/**
* Supplies the content of a file for {@link DiffFormatter}.
- *
+ * <p>
* A content source is not thread-safe. Sources may contain state, including
* information about the last ObjectLoader they returned. Callers must be
* careful to ensure there is no more than one ObjectLoader pending on any
/**
* Compares two {@link Sequence}s to create an {@link EditList} of changes.
- *
+ * <p>
* An algorithm's {@code diff} method must be callable from concurrent threads
* without data collisions. This permits some algorithms to use a singleton
* pattern, with concurrent invocations using the same singleton. Other
/**
* Wraps a {@link Sequence} to assign hash codes to elements.
- *
+ * <p>
* This sequence acts as a proxy for the real sequence, caching element hash
* codes so they don't need to be recomputed each time. Sequences of this type
* must be used with a {@link HashedSequenceComparator}.
- *
+ * <p>
* To construct an instance of this type use {@link HashedSequencePair}.
*
* @param <S>
/**
* Wrap another comparator for use with {@link HashedSequence}.
- *
+ * <p>
* This comparator acts as a proxy for the real comparator, evaluating the
* cached hash code before testing the underlying comparator's equality.
* Comparators of this type must be used with a {@link HashedSequence}.
- *
+ * <p>
* To construct an instance of this type use {@link HashedSequencePair}.
*
* @param <S>
/**
* Wraps two {@link Sequence} instances to cache their element hash codes.
- *
+ * <p>
* This pair wraps two sequences that contain cached hash codes for the input
* sequences.
*
/**
* An extended form of Bram Cohen's patience diff algorithm.
- *
+ * <p>
* This implementation was derived by using the 4 rules that are outlined in
* Bram Cohen's <a href="http://bramcohen.livejournal.com/73318.html">blog</a>,
* and then was further extended to support low-occurrence common elements.
- *
+ * <p>
* The basic idea of the algorithm is to create a histogram of occurrences for
* each element of sequence A. Each element of sequence B is then considered in
* turn. If the element also exists in sequence A, and has a lower occurrence
* lowest number of occurrences is chosen as a split point. The region is split
* around the LCS, and the algorithm is recursively applied to the sections
* before and after the LCS.
- *
+ * <p>
* By always selecting a LCS position with the lowest occurrence count, this
* algorithm behaves exactly like Bram Cohen's patience diff whenever there is a
* unique common element available between the two sequences. When no unique
* elements exist, the lowest occurrence element is chosen instead. This offers
* more readable diffs than simply falling back on the standard Myers' O(ND)
* algorithm would produce.
- *
+ * <p>
* To prevent the algorithm from having an O(N^2) running time, an upper limit
* on the number of unique elements in a histogram bucket is configured by
* {@link #setMaxChainLength(int)}. If sequence A has more than this many
* elements that hash into the same hash bucket, the algorithm passes the region
* to {@link #setFallbackAlgorithm(DiffAlgorithm)}. If no fallback algorithm is
* configured, the region is emitted as a replace edit.
- *
+ * <p>
* During scanning of sequence B, any element of A that occurs more than
* {@link #setMaxChainLength(int)} times is never considered for an LCS match
* position, even if it is common between the two sequences. This limits the
* number of locations in sequence A that must be considered to find the LCS,
* and helps maintain a lower running time bound.
- *
+ * <p>
* So long as {@link #setMaxChainLength(int)} is a small constant (such as 64),
* the algorithm runs in O(N * D) time, where N is the sum of the input lengths
* and D is the number of edits in the resulting EditList. If the supplied
* {@link SequenceComparator} has a good hash function, this implementation
* typically out-performs {@link MyersDiff}, even though its theoretical running
* time is the same.
- *
+ * <p>
* This implementation has an internal limitation that prevents it from handling
* sequences with more than 268,435,456 (2^28) elements.
*/
/**
* Support {@link HistogramDiff} by computing occurrence counts of elements.
- *
+ * <p>
* Each element in the range being considered is put into a hash table, tracking
* the number of times that distinct element appears in the sequence. Once all
* elements have been inserted from sequence A, each element of sequence B is
import org.eclipse.jgit.util.LongList;
/**
- * Diff algorithm, based on "An O(ND) Difference Algorithm and its
- * Variations", by Eugene Myers.
- *
+ * Diff algorithm, based on "An O(ND) Difference Algorithm and its Variations",
+ * by Eugene Myers.
+ * <p>
* The basic idea is to put the line numbers of text A as columns ("x") and the
- * lines of text B as rows ("y"). Now you try to find the shortest "edit path"
- * from the upper left corner to the lower right corner, where you can
- * always go horizontally or vertically, but diagonally from (x,y) to
- * (x+1,y+1) only if line x in text A is identical to line y in text B.
- *
- * Myers' fundamental concept is the "furthest reaching D-path on diagonal k":
- * a D-path is an edit path starting at the upper left corner and containing
- * exactly D non-diagonal elements ("differences"). The furthest reaching
- * D-path on diagonal k is the one that contains the most (diagonal) elements
- * which ends on diagonal k (where k = y - x).
- *
+ * lines of text B as rows ("y"). Now you try to find the shortest "edit path"
+ * from the upper left corner to the lower right corner, where you can always go
+ * horizontally or vertically, but diagonally from (x,y) to (x+1,y+1) only if
+ * line x in text A is identical to line y in text B.
+ * <p>
+ * Myers' fundamental concept is the "furthest reaching D-path on diagonal k": a
+ * D-path is an edit path starting at the upper left corner and containing
+ * exactly D non-diagonal elements ("differences"). The furthest reaching D-path
+ * on diagonal k is the one that contains the most (diagonal) elements which
+ * ends on diagonal k (where k = y - x).
+ * <p>
* Example:
*
+ * <pre>
* H E L L O W O R L D
* ____
* L \___
* O \___
* W \________
- *
- * Since every D-path has exactly D horizontal or vertical elements, it can
- * only end on the diagonals -D, -D+2, ..., D-2, D.
- *
- * Since every furthest reaching D-path contains at least one furthest
- * reaching (D-1)-path (except for D=0), we can construct them recursively.
- *
+ * </pre>
+ * <p>
+ * Since every D-path has exactly D horizontal or vertical elements, it can only
+ * end on the diagonals -D, -D+2, ..., D-2, D.
+ * <p>
+ * Since every furthest reaching D-path contains at least one furthest reaching
+ * (D-1)-path (except for D=0), we can construct them recursively.
+ * <p>
* Since we are really interested in the shortest edit path, we can start
* looking for a 0-path, then a 1-path, and so on, until we find a path that
* ends in the lower right corner.
- *
+ * <p>
* To save space, we do not need to store all paths (which has quadratic space
- * requirements), but generate the D-paths simultaneously from both sides.
- * When the ends meet, we will have found "the middle" of the path. From the
- * end points of that diagonal part, we can generate the rest recursively.
- *
+ * requirements), but generate the D-paths simultaneously from both sides. When
+ * the ends meet, we will have found "the middle" of the path. From the end
+ * points of that diagonal part, we can generate the rest recursively.
+ * <p>
* This only requires linear space.
+ * <p>
+ * The overall (runtime) complexity is:
*
- * The overall (runtime) complexity is
- *
- * O(N * D^2 + 2 * N/2 * (D/2)^2 + 4 * N/4 * (D/4)^2 + ...)
- * = O(N * D^2 * 5 / 4) = O(N * D^2),
- *
- * (With each step, we have to find the middle parts of twice as many regions
- * as before, but the regions (as well as the D) are halved.)
- *
- * So the overall runtime complexity stays the same with linear space,
- * albeit with a larger constant factor.
+ * <pre>
+ * O(N * D^2 + 2 * N/2 * (D/2)^2 + 4 * N/4 * (D/4)^2 + ...)
+ * = O(N * D^2 * 5 / 4) = O(N * D^2),
+ * </pre>
+ * <p>
+ * (With each step, we have to find the middle parts of twice as many regions as
+ * before, but the regions (as well as the D) are halved.)
+ * <p>
+ * So the overall runtime complexity stays the same with linear space, albeit
+ * with a larger constant factor.
*
* @param <S>
* type of sequence.
/**
* Arbitrary sequence of elements.
- *
+ * <p>
* A sequence of elements is defined to contain elements in the index range
* <code>[0, {@link #size()})</code>, like a standard Java List implementation.
* Unlike a List, the members of the sequence are not directly obtainable.
- *
+ * <p>
* Implementations of Sequence are primarily intended for use in content
* difference detection algorithms, to produce an {@link EditList} of
* {@link Edit} instances describing how two Sequence instances differ.
- *
+ * <p>
* To be compared against another Sequence of the same type, a supporting
* {@link SequenceComparator} must also be supplied.
*/
/**
* Equivalence function for a {@link Sequence} compared by difference algorithm.
- *
+ * <p>
* Difference algorithms can use a comparator to compare portions of two
* sequences and discover the minimal edits required to transform from one
* sequence to the other sequence.
- *
+ * <p>
* Indexes within a sequence are zero-based.
*
* @param <S>
/**
* Wraps a {@link Sequence} to have a narrower range of elements.
- *
+ * <p>
* This sequence acts as a proxy for the real sequence, translating element
* indexes on the fly by adding {@code begin} to them. Sequences of this type
* must be used with a {@link SubsequenceComparator}.
public final class Subsequence<S extends Sequence> extends Sequence {
/**
* Construct a subsequence around the A region/base sequence.
- *
+ *
* @param <S>
* the base sequence type.
* @param a
/**
* Construct a subsequence around the B region/base sequence.
- *
+ *
* @param <S>
* the base sequence type.
* @param b
/**
* Adjust the Edit to reflect positions in the base sequence.
- *
+ *
* @param <S>
* the base sequence type.
* @param e
/**
* Wrap another comparator for use with {@link Subsequence}.
- *
+ * <p>
* This comparator acts as a proxy for the real comparator, translating element
* indexes on the fly by adding the subsequence's begin offset to them.
* Comparators of this type must be used with a {@link Subsequence}.
projects / jgit.git / commitdiff