to main bot // in bytes - will cut to half that static volatile int print_log_max = 1024*1024; static volatile int local_log_max = 100*1024; static void print() { print(""); } // slightly overblown signature to return original object... static A print(A o) { String s = String.valueOf(o) + "\n"; StringBuffer loc = local_log; StringBuffer buf = print_log; int loc_max = print_log_max; if (buf != loc && buf != null) { print_append(buf, s, print_log_max); loc_max = local_log_max; } if (loc != null) print_append(loc, s, loc_max); System.out.print(s); return o; } static void print(long l) { print(String.valueOf(l)); } static void print(char c) { print(String.valueOf(c)); } static void print_append(StringBuffer buf, String s, int max) { synchronized(buf) { buf.append(s); max /= 2; if (buf.length() > max) try { int newLength = max/2; int ofs = buf.length()-newLength; String newString = buf.substring(ofs); buf.setLength(0); buf.append("[...] ").append(newString); } catch (Exception e) { buf.setLength(0); } } } static ArrayList litlist(A... a) { return new ArrayList(Arrays.asList(a)); } static ThreadLocal loadPage_charset = new ThreadLocal(); static boolean loadPage_allowGzip = true, loadPage_debug; public static String loadPageSilently(String url) { try { return loadPageSilently(new URL(loadPage_preprocess(url))); } catch (IOException e) { throw new RuntimeException(e); } } public static String loadPageSilently(URL url) { try { IOException e = null; for (int tries = 0; tries < 60; tries++) try { URLConnection con = url.openConnection(); return loadPage(con, url); } catch (IOException _e) { e = _e; print("Retrying because of: " + e); sleepSeconds(1); } throw e; } catch (IOException e) { throw new RuntimeException(e); } } static String loadPage_preprocess(String url) { if (url.startsWith("tb/")) url = "tinybrain.de:8080/" + url; if (url.indexOf("://") < 0) url = "http://" + url; return url; } public static String loadPage(String url) { try { return loadPage(new URL(loadPage_preprocess(url))); } catch (IOException e) { throw new RuntimeException(e); } } public static String loadPage(URL url) { print("Loading: " + url.toExternalForm()); return loadPageSilently(url); } public static String loadPage(URLConnection con, URL url) throws IOException { String computerID = getComputerID(); try { if (computerID != null) con.setRequestProperty("X-ComputerID", computerID); if (loadPage_allowGzip) con.setRequestProperty("Accept-Encoding", "gzip"); } catch (Throwable e) {} // fails if within doPost String contentType = con.getContentType(); if (contentType == null) throw new IOException("Page could not be read: " + url); //print("Content-Type: " + contentType); String charset = loadPage_charset == null ? null : loadPage_charset.get(); if (charset == null) charset = loadPage_guessCharset(contentType); InputStream in = con.getInputStream(); if ("gzip".equals(con.getContentEncoding())) { if (loadPage_debug) print("loadPage: Using gzip."); in = new GZIPInputStream(in); } Reader r = new InputStreamReader(in, charset); StringBuilder buf = new StringBuilder(); while (true) { int ch = r.read(); if (ch < 0) break; //Log.info("Chars read: " + buf.length()); buf.append((char) ch); } return buf.toString(); } static String loadPage_guessCharset(String contentType) { Pattern p = Pattern.compile("text/[a-z]+;\\s+charset=([^\\s]+)\\s*"); Matcher m = p.matcher(contentType); /* If Content-Type doesn't match this pre-conception, choose default and hope for the best. */ return m.matches() ? m.group(1) : "ISO-8859-1"; } static int l(Object[] array) { return array == null ? 0 : array.length; } static int l(byte[] array) { return array == null ? 0 : array.length; } static int l(int[] array) { return array == null ? 0 : array.length; } static int l(char[] array) { return array == null ? 0 : array.length; } static int l(Collection c) { return c == null ? 0 : c.size(); } static int l(Map m) { return m == null ? 0 : m.size(); } static int l(String s) { return s == null ? 0 : s.length(); } static void sleepSeconds(long s) { if (s > 0) sleep(s*1000); } static List toLines(File f) { return toLines(loadTextFile(f)); } public static List toLines(String s) { List lines = new ArrayList(); if (s == null) return lines; int start = 0; while (true) { int i = toLines_nextLineBreak(s, start); if (i < 0) { if (s.length() > start) lines.add(s.substring(start)); break; } lines.add(s.substring(start, i)); if (s.charAt(i) == '\r' && i+1 < s.length() && s.charAt(i+1) == '\n') i += 2; else ++i; start = i; } return lines; } private static int toLines_nextLineBreak(String s, int start) { for (int i = start; i < s.length(); i++) { char c = s.charAt(i); if (c == '\r' || c == '\n') return i; } return -1; } static String getComputerID() { try { return computerID(); } catch (Throwable __e) { throw __e instanceof RuntimeException ? (RuntimeException) __e : new RuntimeException(__e); }} public static String fromLines(List lines) { StringBuilder buf = new StringBuilder(); if (lines != null) for (String line : lines) buf.append(line).append('\n'); return buf.toString(); } public static String loadTextFile(String fileName) { try { return loadTextFile(fileName, null); } catch (IOException e) { throw new RuntimeException(e); } } public static String loadTextFile(String fileName, String defaultContents) throws IOException { if (!new File(fileName).exists()) return defaultContents; FileInputStream fileInputStream = new FileInputStream(fileName); InputStreamReader inputStreamReader = new InputStreamReader(fileInputStream, "UTF-8"); return loadTextFile(inputStreamReader); } public static String loadTextFile(File fileName) { try { return loadTextFile(fileName, null); } catch (IOException e) { throw new RuntimeException(e); } } public static String loadTextFile(File fileName, String defaultContents) throws IOException { try { return loadTextFile(fileName.getPath(), defaultContents); } catch (IOException e) { throw new RuntimeException(e); } } public static String loadTextFile(Reader reader) throws IOException { StringBuilder builder = new StringBuilder(); try { char[] buffer = new char[1024]; int n; while (-1 != (n = reader.read(buffer))) builder.append(buffer, 0, n); } finally { reader.close(); } return builder.toString(); } static void sleep(long ms) { try { Thread.sleep(ms); } catch (Exception e) { throw new RuntimeException(e); } } static void sleep() { try { print("Sleeping."); synchronized(main.class) { main.class.wait(); } } catch (Throwable __e) { throw __e instanceof RuntimeException ? (RuntimeException) __e : new RuntimeException(__e); }} static String _computerID; public static String computerID() { try { if (_computerID == null) { File file = new File(userHome(), ".tinybrain/computer-id"); _computerID = loadTextFile(file.getPath(), null); if (_computerID == null) { _computerID = makeRandomID(12); saveTextFile(file.getPath(), _computerID); } } return _computerID; } catch (Throwable __e) { throw __e instanceof RuntimeException ? (RuntimeException) __e : new RuntimeException(__e); }} static String makeRandomID(int length) { Random random = new Random(); char[] id = new char[length]; for (int i = 0; i < id.length; i++) id[i] = (char) ((int) 'a' + random.nextInt(26)); return new String(id); } static String _userHome; static String userHome() { if (_userHome == null) { if (isAndroid()) _userHome = "/storage/sdcard0/"; else _userHome = System.getProperty("user.home"); //System.out.println("userHome: " + _userHome); } return _userHome; } /** writes safely (to temp file, then rename) */ public static void saveTextFile(String fileName, String contents) throws IOException { File file = new File(fileName); File parentFile = file.getParentFile(); if (parentFile != null) parentFile.mkdirs(); String tempFileName = fileName + "_temp"; if (contents != null) { FileOutputStream fileOutputStream = new FileOutputStream(tempFileName); OutputStreamWriter outputStreamWriter = new OutputStreamWriter(fileOutputStream, "UTF-8"); PrintWriter printWriter = new PrintWriter(outputStreamWriter); printWriter.print(contents); printWriter.close(); } if (file.exists() && !file.delete()) throw new IOException("Can't delete " + fileName); if (contents != null) if (!new File(tempFileName).renameTo(file)) throw new IOException("Can't rename " + tempFileName + " to " + fileName); } public static void saveTextFile(File fileName, String contents) { try { saveTextFile(fileName.getPath(), contents); } catch (IOException e) { throw new RuntimeException(e); } } static boolean isAndroid() { return System.getProperty("java.vendor").toLowerCase().indexOf("android") >= 0; } static class BlockDiff { public CopyBlock asCopyBlock() { return null; } public NewBlock asNewBlock () { return null; } } static class CopyBlock extends BlockDiff { int firstLine, lines; CopyBlock(int firstLine, int lines) { this.firstLine = firstLine; this.lines = lines; } public CopyBlock asCopyBlock() { return this; } public int getFirstLine() { return firstLine; } public int getLines() { return lines; } } static class NewBlock extends BlockDiff { int originalStart; List contents; NewBlock(int originalStart, List contents) { this.originalStart = originalStart; this.contents = contents; } public NewBlock asNewBlock () { return this; } public int getOriginalStart() { return originalStart; } public List getContents() { return contents; } } static class ExplodedLine { int type; String left, right; int leftIndex, rightIndex; ExplodedLine(int type, String left, String right, int leftIndex, int rightIndex) { this.type = type; this.left = left; this.right = right; this.leftIndex = leftIndex; this.rightIndex = rightIndex; } public int getType() { return type; } public String getLeft() { return left; } public String getRight() { return right; } public int getLeftIndex() { return leftIndex; } public int getRightIndex() { return rightIndex; } } static class BlockDiffer { public static final int IDENTICAL = 0; public static final int DIFFERENT = 1; public static final int LEFT_ONLY = 2; public static final int RIGHT_ONLY = 3; private static void printChange(EGDiff.change change) { if (change != null) { System.out.println("line0="+change.line0+", line1="+change.line1 +", inserted="+change.inserted+", deleted="+change.deleted); printChange(change.link); } } /** Generates the text content of a Unified-format context diff between 2 files * (NB the 'files-changed' header must be added separately). */ public static List generateUniDiff(List fileA, List fileB, int contextSize) { EGDiff diff = new EGDiff(fileA.toArray(), fileB.toArray()); EGDiff.change change = diff.diff_2(false); if (change != null) { int inserted, deleted; List hunkLines = new ArrayList(); int cumulExtraLinesBwrtA = 0; // Each hunk is generated with a header do { int line0 = change.line0, line1 = change.line1; int changeStart = ((line1 < line0) ? line1 : line0); int contextStart = ((changeStart > contextSize) ? changeStart - contextSize : 0); int headerPosn = hunkLines.size(); // Provide the first lines of context for (int i = contextStart; i < changeStart; i++) //System.out.println(" " + fileA.get(i)); hunkLines.add(" " + fileA.get(i)); boolean hunkFinish = false; // Step through each change giving the change lines and following context do { inserted = change.inserted; deleted = change.deleted; line0 = change.line0; line1 = change.line1; if (line1 < line0) // An insert comes earlier while (inserted-- > 0) hunkLines.add("+" + fileB.get(line1++)); while (deleted-- > 0) hunkLines.add("-" + fileA.get(line0++)); while (inserted-- > 0) hunkLines.add("+" + fileB.get(line1++)); // Lines following are trailing context, identical in fileA and fileB // The next change may overlap the context, so check and if so, form one hunk EGDiff.change nextChange = change.link; int nextChangeStart = fileA.size(); if (nextChange != null) nextChangeStart = ((nextChange.line1 < nextChange.line0) ? nextChange.line1 : nextChange.line0); if (nextChangeStart - line0 > contextSize * 2) { // A separate hunk nextChangeStart = line0 + contextSize; hunkFinish = true; } if (nextChangeStart > fileA.size()) nextChangeStart = fileA.size(); // Limit to file size while (line0 < nextChangeStart) { hunkLines.add(" " + fileA.get(line0++)); line1++; // Keep in sync with trailing context } change = change.link; } while (!hunkFinish && change != null); int hunkStartB = contextStart + cumulExtraLinesBwrtA; int hunkTotA = line0 - contextStart; int hunkTotB = line1 - hunkStartB; hunkLines.add(headerPosn, "@@ -" + (contextStart + 1) + ',' + hunkTotA + " +" + (hunkStartB + 1) + ',' + hunkTotB + " @@"); cumulExtraLinesBwrtA += hunkTotB - hunkTotA; } while (change != null); return hunkLines; } return null; } /* For testing: private static void printUniDiff(List fileA, List fileB, int contextSize) { List uniDiff = generateUniDiff(fileA, fileB, contextSize); if (uniDiff != null) for (int j = 0; j < uniDiff.size(); j++) System.out.println(uniDiff.get(j)); } */ public static List diffLines(List lines, List reference) { List diffs = new ArrayList(); EGDiff diff = new EGDiff(reference.toArray(), lines.toArray()); EGDiff.change change = diff.diff_2(false); //printChange(change); //printUniDiff(reference, lines, 3); int l0 = 0, l1 = 0; while (change != null) { if (change.line0 > l0 && change.line1 > l1) diffs.add(new CopyBlock(l0, change.line0-l0)); if (change.inserted != 0) diffs.add(new NewBlock(change.line1, lines.subList(change.line1, change.line1+change.inserted))); l0 = change.line0 + change.deleted; l1 = change.line1 + change.inserted; change = change.link; } if (l0 < reference.size()) diffs.add(new CopyBlock(l0, reference.size()-l0)); return diffs; } /** fills files with empty lines to align matching blocks * * @param file1 first file * @param file2 second file * @return an array with two lists */ public static List explode(List file1, List file2) { List lines = new ArrayList(); List diffs = BlockDiffer.diffLines(file2, file1); int lastLineCopied = 0, rightOnlyStart = -1, rightPosition = 0; for (int i = 0; i < diffs.size(); i++) { BlockDiff diff = diffs.get(i); if (diff instanceof CopyBlock) { CopyBlock copyBlock = (CopyBlock) diff; if (lastLineCopied < copyBlock.getFirstLine()) { if (rightOnlyStart >= 0) { int overlap = Math.min(lines.size()-rightOnlyStart, copyBlock.getFirstLine()-lastLineCopied); //lines.subList(rightOnlyStart, rightOnlyStart+overlap).clear(); convertRightOnlyToDifferent(lines, rightOnlyStart, overlap, file1, lastLineCopied); lastLineCopied += overlap; } addBlock(lines, LEFT_ONLY, file1, lastLineCopied, copyBlock.getFirstLine(), lastLineCopied, -1); } addBlock(lines, IDENTICAL, file1, copyBlock.getFirstLine(), copyBlock.getFirstLine()+copyBlock.getLines(), copyBlock.getFirstLine(), rightPosition); rightPosition += copyBlock.getLines(); lastLineCopied = copyBlock.getFirstLine()+copyBlock.getLines(); rightOnlyStart = -1; } else if (diff instanceof NewBlock) { NewBlock newBlock = (NewBlock) diff; /*if (nextDiff instanceof BlockDiffer.CopyBlock) { BlockDiffer.CopyBlock copyBlock = (BlockDiffer.CopyBlock) nextDiff; copyBlock.getFirstLine()-lastLineCopied*/ rightOnlyStart = lines.size(); addBlock(lines, RIGHT_ONLY, newBlock.getContents(), 0, newBlock.getContents().size(), -1, rightPosition); rightPosition += newBlock.getContents().size(); } } if (rightOnlyStart >= 0) { int overlap = Math.min(lines.size()-rightOnlyStart, file1.size()-lastLineCopied); //lines.subList(rightOnlyStart, rightOnlyStart+overlap).clear(); convertRightOnlyToDifferent(lines, rightOnlyStart, overlap, file1, lastLineCopied); lastLineCopied += overlap; } addBlock(lines, LEFT_ONLY, file1, lastLineCopied, file1.size(), lastLineCopied, -1); return lines; } private static void convertRightOnlyToDifferent(List lines, int start, int numLines, List leftLines, int leftStart) { for (int i = 0; i < numLines; i++) { ExplodedLine line = lines.get(start+i); lines.set(start+i, new ExplodedLine(DIFFERENT, leftLines.get(i+leftStart), line.getRight(), i+leftStart, line.getRightIndex())); } } private static void addBlock(List lines, int type, List srcLines, int start, int end, int leftStart, int rightStart) { for (int i = start; i < end; i++) lines.add(new ExplodedLine(type, type == RIGHT_ONLY ? "" : srcLines.get(i), type == LEFT_ONLY ? "" : srcLines.get(i), type == RIGHT_ONLY ? -1 : i - start + leftStart, type == LEFT_ONLY ? -1 : i - start + rightStart)); } public static List condense(List lines) { List result = new ArrayList(); for (Iterator i = lines.iterator(); i.hasNext();) { ExplodedLine line = i.next(); if (line.getType() == IDENTICAL) { if (result.isEmpty() || result.get(result.size()-1).getType() != IDENTICAL) result.add(new ExplodedLine(IDENTICAL, "[...]", "[...]", -1, -1)); } else result.add(line); } return result; } } static long now_virtualTime; static long now() { return now_virtualTime != 0 ? now_virtualTime : System.currentTimeMillis(); } static ThreadLocal saveTiming_last = new ThreadLocal(); static void saveTiming(long ms) { print(ms + " ms"); saveTiming_last.set(ms); } /** * A class to compare vectors of objects. The result of comparison * is a list of change objects which form an * edit script. The objects compared are traditionally lines * of text from two files. Comparison options such as "ignore * whitespace" are implemented by modifying the equals * and hashcode methods for the objects compared. *

* The basic algorithm is described in:
* "An O(ND) Difference Algorithm and its Variations", Eugene Myers, * Algorithmica Vol. 1 No. 2, 1986, p 251. *

* This class outputs different results from GNU diff 1.15 on some * inputs. Our results are actually better (smaller change list, smaller * total size of changes), but it would be nice to know why. Perhaps * there is a memory overwrite bug in GNU diff 1.15. * * @author Stuart D. Gathman, translated from GNU diff 1.15 * Copyright (C) 2000 Business Management Systems, Inc. *

* 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 1, 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. */ static class EGDiff { /** * Prepare to find differences between two arrays. Each element of * the arrays is translated to an "equivalence number" based on * the result of equals. The original Object arrays * are no longer needed for computing the differences. They will * be needed again later to print the results of the comparison as * an edit script, if desired. */ public EGDiff(Object[] a, Object[] b) { Hashtable h = new Hashtable(a.length + b.length); filevec[0] = new file_data(a, h); filevec[1] = new file_data(b, h); } /** * 1 more than the maximum equivalence value used for this or its * sibling file. */ private int equiv_max = 1; /** * When set to true, the comparison uses a heuristic to speed it up. * With this heuristic, for files with a constant small density * of changes, the algorithm is linear in the file size. */ public boolean heuristic = false; /** * When set to true, the algorithm returns a guarranteed minimal * set of changes. This makes things slower, sometimes much slower. */ public boolean no_discards = false; private int[] xvec, yvec; /* Vectors being compared. */ private int[] fdiag; /* Vector, indexed by diagonal, containing the X coordinate of the point furthest along the given diagonal in the forward search of the edit matrix. */ private int[] bdiag; /* Vector, indexed by diagonal, containing the X coordinate of the point furthest along the given diagonal in the backward search of the edit matrix. */ private int fdiagoff, bdiagoff; private final file_data[] filevec = new file_data[2]; private int cost; /** * Find the midpoint of the shortest edit script for a specified * portion of the two files. *

* We scan from the beginnings of the files, and simultaneously from the ends, * doing a breadth-first search through the space of edit-sequence. * When the two searches meet, we have found the midpoint of the shortest * edit sequence. *

* The value returned is the number of the diagonal on which the midpoint lies. * The diagonal number equals the number of inserted lines minus the number * of deleted lines (counting only lines before the midpoint). * The edit cost is stored into COST; this is the total number of * lines inserted or deleted (counting only lines before the midpoint). *

* This function assumes that the first lines of the specified portions * of the two files do not match, and likewise that the last lines do not * match. The caller must trim matching lines from the beginning and end * of the portions it is going to specify. *

* Note that if we return the "wrong" diagonal value, or if * the value of bdiag at that diagonal is "wrong", * the worst this can do is cause suboptimal diff output. * It cannot cause incorrect diff output. */ private int diag(int xoff, int xlim, int yoff, int ylim) { final int[] fd = fdiag; // Give the compiler a chance. final int[] bd = bdiag; // Additional help for the compiler. final int[] xv = xvec; // Still more help for the compiler. final int[] yv = yvec; // And more and more . . . final int dmin = xoff - ylim; // Minimum valid diagonal. final int dmax = xlim - yoff; // Maximum valid diagonal. final int fmid = xoff - yoff; // Center diagonal of top-down search. final int bmid = xlim - ylim; // Center diagonal of bottom-up search. int fmin = fmid, fmax = fmid; // Limits of top-down search. int bmin = bmid, bmax = bmid; // Limits of bottom-up search. /* True if southeast corner is on an odd diagonal with respect to the northwest. */ final boolean odd = (fmid - bmid & 1) != 0; fd[fdiagoff + fmid] = xoff; bd[bdiagoff + bmid] = xlim; for (int c = 1; ; ++c) { int d; /* Active diagonal. */ boolean big_snake = false; /* Extend the top-down search by an edit step in each diagonal. */ if (fmin > dmin) fd[fdiagoff + --fmin - 1] = -1; else ++fmin; if (fmax < dmax) fd[fdiagoff + ++fmax + 1] = -1; else --fmax; for (d = fmax; d >= fmin; d -= 2) { int x, y, oldx, tlo = fd[fdiagoff + d - 1], thi = fd[fdiagoff + d + 1]; if (tlo >= thi) x = tlo + 1; else x = thi; oldx = x; y = x - d; while (x < xlim && y < ylim && xv[x] == yv[y]) { ++x; ++y; } if (x - oldx > 20) big_snake = true; fd[fdiagoff + d] = x; if (odd && bmin <= d && d <= bmax && bd[bdiagoff + d] <= fd[fdiagoff + d]) { cost = 2 * c - 1; return d; } } /* Similar extend the bottom-up search. */ if (bmin > dmin) bd[bdiagoff + --bmin - 1] = Integer.MAX_VALUE; else ++bmin; if (bmax < dmax) bd[bdiagoff + ++bmax + 1] = Integer.MAX_VALUE; else --bmax; for (d = bmax; d >= bmin; d -= 2) { int x, y, oldx, tlo = bd[bdiagoff + d - 1], thi = bd[bdiagoff + d + 1]; if (tlo < thi) x = tlo; else x = thi - 1; oldx = x; y = x - d; while (x > xoff && y > yoff && xv[x - 1] == yv[y - 1]) { --x; --y; } if (oldx - x > 20) big_snake = true; bd[bdiagoff + d] = x; if (!odd && fmin <= d && d <= fmax && bd[bdiagoff + d] <= fd[fdiagoff + d]) { cost = 2 * c; return d; } } /* Heuristic: check occasionally for a diagonal that has made lots of progress compared with the edit distance. If we have any such, find the one that has made the most progress and return it as if it had succeeded. With this heuristic, for files with a constant small density of changes, the algorithm is linear in the file size. */ if (c > 200 && big_snake && heuristic) { int best = 0; int bestpos = -1; for (d = fmax; d >= fmin; d -= 2) { int dd = d - fmid; if ((fd[fdiagoff + d] - xoff) * 2 - dd > 12 * (c + (dd > 0 ? dd : -dd))) { if (fd[fdiagoff + d] * 2 - dd > best && fd[fdiagoff + d] - xoff > 20 && fd[fdiagoff + d] - d - yoff > 20) { int k; int x = fd[fdiagoff + d]; /* We have a good enough best diagonal; now insist that it end with a significant snake. */ for (k = 1; k <= 20; k++) if (xvec[x - k] != yvec[x - d - k]) break; if (k == 21) { best = fd[fdiagoff + d] * 2 - dd; bestpos = d; } } } } if (best > 0) { cost = 2 * c - 1; return bestpos; } best = 0; for (d = bmax; d >= bmin; d -= 2) { int dd = d - bmid; if ((xlim - bd[bdiagoff + d]) * 2 + dd > 12 * (c + (dd > 0 ? dd : -dd))) { if ((xlim - bd[bdiagoff + d]) * 2 + dd > best && xlim - bd[bdiagoff + d] > 20 && ylim - (bd[bdiagoff + d] - d) > 20) { /* We have a good enough best diagonal; now insist that it end with a significant snake. */ int k; int x = bd[bdiagoff + d]; for (k = 0; k < 20; k++) if (xvec[x + k] != yvec[x - d + k]) break; if (k == 20) { best = (xlim - bd[bdiagoff + d]) * 2 + dd; bestpos = d; } } } } if (best > 0) { cost = 2 * c - 1; return bestpos; } } } } /** * Compare in detail contiguous subsequences of the two files * which are known, as a whole, to match each other. *

* The results are recorded in the vectors filevec[N].changed_flag, by * storing a 1 in the element for each line that is an insertion or deletion. *

* The subsequence of file 0 is [XOFF, XLIM) and likewise for file 1. *

* Note that XLIM, YLIM are exclusive bounds. * All line numbers are origin-0 and discarded lines are not counted. */ private void compareseq(int xoff, int xlim, int yoff, int ylim) { /* Slide down the bottom initial diagonal. */ while (xoff < xlim && yoff < ylim && xvec[xoff] == yvec[yoff]) { ++xoff; ++yoff; } /* Slide up the top initial diagonal. */ while (xlim > xoff && ylim > yoff && xvec[xlim - 1] == yvec[ylim - 1]) { --xlim; --ylim; } /* Handle simple cases. */ if (xoff == xlim) while (yoff < ylim) filevec[1].changed_flag[1 + filevec[1].realindexes[yoff++]] = true; else if (yoff == ylim) while (xoff < xlim) filevec[0].changed_flag[1 + filevec[0].realindexes[xoff++]] = true; else { /* Find a point of correspondence in the middle of the files. */ int d = diag(xoff, xlim, yoff, ylim); int c = cost; int b = bdiag[bdiagoff + d]; if (c == 1) { /* This should be impossible, because it implies that one of the two subsequences is empty, and that case was handled above without calling `diag'. Let's verify that this is true. */ throw new IllegalArgumentException("Empty subsequence"); } else { /* Use that point to split this problem into two subproblems. */ compareseq(xoff, b, yoff, b - d); /* This used to use f instead of b, but that is incorrect! It is not necessarily the case that diagonal d has a snake from b to f. */ compareseq(b, xlim, b - d, ylim); } } } /** * Discard lines from one file that have no matches in the other file. */ private void discard_confusing_lines() { filevec[0].discard_confusing_lines(filevec[1]); filevec[1].discard_confusing_lines(filevec[0]); } private boolean inhibit = false; /** * Adjust inserts/deletes of blank lines to join changes * as much as possible. */ private void shift_boundaries() { if (inhibit) return; filevec[0].shift_boundaries(filevec[1]); filevec[1].shift_boundaries(filevec[0]); } public interface ScriptBuilder { /** * Scan the tables of which lines are inserted and deleted, * producing an edit script. * * @param changed0 true for lines in first file which do not match 2nd * @param len0 number of lines in first file * @param changed1 true for lines in 2nd file which do not match 1st * @param len1 number of lines in 2nd file * @return a linked list of changes - or null */ public change build_script(boolean[] changed0, int len0, boolean[] changed1, int len1); } /** * Scan the tables of which lines are inserted and deleted, * producing an edit script in reverse order. */ static class ReverseScript implements ScriptBuilder { public change build_script(final boolean[] changed0, int len0, final boolean[] changed1, int len1) { change script = null; int i0 = 0, i1 = 0; while (i0 < len0 || i1 < len1) { if (changed0[1 + i0] || changed1[1 + i1]) { int line0 = i0, line1 = i1; /* Find # lines changed here in each file. */ while (changed0[1 + i0]) ++i0; while (changed1[1 + i1]) ++i1; /* Record this change. */ script = new change(line0, line1, i0 - line0, i1 - line1, script); } /* We have reached lines in the two files that match each other. */ i0++; i1++; } return script; } } static class ForwardScript implements ScriptBuilder { /** * Scan the tables of which lines are inserted and deleted, * producing an edit script in forward order. */ public change build_script(final boolean[] changed0, int len0, final boolean[] changed1, int len1) { change script = null; int i0 = len0, i1 = len1; while (i0 >= 0 || i1 >= 0) { if (changed0[i0] || changed1[i1]) { int line0 = i0, line1 = i1; /* Find # lines changed here in each file. */ while (changed0[i0]) --i0; while (changed1[i1]) --i1; /* Record this change. */ script = new change(i0, i1, line0 - i0, line1 - i1, script); } /* We have reached lines in the two files that match each other. */ i0--; i1--; } return script; } } /** * Standard ScriptBuilders. */ public final static ScriptBuilder forwardScript = new ForwardScript(), reverseScript = new ReverseScript(); /* Report the differences of two files. DEPTH is the current directory depth. */ public final change diff_2(final boolean reverse) { return diff(reverse ? reverseScript : forwardScript); } /** * Get the results of comparison as an edit script. The script * is described by a list of changes. The standard ScriptBuilder * implementations provide for forward and reverse edit scripts. * Alternate implementations could, for instance, list common elements * instead of differences. * * @param bld an object to build the script from change flags * @return the head of a list of changes */ public change diff(final ScriptBuilder bld) { /* Some lines are obviously insertions or deletions because they don't match anything. Detect them now, and avoid even thinking about them in the main comparison algorithm. */ discard_confusing_lines(); /* Now do the main comparison algorithm, considering just the undiscarded lines. */ xvec = filevec[0].undiscarded; yvec = filevec[1].undiscarded; int diags = filevec[0].nondiscarded_lines + filevec[1].nondiscarded_lines + 3; fdiag = new int[diags]; fdiagoff = filevec[1].nondiscarded_lines + 1; bdiag = new int[diags]; bdiagoff = filevec[1].nondiscarded_lines + 1; compareseq(0, filevec[0].nondiscarded_lines, 0, filevec[1].nondiscarded_lines); fdiag = null; bdiag = null; /* Modify the results slightly to make them prettier in cases where that can validly be done. */ shift_boundaries(); /* Get the results of comparison in the form of a chain of `struct change's -- an edit script. */ return bld.build_script(filevec[0].changed_flag, filevec[0].buffered_lines, filevec[1].changed_flag, filevec[1].buffered_lines); } /** * The result of comparison is an "edit script": a chain of change objects. * Each change represents one place where some lines are deleted * and some are inserted. *

* LINE0 and LINE1 are the first affected lines in the two files (origin 0). * DELETED is the number of lines deleted here from file 0. * INSERTED is the number of lines inserted here in file 1. *

* If DELETED is 0 then LINE0 is the number of the line before * which the insertion was done; vice versa for INSERTED and LINE1. */ public static class change { /** * Previous or next edit command. */ public change link; /** * # lines of file 1 changed here. */ public final int inserted; /** * # lines of file 0 changed here. */ public final int deleted; /** * Line number of 1st deleted line. */ public final int line0; /** * Line number of 1st inserted line. */ public final int line1; /** * Cons an additional entry onto the front of an edit script OLD. * LINE0 and LINE1 are the first affected lines in the two files (origin 0). * DELETED is the number of lines deleted here from file 0. * INSERTED is the number of lines inserted here in file 1. *

* If DELETED is 0 then LINE0 is the number of the line before * which the insertion was done; vice versa for INSERTED and LINE1. */ public change(int line0, int line1, int deleted, int inserted, change old) { this.line0 = line0; this.line1 = line1; this.inserted = inserted; this.deleted = deleted; this.link = old; //System.err.println(line0+","+line1+","+inserted+","+deleted); } } /** * Data on one input file being compared. */ class file_data { /** * Allocate changed array for the results of comparison. */ void clear() { /* Allocate a flag for each line of each file, saying whether that line is an insertion or deletion. Allocate an extra element, always zero, at each end of each vector. */ changed_flag = new boolean[buffered_lines + 2]; } /** * Return equiv_count[I] as the number of lines in this file * that fall in equivalence class I. * * @return the array of equivalence class counts. */ int[] equivCount() { int[] equiv_count = new int[equiv_max]; for (int i = 0; i < buffered_lines; ++i) ++equiv_count[equivs[i]]; return equiv_count; } /** * Discard lines that have no matches in another file. *

* A line which is discarded will not be considered by the actual * comparison algorithm; it will be as if that line were not in the file. * The file's `realindexes' table maps virtual line numbers * (which don't count the discarded lines) into real line numbers; * this is how the actual comparison algorithm produces results * that are comprehensible when the discarded lines are counted. *

* When we discard a line, we also mark it as a deletion or insertion * so that it will be printed in the output. * * @param f the other file */ void discard_confusing_lines(file_data f) { clear(); /* Set up table of which lines are going to be discarded. */ final byte[] discarded = discardable(f.equivCount()); /* Don't really discard the provisional lines except when they occur in a run of discardables, with nonprovisionals at the beginning and end. */ filterDiscards(discarded); /* Actually discard the lines. */ discard(discarded); } /** * Mark to be discarded each line that matches no line of another file. * If a line matches many lines, mark it as provisionally discardable. * * @param counts The count of each equivalence number for the other file. * @return 0=nondiscardable, 1=discardable or 2=provisionally discardable * for each line */ private byte[] discardable(final int[] counts) { final int end = buffered_lines; final byte[] discards = new byte[end]; final int[] equivs = this.equivs; int many = 5; int tem = end / 64; /* Multiply MANY by approximate square root of number of lines. That is the threshold for provisionally discardable lines. */ while ((tem = tem >> 2) > 0) many *= 2; for (int i = 0; i < end; i++) { int nmatch; if (equivs[i] == 0) continue; nmatch = counts[equivs[i]]; if (nmatch == 0) discards[i] = 1; else if (nmatch > many) discards[i] = 2; } return discards; } /** * Don't really discard the provisional lines except when they occur * in a run of discardables, with nonprovisionals at the beginning * and end. */ private void filterDiscards(final byte[] discards) { final int end = buffered_lines; for (int i = 0; i < end; i++) { /* Cancel provisional discards not in middle of run of discards. */ if (discards[i] == 2) discards[i] = 0; else if (discards[i] != 0) { /* We have found a nonprovisional discard. */ int j; int length; int provisional = 0; /* Find end of this run of discardable lines. Count how many are provisionally discardable. */ for (j = i; j < end; j++) { if (discards[j] == 0) break; if (discards[j] == 2) ++provisional; } /* Cancel provisional discards at end, and shrink the run. */ while (j > i && discards[j - 1] == 2) { discards[--j] = 0; --provisional; } /* Now we have the length of a run of discardable lines whose first and last are not provisional. */ length = j - i; /* If 1/4 of the lines in the run are provisional, cancel discarding of all provisional lines in the run. */ if (provisional * 4 > length) { while (j > i) if (discards[--j] == 2) discards[j] = 0; } else { int consec; int minimum = 1; int tem = length / 4; /* MINIMUM is approximate square root of LENGTH/4. A subrun of two or more provisionals can stand when LENGTH is at least 16. A subrun of 4 or more can stand when LENGTH >= 64. */ while ((tem = tem >> 2) > 0) minimum *= 2; minimum++; /* Cancel any subrun of MINIMUM or more provisionals within the larger run. */ for (j = 0, consec = 0; j < length; j++) if (discards[i + j] != 2) consec = 0; else if (minimum == ++consec) /* Back up to start of subrun, to cancel it all. */ j -= consec; else if (minimum < consec) discards[i + j] = 0; /* Scan from beginning of run until we find 3 or more nonprovisionals in a row or until the first nonprovisional at least 8 lines in. Until that point, cancel any provisionals. */ for (j = 0, consec = 0; j < length; j++) { if (j >= 8 && discards[i + j] == 1) break; if (discards[i + j] == 2) { consec = 0; discards[i + j] = 0; } else if (discards[i + j] == 0) consec = 0; else consec++; if (consec == 3) break; } /* I advances to the last line of the run. */ i += length - 1; /* Same thing, from end. */ for (j = 0, consec = 0; j < length; j++) { if (j >= 8 && discards[i - j] == 1) break; if (discards[i - j] == 2) { consec = 0; discards[i - j] = 0; } else if (discards[i - j] == 0) consec = 0; else consec++; if (consec == 3) break; } } } } } /** * Actually discard the lines. * * @param discards flags lines to be discarded */ private void discard(final byte[] discards) { final int end = buffered_lines; int j = 0; for (int i = 0; i < end; ++i) if (no_discards || discards[i] == 0) { undiscarded[j] = equivs[i]; realindexes[j++] = i; } else changed_flag[1 + i] = true; nondiscarded_lines = j; } file_data(Object[] data, Hashtable h) { buffered_lines = data.length; equivs = new int[buffered_lines]; undiscarded = new int[buffered_lines]; realindexes = new int[buffered_lines]; for (int i = 0; i < data.length; ++i) { Integer ir = (Integer) h.get(data[i]); if (ir == null) h.put(data[i], new Integer(equivs[i] = equiv_max++)); else equivs[i] = ir.intValue(); } } /** * Adjust inserts/deletes of blank lines to join changes * as much as possible. *

* We do something when a run of changed lines include a blank * line at one end and have an excluded blank line at the other. * We are free to choose which blank line is included. * `compareseq' always chooses the one at the beginning, * but usually it is cleaner to consider the following blank line * to be the "change". The only exception is if the preceding blank line * would join this change to other changes. * * @param f the file being compared against */ void shift_boundaries(file_data f) { final boolean[] changed = changed_flag; final boolean[] other_changed = f.changed_flag; int i = 0; int j = 0; int i_end = buffered_lines; int preceding = -1; int other_preceding = -1; for (; ;) { int start, end, other_start; /* Scan forwards to find beginning of another run of changes. Also keep track of the corresponding point in the other file. */ while (i < i_end && !changed[1 + i]) { while (other_changed[1 + j++]) /* Non-corresponding lines in the other file will count as the preceding batch of changes. */ other_preceding = j; i++; } if (i == i_end) break; start = i; other_start = j; for (; ;) { /* Now find the end of this run of changes. */ while (i < i_end && changed[1 + i]) i++; end = i; /* If the first changed line matches the following unchanged one, and this run does not follow right after a previous run, and there are no lines deleted from the other file here, then classify the first changed line as unchanged and the following line as changed in its place. */ /* You might ask, how could this run follow right after another? Only because the previous run was shifted here. */ if (end != i_end && equivs[start] == equivs[end] && !other_changed[1 + j] && end != i_end && !((preceding >= 0 && start == preceding) || (other_preceding >= 0 && other_start == other_preceding))) { changed[1 + end] = true; changed[1 + start++] = false; ++i; /* Since one line-that-matches is now before this run instead of after, we must advance in the other file to keep in synch. */ ++j; } else break; } preceding = i; other_preceding = j; } } /** * Number of elements (lines) in this file. */ final int buffered_lines; /** * Vector, indexed by line number, containing an equivalence code for * each line. It is this vector that is actually compared with that * of another file to generate differences. */ private final int[] equivs; /** * Vector, like the previous one except that * the elements for discarded lines have been squeezed out. */ final int[] undiscarded; /** * Vector mapping virtual line numbers (not counting discarded lines) * to real ones (counting those lines). Both are origin-0. */ final int[] realindexes; /** * Total number of nondiscarded lines. */ int nondiscarded_lines; /** * Array, indexed by real origin-1 line number, * containing true for a line that is an insertion or a deletion. * The results of comparison are stored here. */ boolean[] changed_flag; } } static List diff(List a, List b) { Set set = asSet(b); List l = new ArrayList(); for (String s : a) if (!set.contains(s)) l.add(s); return l; } static Set asSet(Object[] array) { HashSet set = new HashSet(); for (Object o : array) if (o != null) set.add(o); return set; } static Set asSet(String[] array) { TreeSet set = new TreeSet(); for (String o : array) if (o != null) set.add(o); return set; } static Set asSet(Collection l) { TreeSet set = new TreeSet(); for (String o : l) if (o != null) set.add(o); return set; } }