feat: updates & github event support

This commit is contained in:
Joel Male
2024-03-28 12:00:41 +10:00
parent 8561675bfc
commit 3cee2ba3f1
1640 changed files with 100049 additions and 152370 deletions
-2
View File
@@ -1,2 +0,0 @@
language: node_js
node_js: '0.10'
+1 -1
View File
@@ -1,4 +1,4 @@
# Fast Diff [![Build Status](https://travis-ci.org/jhchen/fast-diff.svg)](https://travis-ci.org/jhchen/fast-diff)
# Fast Diff ![Build Status](https://github.com/jhchen/fast-diff/actions/workflows/test.yml/badge.svg)
This is a simplified import of the excellent [diff-match-patch](https://code.google.com/p/google-diff-match-patch/) library by [Neil Fraser](https://neil.fraser.name/) into the Node.js environment. The match and patch parts are removed, as well as all the extra diff options. What remains is incredibly fast diffing between two strings.
+12 -11
View File
@@ -1,20 +1,21 @@
declare function diff(
text1: string,
text2: string,
cursorPos?: number | diff.CursorInfo
text1: string,
text2: string,
cursorPos?: number | diff.CursorInfo,
cleanup?: boolean
): diff.Diff[];
declare namespace diff {
type Diff = [-1 | 0 | 1, string];
type Diff = [-1 | 0 | 1, string];
const DELETE: -1;
const INSERT: 1;
const EQUAL: 0;
const DELETE: -1;
const INSERT: 1;
const EQUAL: 0;
interface CursorInfo {
oldRange: { index: number; length: number };
newRange: { index: number; length: number };
}
interface CursorInfo {
oldRange: { index: number; length: number };
newRange: { index: number; length: number };
}
}
export = diff;
+431 -67
View File
@@ -23,7 +23,6 @@
* limitations under the License.
*/
/**
* The data structure representing a diff is an array of tuples:
* [[DIFF_DELETE, 'Hello'], [DIFF_INSERT, 'Goodbye'], [DIFF_EQUAL, ' world.']]
@@ -33,16 +32,16 @@ var DIFF_DELETE = -1;
var DIFF_INSERT = 1;
var DIFF_EQUAL = 0;
/**
* Find the differences between two texts. Simplifies the problem by stripping
* any common prefix or suffix off the texts before diffing.
* @param {string} text1 Old string to be diffed.
* @param {string} text2 New string to be diffed.
* @param {Int|Object} [cursor_pos] Edit position in text1 or object with more info
* @param {boolean} [cleanup] Apply semantic cleanup before returning.
* @return {Array} Array of diff tuples.
*/
function diff_main(text1, text2, cursor_pos, _fix_unicode) {
function diff_main(text1, text2, cursor_pos, cleanup, _fix_unicode) {
// Check for equality
if (text1 === text2) {
if (text1) {
@@ -81,9 +80,11 @@ function diff_main(text1, text2, cursor_pos, _fix_unicode) {
diffs.push([DIFF_EQUAL, commonsuffix]);
}
diff_cleanupMerge(diffs, _fix_unicode);
if (cleanup) {
diff_cleanupSemantic(diffs);
}
return diffs;
};
}
/**
* Find the differences between two texts. Assumes that the texts do not
@@ -113,7 +114,7 @@ function diff_compute_(text1, text2) {
diffs = [
[DIFF_INSERT, longtext.substring(0, i)],
[DIFF_EQUAL, shorttext],
[DIFF_INSERT, longtext.substring(i + shorttext.length)]
[DIFF_INSERT, longtext.substring(i + shorttext.length)],
];
// Swap insertions for deletions if diff is reversed.
if (text1.length > text2.length) {
@@ -125,7 +126,10 @@ function diff_compute_(text1, text2) {
if (shorttext.length === 1) {
// Single character string.
// After the previous speedup, the character can't be an equality.
return [[DIFF_DELETE, text1], [DIFF_INSERT, text2]];
return [
[DIFF_DELETE, text1],
[DIFF_INSERT, text2],
];
}
// Check to see if the problem can be split in two.
@@ -145,8 +149,7 @@ function diff_compute_(text1, text2) {
}
return diff_bisect_(text1, text2);
};
}
/**
* Find the 'middle snake' of a diff, split the problem in two
@@ -177,7 +180,7 @@ function diff_bisect_(text1, text2) {
var delta = text1_length - text2_length;
// If the total number of characters is odd, then the front path will collide
// with the reverse path.
var front = (delta % 2 !== 0);
var front = delta % 2 !== 0;
// Offsets for start and end of k loop.
// Prevents mapping of space beyond the grid.
var k1start = 0;
@@ -196,7 +199,8 @@ function diff_bisect_(text1, text2) {
}
var y1 = x1 - k1;
while (
x1 < text1_length && y1 < text2_length &&
x1 < text1_length &&
y1 < text2_length &&
text1.charAt(x1) === text2.charAt(y1)
) {
x1++;
@@ -233,8 +237,10 @@ function diff_bisect_(text1, text2) {
}
var y2 = x2 - k2;
while (
x2 < text1_length && y2 < text2_length &&
text1.charAt(text1_length - x2 - 1) === text2.charAt(text2_length - y2 - 1)
x2 < text1_length &&
y2 < text2_length &&
text1.charAt(text1_length - x2 - 1) ===
text2.charAt(text2_length - y2 - 1)
) {
x2++;
y2++;
@@ -263,9 +269,11 @@ function diff_bisect_(text1, text2) {
}
// Diff took too long and hit the deadline or
// number of diffs equals number of characters, no commonality at all.
return [[DIFF_DELETE, text1], [DIFF_INSERT, text2]];
};
return [
[DIFF_DELETE, text1],
[DIFF_INSERT, text2],
];
}
/**
* Given the location of the 'middle snake', split the diff in two parts
@@ -287,8 +295,7 @@ function diff_bisectSplit_(text1, text2, x, y) {
var diffsb = diff_main(text1b, text2b);
return diffs.concat(diffsb);
};
}
/**
* Determine the common prefix of two strings.
@@ -326,8 +333,57 @@ function diff_commonPrefix(text1, text2) {
}
return pointermid;
};
}
/**
* Determine if the suffix of one string is the prefix of another.
* @param {string} text1 First string.
* @param {string} text2 Second string.
* @return {number} The number of characters common to the end of the first
* string and the start of the second string.
* @private
*/
function diff_commonOverlap_(text1, text2) {
// Cache the text lengths to prevent multiple calls.
var text1_length = text1.length;
var text2_length = text2.length;
// Eliminate the null case.
if (text1_length == 0 || text2_length == 0) {
return 0;
}
// Truncate the longer string.
if (text1_length > text2_length) {
text1 = text1.substring(text1_length - text2_length);
} else if (text1_length < text2_length) {
text2 = text2.substring(0, text1_length);
}
var text_length = Math.min(text1_length, text2_length);
// Quick check for the worst case.
if (text1 == text2) {
return text_length;
}
// Start by looking for a single character match
// and increase length until no match is found.
// Performance analysis: http://neil.fraser.name/news/2010/11/04/
var best = 0;
var length = 1;
while (true) {
var pattern = text1.substring(text_length - length);
var found = text2.indexOf(pattern);
if (found == -1) {
return best;
}
length += found;
if (
found == 0 ||
text1.substring(text_length - length) == text2.substring(0, length)
) {
best = length;
length++;
}
}
}
/**
* Determine the common suffix of two strings.
@@ -364,8 +420,7 @@ function diff_commonSuffix(text1, text2) {
}
return pointermid;
};
}
/**
* Do the two texts share a substring which is at least half the length of the
@@ -381,7 +436,7 @@ function diff_halfMatch_(text1, text2) {
var longtext = text1.length > text2.length ? text1 : text2;
var shorttext = text1.length > text2.length ? text2 : text1;
if (longtext.length < 4 || shorttext.length * 2 < longtext.length) {
return null; // Pointless.
return null; // Pointless.
}
/**
@@ -400,16 +455,21 @@ function diff_halfMatch_(text1, text2) {
// Start with a 1/4 length substring at position i as a seed.
var seed = longtext.substring(i, i + Math.floor(longtext.length / 4));
var j = -1;
var best_common = '';
var best_common = "";
var best_longtext_a, best_longtext_b, best_shorttext_a, best_shorttext_b;
while ((j = shorttext.indexOf(seed, j + 1)) !== -1) {
var prefixLength = diff_commonPrefix(
longtext.substring(i), shorttext.substring(j));
longtext.substring(i),
shorttext.substring(j)
);
var suffixLength = diff_commonSuffix(
longtext.substring(0, i), shorttext.substring(0, j));
longtext.substring(0, i),
shorttext.substring(0, j)
);
if (best_common.length < suffixLength + prefixLength) {
best_common = shorttext.substring(
j - suffixLength, j) + shorttext.substring(j, j + prefixLength);
best_common =
shorttext.substring(j - suffixLength, j) +
shorttext.substring(j, j + prefixLength);
best_longtext_a = longtext.substring(0, i - suffixLength);
best_longtext_b = longtext.substring(i + prefixLength);
best_shorttext_a = shorttext.substring(0, j - suffixLength);
@@ -418,8 +478,11 @@ function diff_halfMatch_(text1, text2) {
}
if (best_common.length * 2 >= longtext.length) {
return [
best_longtext_a, best_longtext_b,
best_shorttext_a, best_shorttext_b, best_common
best_longtext_a,
best_longtext_b,
best_shorttext_a,
best_shorttext_b,
best_common,
];
} else {
return null;
@@ -427,9 +490,17 @@ function diff_halfMatch_(text1, text2) {
}
// First check if the second quarter is the seed for a half-match.
var hm1 = diff_halfMatchI_(longtext, shorttext, Math.ceil(longtext.length / 4));
var hm1 = diff_halfMatchI_(
longtext,
shorttext,
Math.ceil(longtext.length / 4)
);
// Check again based on the third quarter.
var hm2 = diff_halfMatchI_(longtext, shorttext, Math.ceil(longtext.length / 2));
var hm2 = diff_halfMatchI_(
longtext,
shorttext,
Math.ceil(longtext.length / 2)
);
var hm;
if (!hm1 && !hm2) {
return null;
@@ -457,8 +528,267 @@ function diff_halfMatch_(text1, text2) {
}
var mid_common = hm[4];
return [text1_a, text1_b, text2_a, text2_b, mid_common];
};
}
/**
* Reduce the number of edits by eliminating semantically trivial equalities.
* @param {!Array.<!diff_match_patch.Diff>} diffs Array of diff tuples.
*/
function diff_cleanupSemantic(diffs) {
var changes = false;
var equalities = []; // Stack of indices where equalities are found.
var equalitiesLength = 0; // Keeping our own length var is faster in JS.
/** @type {?string} */
var lastequality = null;
// Always equal to diffs[equalities[equalitiesLength - 1]][1]
var pointer = 0; // Index of current position.
// Number of characters that changed prior to the equality.
var length_insertions1 = 0;
var length_deletions1 = 0;
// Number of characters that changed after the equality.
var length_insertions2 = 0;
var length_deletions2 = 0;
while (pointer < diffs.length) {
if (diffs[pointer][0] == DIFF_EQUAL) {
// Equality found.
equalities[equalitiesLength++] = pointer;
length_insertions1 = length_insertions2;
length_deletions1 = length_deletions2;
length_insertions2 = 0;
length_deletions2 = 0;
lastequality = diffs[pointer][1];
} else {
// An insertion or deletion.
if (diffs[pointer][0] == DIFF_INSERT) {
length_insertions2 += diffs[pointer][1].length;
} else {
length_deletions2 += diffs[pointer][1].length;
}
// Eliminate an equality that is smaller or equal to the edits on both
// sides of it.
if (
lastequality &&
lastequality.length <=
Math.max(length_insertions1, length_deletions1) &&
lastequality.length <= Math.max(length_insertions2, length_deletions2)
) {
// Duplicate record.
diffs.splice(equalities[equalitiesLength - 1], 0, [
DIFF_DELETE,
lastequality,
]);
// Change second copy to insert.
diffs[equalities[equalitiesLength - 1] + 1][0] = DIFF_INSERT;
// Throw away the equality we just deleted.
equalitiesLength--;
// Throw away the previous equality (it needs to be reevaluated).
equalitiesLength--;
pointer = equalitiesLength > 0 ? equalities[equalitiesLength - 1] : -1;
length_insertions1 = 0; // Reset the counters.
length_deletions1 = 0;
length_insertions2 = 0;
length_deletions2 = 0;
lastequality = null;
changes = true;
}
}
pointer++;
}
// Normalize the diff.
if (changes) {
diff_cleanupMerge(diffs);
}
diff_cleanupSemanticLossless(diffs);
// Find any overlaps between deletions and insertions.
// e.g: <del>abcxxx</del><ins>xxxdef</ins>
// -> <del>abc</del>xxx<ins>def</ins>
// e.g: <del>xxxabc</del><ins>defxxx</ins>
// -> <ins>def</ins>xxx<del>abc</del>
// Only extract an overlap if it is as big as the edit ahead or behind it.
pointer = 1;
while (pointer < diffs.length) {
if (
diffs[pointer - 1][0] == DIFF_DELETE &&
diffs[pointer][0] == DIFF_INSERT
) {
var deletion = diffs[pointer - 1][1];
var insertion = diffs[pointer][1];
var overlap_length1 = diff_commonOverlap_(deletion, insertion);
var overlap_length2 = diff_commonOverlap_(insertion, deletion);
if (overlap_length1 >= overlap_length2) {
if (
overlap_length1 >= deletion.length / 2 ||
overlap_length1 >= insertion.length / 2
) {
// Overlap found. Insert an equality and trim the surrounding edits.
diffs.splice(pointer, 0, [
DIFF_EQUAL,
insertion.substring(0, overlap_length1),
]);
diffs[pointer - 1][1] = deletion.substring(
0,
deletion.length - overlap_length1
);
diffs[pointer + 1][1] = insertion.substring(overlap_length1);
pointer++;
}
} else {
if (
overlap_length2 >= deletion.length / 2 ||
overlap_length2 >= insertion.length / 2
) {
// Reverse overlap found.
// Insert an equality and swap and trim the surrounding edits.
diffs.splice(pointer, 0, [
DIFF_EQUAL,
deletion.substring(0, overlap_length2),
]);
diffs[pointer - 1][0] = DIFF_INSERT;
diffs[pointer - 1][1] = insertion.substring(
0,
insertion.length - overlap_length2
);
diffs[pointer + 1][0] = DIFF_DELETE;
diffs[pointer + 1][1] = deletion.substring(overlap_length2);
pointer++;
}
}
pointer++;
}
pointer++;
}
}
var nonAlphaNumericRegex_ = /[^a-zA-Z0-9]/;
var whitespaceRegex_ = /\s/;
var linebreakRegex_ = /[\r\n]/;
var blanklineEndRegex_ = /\n\r?\n$/;
var blanklineStartRegex_ = /^\r?\n\r?\n/;
/**
* Look for single edits surrounded on both sides by equalities
* which can be shifted sideways to align the edit to a word boundary.
* e.g: The c<ins>at c</ins>ame. -> The <ins>cat </ins>came.
* @param {!Array.<!diff_match_patch.Diff>} diffs Array of diff tuples.
*/
function diff_cleanupSemanticLossless(diffs) {
/**
* Given two strings, compute a score representing whether the internal
* boundary falls on logical boundaries.
* Scores range from 6 (best) to 0 (worst).
* Closure, but does not reference any external variables.
* @param {string} one First string.
* @param {string} two Second string.
* @return {number} The score.
* @private
*/
function diff_cleanupSemanticScore_(one, two) {
if (!one || !two) {
// Edges are the best.
return 6;
}
// Each port of this function behaves slightly differently due to
// subtle differences in each language's definition of things like
// 'whitespace'. Since this function's purpose is largely cosmetic,
// the choice has been made to use each language's native features
// rather than force total conformity.
var char1 = one.charAt(one.length - 1);
var char2 = two.charAt(0);
var nonAlphaNumeric1 = char1.match(nonAlphaNumericRegex_);
var nonAlphaNumeric2 = char2.match(nonAlphaNumericRegex_);
var whitespace1 = nonAlphaNumeric1 && char1.match(whitespaceRegex_);
var whitespace2 = nonAlphaNumeric2 && char2.match(whitespaceRegex_);
var lineBreak1 = whitespace1 && char1.match(linebreakRegex_);
var lineBreak2 = whitespace2 && char2.match(linebreakRegex_);
var blankLine1 = lineBreak1 && one.match(blanklineEndRegex_);
var blankLine2 = lineBreak2 && two.match(blanklineStartRegex_);
if (blankLine1 || blankLine2) {
// Five points for blank lines.
return 5;
} else if (lineBreak1 || lineBreak2) {
// Four points for line breaks.
return 4;
} else if (nonAlphaNumeric1 && !whitespace1 && whitespace2) {
// Three points for end of sentences.
return 3;
} else if (whitespace1 || whitespace2) {
// Two points for whitespace.
return 2;
} else if (nonAlphaNumeric1 || nonAlphaNumeric2) {
// One point for non-alphanumeric.
return 1;
}
return 0;
}
var pointer = 1;
// Intentionally ignore the first and last element (don't need checking).
while (pointer < diffs.length - 1) {
if (
diffs[pointer - 1][0] == DIFF_EQUAL &&
diffs[pointer + 1][0] == DIFF_EQUAL
) {
// This is a single edit surrounded by equalities.
var equality1 = diffs[pointer - 1][1];
var edit = diffs[pointer][1];
var equality2 = diffs[pointer + 1][1];
// First, shift the edit as far left as possible.
var commonOffset = diff_commonSuffix(equality1, edit);
if (commonOffset) {
var commonString = edit.substring(edit.length - commonOffset);
equality1 = equality1.substring(0, equality1.length - commonOffset);
edit = commonString + edit.substring(0, edit.length - commonOffset);
equality2 = commonString + equality2;
}
// Second, step character by character right, looking for the best fit.
var bestEquality1 = equality1;
var bestEdit = edit;
var bestEquality2 = equality2;
var bestScore =
diff_cleanupSemanticScore_(equality1, edit) +
diff_cleanupSemanticScore_(edit, equality2);
while (edit.charAt(0) === equality2.charAt(0)) {
equality1 += edit.charAt(0);
edit = edit.substring(1) + equality2.charAt(0);
equality2 = equality2.substring(1);
var score =
diff_cleanupSemanticScore_(equality1, edit) +
diff_cleanupSemanticScore_(edit, equality2);
// The >= encourages trailing rather than leading whitespace on edits.
if (score >= bestScore) {
bestScore = score;
bestEquality1 = equality1;
bestEdit = edit;
bestEquality2 = equality2;
}
}
if (diffs[pointer - 1][1] != bestEquality1) {
// We have an improvement, save it back to the diff.
if (bestEquality1) {
diffs[pointer - 1][1] = bestEquality1;
} else {
diffs.splice(pointer - 1, 1);
pointer--;
}
diffs[pointer][1] = bestEdit;
if (bestEquality2) {
diffs[pointer + 1][1] = bestEquality2;
} else {
diffs.splice(pointer + 1, 1);
pointer--;
}
}
}
pointer++;
}
}
/**
* Reorder and merge like edit sections. Merge equalities.
@@ -467,12 +797,12 @@ function diff_halfMatch_(text1, text2) {
* @param {boolean} fix_unicode Whether to normalize to a unicode-correct diff
*/
function diff_cleanupMerge(diffs, fix_unicode) {
diffs.push([DIFF_EQUAL, '']); // Add a dummy entry at the end.
diffs.push([DIFF_EQUAL, ""]); // Add a dummy entry at the end.
var pointer = 0;
var count_delete = 0;
var count_insert = 0;
var text_delete = '';
var text_insert = '';
var text_delete = "";
var text_insert = "";
var commonlength;
while (pointer < diffs.length) {
if (pointer < diffs.length - 1 && !diffs[pointer][1]) {
@@ -481,7 +811,6 @@ function diff_cleanupMerge(diffs, fix_unicode) {
}
switch (diffs[pointer][0]) {
case DIFF_INSERT:
count_insert++;
text_insert += diffs[pointer][1];
pointer++;
@@ -504,9 +833,15 @@ function diff_cleanupMerge(diffs, fix_unicode) {
// inserting 'AC', and then the common suffix 'AC' will be eliminated. in this
// particular case, both equalities go away, we absorb any previous inequalities,
// and we keep scanning for the next equality before rewriting the tuples.
if (previous_equality >= 0 && ends_with_pair_start(diffs[previous_equality][1])) {
if (
previous_equality >= 0 &&
ends_with_pair_start(diffs[previous_equality][1])
) {
var stray = diffs[previous_equality][1].slice(-1);
diffs[previous_equality][1] = diffs[previous_equality][1].slice(0, -1);
diffs[previous_equality][1] = diffs[previous_equality][1].slice(
0,
-1
);
text_delete = stray + text_delete;
text_insert = stray + text_insert;
if (!diffs[previous_equality][1]) {
@@ -546,9 +881,15 @@ function diff_cleanupMerge(diffs, fix_unicode) {
commonlength = diff_commonPrefix(text_insert, text_delete);
if (commonlength !== 0) {
if (previous_equality >= 0) {
diffs[previous_equality][1] += text_insert.substring(0, commonlength);
diffs[previous_equality][1] += text_insert.substring(
0,
commonlength
);
} else {
diffs.splice(0, 0, [DIFF_EQUAL, text_insert.substring(0, commonlength)]);
diffs.splice(0, 0, [
DIFF_EQUAL,
text_insert.substring(0, commonlength),
]);
pointer++;
}
text_insert = text_insert.substring(commonlength);
@@ -558,9 +899,16 @@ function diff_cleanupMerge(diffs, fix_unicode) {
commonlength = diff_commonSuffix(text_insert, text_delete);
if (commonlength !== 0) {
diffs[pointer][1] =
text_insert.substring(text_insert.length - commonlength) + diffs[pointer][1];
text_insert = text_insert.substring(0, text_insert.length - commonlength);
text_delete = text_delete.substring(0, text_delete.length - commonlength);
text_insert.substring(text_insert.length - commonlength) +
diffs[pointer][1];
text_insert = text_insert.substring(
0,
text_insert.length - commonlength
);
text_delete = text_delete.substring(
0,
text_delete.length - commonlength
);
}
}
// Delete the offending records and add the merged ones.
@@ -575,7 +923,12 @@ function diff_cleanupMerge(diffs, fix_unicode) {
diffs.splice(pointer - n, n, [DIFF_DELETE, text_delete]);
pointer = pointer - n + 1;
} else {
diffs.splice(pointer - n, n, [DIFF_DELETE, text_delete], [DIFF_INSERT, text_insert]);
diffs.splice(
pointer - n,
n,
[DIFF_DELETE, text_delete],
[DIFF_INSERT, text_insert]
);
pointer = pointer - n + 2;
}
}
@@ -588,13 +941,13 @@ function diff_cleanupMerge(diffs, fix_unicode) {
}
count_insert = 0;
count_delete = 0;
text_delete = '';
text_insert = '';
text_delete = "";
text_insert = "";
break;
}
}
if (diffs[diffs.length - 1][1] === '') {
diffs.pop(); // Remove the dummy entry at the end.
if (diffs[diffs.length - 1][1] === "") {
diffs.pop(); // Remove the dummy entry at the end.
}
// Second pass: look for single edits surrounded on both sides by equalities
@@ -604,20 +957,30 @@ function diff_cleanupMerge(diffs, fix_unicode) {
pointer = 1;
// Intentionally ignore the first and last element (don't need checking).
while (pointer < diffs.length - 1) {
if (diffs[pointer - 1][0] === DIFF_EQUAL &&
diffs[pointer + 1][0] === DIFF_EQUAL) {
if (
diffs[pointer - 1][0] === DIFF_EQUAL &&
diffs[pointer + 1][0] === DIFF_EQUAL
) {
// This is a single edit surrounded by equalities.
if (diffs[pointer][1].substring(diffs[pointer][1].length -
diffs[pointer - 1][1].length) === diffs[pointer - 1][1]) {
if (
diffs[pointer][1].substring(
diffs[pointer][1].length - diffs[pointer - 1][1].length
) === diffs[pointer - 1][1]
) {
// Shift the edit over the previous equality.
diffs[pointer][1] = diffs[pointer - 1][1] +
diffs[pointer][1].substring(0, diffs[pointer][1].length -
diffs[pointer - 1][1].length);
diffs[pointer][1] =
diffs[pointer - 1][1] +
diffs[pointer][1].substring(
0,
diffs[pointer][1].length - diffs[pointer - 1][1].length
);
diffs[pointer + 1][1] = diffs[pointer - 1][1] + diffs[pointer + 1][1];
diffs.splice(pointer - 1, 1);
changes = true;
} else if (diffs[pointer][1].substring(0, diffs[pointer + 1][1].length) ==
diffs[pointer + 1][1]) {
} else if (
diffs[pointer][1].substring(0, diffs[pointer + 1][1].length) ==
diffs[pointer + 1][1]
) {
// Shift the edit over the next equality.
diffs[pointer - 1][1] += diffs[pointer + 1][1];
diffs[pointer][1] =
@@ -633,14 +996,14 @@ function diff_cleanupMerge(diffs, fix_unicode) {
if (changes) {
diff_cleanupMerge(diffs, fix_unicode);
}
};
}
function is_surrogate_pair_start(charCode) {
return charCode >= 0xD800 && charCode <= 0xDBFF;
return charCode >= 0xd800 && charCode <= 0xdbff;
}
function is_surrogate_pair_end(charCode) {
return charCode >= 0xDC00 && charCode <= 0xDFFF;
return charCode >= 0xdc00 && charCode <= 0xdfff;
}
function starts_with_pair_end(str) {
@@ -669,16 +1032,17 @@ function make_edit_splice(before, oldMiddle, newMiddle, after) {
[DIFF_EQUAL, before],
[DIFF_DELETE, oldMiddle],
[DIFF_INSERT, newMiddle],
[DIFF_EQUAL, after]
[DIFF_EQUAL, after],
]);
}
function find_cursor_edit_diff(oldText, newText, cursor_pos) {
// note: this runs after equality check has ruled out exact equality
var oldRange = typeof cursor_pos === 'number' ?
{ index: cursor_pos, length: 0 } : cursor_pos.oldRange;
var newRange = typeof cursor_pos === 'number' ?
null : cursor_pos.newRange;
var oldRange =
typeof cursor_pos === "number"
? { index: cursor_pos, length: 0 }
: cursor_pos.oldRange;
var newRange = typeof cursor_pos === "number" ? null : cursor_pos.newRange;
// take into account the old and new selection to generate the best diff
// possible for a text edit. for example, a text change from "xxx" to "xx"
// could be a delete or forwards-delete of any one of the x's, or the
@@ -761,10 +1125,10 @@ function find_cursor_edit_diff(oldText, newText, cursor_pos) {
return null;
}
function diff(text1, text2, cursor_pos) {
function diff(text1, text2, cursor_pos, cleanup) {
// only pass fix_unicode=true at the top level, not when diff_main is
// recursively invoked
return diff_main(text1, text2, cursor_pos, true);
return diff_main(text1, text2, cursor_pos, cleanup, true);
}
diff.INSERT = DIFF_INSERT;
+8 -4
View File
@@ -1,13 +1,17 @@
{
"name": "fast-diff",
"version": "1.2.0",
"version": "1.3.0",
"description": "Fast Javascript text diff",
"author": "Jason Chen <jhchen7@gmail.com>",
"main": "diff.js",
"types": "diff.d.ts",
"files": [
"diff.d.ts"
],
"devDependencies": {
"lodash": "~3.9.3",
"seedrandom": "~2.4.0"
"lodash": "~4.17.21",
"nyc": "~15.1.0",
"seedrandom": "~3.0.5"
},
"repository": {
"type": "git",
@@ -17,7 +21,7 @@
"url": "https://github.com/jhchen/fast-diff/issues"
},
"scripts": {
"test": "node test.js"
"test": "nyc node test.js"
},
"license": "Apache-2.0",
"keywords": [
-334
View File
@@ -1,334 +0,0 @@
var _ = require('lodash');
var seedrandom = require('seedrandom');
var diff = require('./diff.js');
var ITERATIONS = 10000;
var ALPHABET = 'GATTACA';
var LENGTH = 100;
var EMOJI_MAX_LENGTH = 50;
var seed = Math.floor(Math.random() * 10000);
var random = seedrandom(seed);
console.log('Running regression tests...');
[
['GAATAAAAAAAGATTAACAT', 'AAAAACTTGTAATTAACAAC'],
['🔘🤘🔗🔗', '🔗🤗🤗__🤗🤘🤘🤗🔗🤘🔗'],
['🔗🤗🤗__🤗🤘🤘🤗🔗🤘🔗', '🤗🤘🔘'],
['🤘🤘🔘🔘_🔘🔗🤘🤗🤗__🔗🤘', '🤘🔘🤘🔗🤘🤘🔗🤗🤘🔘🔘'],
['🤗🤘🤗🔘🤘🔘🤗_🤗🔗🤘🤗_🤘🔗🤗🤘🔗🤘🤘🤘🔗🤗🔗🔗🔗🤗_🤘🔗🤗🤗🔘🤗🤗🤘🤗',
'_🤗🤘_🤘🤘🔘🤗🔘🤘_🔘🤗🔗🔘🔗🤘🔗🤘🤗🔗🔗🔗🤘🔘_🤗🤘🤘🤘__🤘_🔘🤘🤘_🔗🤘🔘'],
['🔗🤘🤗🔘🔘🤗', '🤘🤘🤘🤗🔘🔗🔗'],
['🔘_🔗🔗🔗🤗🔗', '🤘🤗🔗🤗_🤘🔘_'],
].forEach(function (data) {
var result = diff(data[0], data[1]);
applyDiff(result, data[0], data[1]);
});
console.log('Running computing ' + ITERATIONS + ' diffs with seed ' + seed + '...');
console.log('Generating strings...');
var strings = [];
for (var i = 0; i <= ITERATIONS; ++i) {
var chars = [];
for (var l = 0; l < LENGTH; ++l) {
var letter = ALPHABET.substr(Math.floor(random() * ALPHABET.length), 1);
chars.push(letter);
}
strings.push(chars.join(''));
}
console.log('Running fuzz tests *without* cursor information...');
for (var i = 0; i < ITERATIONS; ++i) {
var result = diff(strings[i], strings[i + 1]);
applyDiff(result, strings[i], strings[i + 1]);
}
console.log('Running fuzz tests *with* cursor information');
for (var i = 0; i < ITERATIONS; ++i) {
var cursor_pos = Math.floor(random() * strings[i].length + 1);
var diffs = diff(strings[i], strings[i + 1], cursor_pos);
applyDiff(diffs, strings[i], strings[i + 1]);
}
function parseDiff(str) {
if (!str) {
return [];
}
return str.split(/(?=[+\-=])/).map(function (piece) {
var symbol = piece.charAt(0);
var text = piece.slice(1);
return [
symbol === '+' ? diff.INSERT : symbol === '-' ? diff.DELETE : diff.EQUAL,
text
]
});
}
console.log('Running cursor tests');
[
['', 0, '', null, ''],
['', 0, 'a', null, '+a'],
['a', 0, 'aa', null, '+a=a'],
['a', 1, 'aa', null, '=a+a'],
['aa', 0, 'aaa', null, '+a=aa'],
['aa', 1, 'aaa', null, '=a+a=a'],
['aa', 2, 'aaa', null, '=aa+a'],
['aaa', 0, 'aaaa', null, '+a=aaa'],
['aaa', 1, 'aaaa', null, '=a+a=aa'],
['aaa', 2, 'aaaa', null, '=aa+a=a'],
['aaa', 3, 'aaaa', null, '=aaa+a'],
['a', 0, '', null, '-a'],
['a', 1, '', null, '-a'],
['aa', 0, 'a', null, '-a=a'],
['aa', 1, 'a', null, '-a=a'],
['aa', 2, 'a', null, '=a-a'],
['aaa', 0, 'aa', null, '-a=aa'],
['aaa', 1, 'aa', null, '-a=aa'],
['aaa', 2, 'aa', null, '=a-a=a'],
['aaa', 3, 'aa', null, '=aa-a'],
['', 0, '', 0, ''],
['', 0, 'a', 1, '+a'],
['a', 0, 'aa', 1, '+a=a'],
['a', 1, 'aa', 2, '=a+a'],
['aa', 0, 'aaa', 1, '+a=aa'],
['aa', 1, 'aaa', 2, '=a+a=a'],
['aa', 2, 'aaa', 3, '=aa+a'],
['aaa', 0, 'aaaa', 1, '+a=aaa'],
['aaa', 1, 'aaaa', 2, '=a+a=aa'],
['aaa', 2, 'aaaa', 3, '=aa+a=a'],
['aaa', 3, 'aaaa', 4, '=aaa+a'],
['a', 1, '', 0, '-a'],
['aa', 1, 'a', 0, '-a=a'],
['aa', 2, 'a', 1, '=a-a'],
['aaa', 1, 'aa', 0, '-a=aa'],
['aaa', 2, 'aa', 1, '=a-a=a'],
['aaa', 3, 'aa', 2, '=aa-a'],
['a', 1, '', 0, '-a'],
['aa', 1, 'a', 0, '-a=a'],
['aa', 2, 'a', 1, '=a-a'],
['aaa', 1, 'aa', 0, '-a=aa'],
['aaa', 2, 'aa', 1, '=a-a=a'],
['aaa', 3, 'aa', 2, '=aa-a'],
// forward-delete
['a', 0, '', 0, '-a'],
['aa', 0, 'a', 0, '-a=a'],
['aa', 1, 'a', 1, '=a-a'],
['aaa', 0, 'aa', 0, '-a=aa'],
['aaa', 1, 'aa', 1, '=a-a=a'],
['aaa', 2, 'aa', 2, '=aa-a'],
['bob', 0, 'bobob', null, '+bo=bob'],
['bob', 1, 'bobob', null, '=b+ob=ob'],
['bob', 2, 'bobob', null, '=bo+bo=b'],
['bob', 3, 'bobob', null, '=bob+ob'],
['bob', 0, 'bobob', 2, '+bo=bob'],
['bob', 1, 'bobob', 3, '=b+ob=ob'],
['bob', 2, 'bobob', 4, '=bo+bo=b'],
['bob', 3, 'bobob', 5, '=bob+ob'],
['bobob', 2, 'bob', null, '-bo=bob'],
['bobob', 3, 'bob', null, '=b-ob=ob'],
['bobob', 4, 'bob', null, '=bo-bo=b'],
['bobob', 5, 'bob', null, '=bob-ob'],
['bobob', 2, 'bob', 0, '-bo=bob'],
['bobob', 3, 'bob', 1, '=b-ob=ob'],
['bobob', 4, 'bob', 2, '=bo-bo=b'],
['bobob', 5, 'bob', 3, '=bob-ob'],
['bob', 1, 'b', null, '=b-ob'],
['hello', [0, 5], 'h', 1, '-hello+h'],
['yay', [0, 3], 'y', 1, '-yay+y'],
['bobob', [1, 4], 'bob', 2, '=b-obo+o=b'],
].forEach(function (data) {
var oldText = data[0];
var newText = data[2];
var oldRange = typeof data[1] === 'number' ?
{ index: data[1], length: 0 } :
{ index: data[1][0], length: data[1][1] - data[1][0] };
var newRange = typeof data[3] === 'number' ?
{ index: data[3], length: 0 } :
data[3] === null ? null : { index: data[3][0], length: data[3][1] - data[3][0] };
var expected = parseDiff(data[4]);
if (newRange === null && typeof data[1] !== 'number') {
throw new Error('invalid test case');
}
var cursorInfo = newRange === null ? data[1] : {
oldRange: oldRange,
newRange: newRange,
};
doCursorTest(oldText, newText, cursorInfo, expected);
doCursorTest('x' + oldText, 'x' + newText, shiftCursorInfo(cursorInfo, 1), diffPrepend(expected, 'x'));
doCursorTest(oldText + 'x', newText + 'x', cursorInfo, diffAppend(expected, 'x'));
});
function diffPrepend(tuples, text) {
if (tuples.length > 0 && tuples[0][0] === diff.EQUAL) {
return [[diff.EQUAL, text + tuples[0][1]]].concat(tuples.slice(1));
} else {
return [[diff.EQUAL, text]].concat(tuples);
}
}
function diffAppend(tuples, text) {
var lastTuple = tuples[tuples.length - 1];
if (lastTuple && lastTuple[0] === diff.EQUAL) {
return tuples.slice(0, -1).concat([[diff.EQUAL, lastTuple[1] + text]]);
} else {
return tuples.concat([[diff.EQUAL, text]]);
}
}
function shiftCursorInfo(cursorInfo, amount) {
if (typeof cursorInfo === 'number') {
return cursorInfo + amount;
} else {
return {
oldRange: {
index: cursorInfo.oldRange.index + amount,
length: cursorInfo.oldRange.length,
},
newRange: {
index: cursorInfo.newRange.index + amount,
length: cursorInfo.newRange.length,
},
}
}
}
function doCursorTest(oldText, newText, cursorInfo, expected) {
var result = diff(oldText, newText, cursorInfo);
if (!_.isEqual(result, expected)) {
console.log([oldText, newText, cursorInfo]);
console.log(result, '!==', expected);
throw new Error('cursor test failed');
}
}
console.log('Running emoji tests');
[
['🐶', '🐯', '-🐶+🐯'],
['👨🏽', '👩🏽', '-👨+👩=🏽'],
['👩🏼', '👩🏽', '=👩-🏼+🏽'],
['🍏🍎', '🍎', '-🍏=🍎'],
['🍎', '🍏🍎', '+🍏=🍎'],
].forEach(function (data) {
var oldText = data[0];
var newText = data[1];
var expected = parseDiff(data[2]);
doEmojiTest(oldText, newText, expected);
doEmojiTest('x' + oldText, 'x' + newText, diffPrepend(expected, 'x'));
doEmojiTest(oldText + 'x', newText + 'x', diffAppend(expected, 'x'));
});
function doEmojiTest(oldText, newText, expected) {
var result = diff(oldText, newText);
if (!_.isEqual(result, expected)) {
console.log(oldText, newText, expected);
console.log(result, '!==', expected);
throw new Error('Emoji simple test case failed');
}
}
// emojis chosen to share high and low surrogates!
var EMOJI_ALPHABET = ['_', '🤗', '🔗', '🤘', '🔘'];
console.log('Generating emoji strings...');
var emoji_strings = [];
for (var i = 0; i <= ITERATIONS; ++i) {
var letters = [];
var len = Math.floor(random() * EMOJI_MAX_LENGTH);
for (var l = 0; l < len; ++l) {
var letter = EMOJI_ALPHABET[Math.floor(random() * EMOJI_ALPHABET.length)];
letters.push(letter);
}
emoji_strings.push(letters.join(''));
}
console.log('Running emoji fuzz tests...');
for (var i = 0; i < ITERATIONS; ++i) {
var oldText = emoji_strings[i];
var newText = emoji_strings[i + 1];
var result = diff(oldText, newText);
applyDiff(result, oldText, newText);
}
// Applies a diff to text, throwing an error if diff is invalid or incorrect
function applyDiff(diffs, text, expectedResult) {
var pos = 0;
function throwError(message) {
console.log(diffs, text, expectedResult);
throw new Error(message);
}
function expect(expected) {
var found = text.substr(pos, expected.length);
if (found !== expected) {
throwError('Expected "' + expected + '", found "' + found + '"');
}
}
var result = '';
var inserts_since_last_equality = 0;
var deletes_since_last_equality = 0;
for (var i = 0; i < diffs.length; i++) {
var d = diffs[i];
if (!d[1]) {
throwError('Empty tuple in diff')
}
var firstCharCode = d[1].charCodeAt(0);
var lastCharCode = d[1].slice(-1).charCodeAt(0);
if (firstCharCode >= 0xDC00 && firstCharCode <= 0xDFFF ||
lastCharCode >= 0xD800 && lastCharCode <= 0xDBFF) {
throwError('Bad unicode diff tuple')
}
switch (d[0]) {
case diff.EQUAL:
if (i !== 0 && !inserts_since_last_equality && !deletes_since_last_equality) {
throwError('two consecutive equalities in diff');
}
inserts_since_last_equality = 0;
deletes_since_last_equality = 0;
expect(d[1]);
result += d[1];
pos += d[1].length;
break;
case diff.DELETE:
if (deletes_since_last_equality) {
throwError('multiple deletes between equalities')
}
if (inserts_since_last_equality) {
throwError('delete following insert in diff')
}
deletes_since_last_equality++;
expect(d[1]);
pos += d[1].length;
break
case diff.INSERT:
if (inserts_since_last_equality) {
throwError('multiple inserts between equalities')
}
inserts_since_last_equality++;
result += d[1];
break;
}
}
if (pos !== text.length) {
throwError('Diff did not consume entire input text');
}
if (result !== expectedResult) {
console.log(diffs, text, expectedResult, result);
throw new Error('Diff not correct')
}
return result;
}
console.log("Success!");