diff.java

JSPWiki,100%Java开发的一套完整WIKI程序

     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 f = fdiag[fdiagoff + d];
            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]);
    }

    /** Scan the tables of which lines are inserted and deleted,
     producing an edit script in reverse order.  */

    private change build_reverse_script() {
        change script = null;
        final boolean[] changed0 = filevec[0].changed_flag;
        final boolean[] changed1 = filevec[1].changed_flag;
        final int len0 = filevec[0].buffered_lines;
        final int len1 = filevec[1].buffered_lines;

        /* Note that changedN[len0] does exist, and contains 0.  */

        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;
    }

    /** Scan the tables of which lines are inserted and deleted,
     producing an edit script in forward order.  */

    private change build_script() {
        change script = null;
        final boolean[] changed0 = filevec[0].changed_flag;
        final boolean[] changed1 = filevec[1].changed_flag;
        final int len0 = filevec[0].buffered_lines;
        final int len1 = filevec[1].buffered_lines;
        int i0 = len0, i1 = len1;

        /* Note that changedN[-1] does exist, and contains 0.  */

        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;
    }

    /* Report the differences of two files.  DEPTH is the current directory
       depth. */
    public change diff_2(final boolean reverse) {

        /* 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.  */

        if (reverse)
            return build_reverse_script();
        else
            return build_script();
    }

    /** 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.  */
        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.
         <p>
         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.