package/package.json000644 001751 000062 0000001613 12066005331013010 0ustar00000000 000000 { "contributors": [ { "name": "Shimon Doodkin", "email": "helpmepro1@gmail.com", "homepage": "http://doodkin.com" }, { "name": "Ryan Graham", "email": "r.m.graham@gmail.com", "homepage": "https://github.com/rmg" } ], "author": "Neil Fraser (http://neil.fraser.name/)", "name": "googlediff", "description": "Simple packaging of Neil Fraser's original google-diff-match-patch", "version": "0.1.0", "homepage": "http://code.google.com/p/google-diff-match-patch/", "repository": { "type": "git", "url": "git://github.com/shimondoodkin/googlediff.git" }, "scripts": { "test": "node test.js" }, "engines": { "node": ">=0.8.0" }, "dependencies": { }, "devDependencies": {}, "licenses": [ { "type": "Apache", "url": "http://www.apache.org/licenses/LICENSE-2.0" } ] } package/README000644 001751 000062 0000003117 12066005331011403 0ustar00000000 000000 node module for: http://code.google.com/p/google-diff-match-patch/ by Neil Fraser and contributors. drop in the original code form svn of Neil Fraser's diff_match_patch. uses jsinc and a simple index.js file and thus it is easily updatable to update just copy in here the 'demos' and 'javascript' folders this module is also awesome because you can use exactrly the same code on the serverside and on the client side. this module is built using https://github.com/shimondoodkin/jsinc the test tests just that the modue is usable. see demos to learn how it works. by Shimon Dookdin may be intalled with npm: npm install googlediff demo: var diff_match_patch=require('googlediff'); var dmp =new diff_match_patch(); function launch() { var text1 = "this is some test. blah blah blah"; var text2 = "this is other text. blah blah blah"; //dmp.Diff_Timeout = 1; // set 0 for no timeout var ms_start = (new Date()).getTime(); var d = dmp.diff_main(text1, text2); var ms_end = (new Date()).getTime(); if (true) { dmp.diff_cleanupSemantic(d); } if (false) { dmp.Diff_EditCost = 4; dmp.diff_cleanupEfficiency(d); } //var ds = dmp.diff_prettyHtml(d); //console.log(ds + '
Time: ' + (ms_end - ms_start) / 1000 + 's'); console.log(d) } launch() output: [ [ 0, 'this is ' ], [ -1, 'some' ], [ 1, 'other' ], [ 0, ' te' ], [ -1, 's' ], [ 1, 'x' ], [ 0, 't. blah blah blah' ] ] package/index.js000644 001751 000062 0000000134 12066005331012164 0ustar00000000 000000 module.exports = require('./javascript/diff_match_patch_uncompressed.js').diff_match_patch; package/test.js000644 001751 000062 0000001432 12066005331012036 0ustar00000000 000000 var assert = require('assert'); var diff_match_patch=require('./index'); var dmp = new diff_match_patch(); console.dir(diff_match_patch); function launch() { var text1 = "this is some test. blah blah blah"; var text2 = "this is other text. blah blah blah"; dmp.Diff_Timeout = 1; var ms_start = (new Date()).getTime(); var d = dmp.diff_main(text1, text2); var ms_end = (new Date()).getTime(); if (true) { dmp.diff_cleanupSemantic(d); } if (false) { dmp.Diff_EditCost = 4; dmp.diff_cleanupEfficiency(d); } //var ds = dmp.diff_prettyHtml(d); //console.log(ds + '
Time: ' + (ms_end - ms_start) / 1000 + 's'); console.log(d); var patch = dmp.patch_make(d); var results = dmp.patch_apply(patch, text1); assert.equal(results[0], text2); } launch() package/javascript/diff_match_patch.js000755 001751 000062 0000045370 12066005331016504 0ustar00000000 000000 (function(){function diff_match_patch(){this.Diff_Timeout=1;this.Diff_EditCost=4;this.Match_Threshold=0.5;this.Match_Distance=1E3;this.Patch_DeleteThreshold=0.5;this.Patch_Margin=4;this.Match_MaxBits=32} diff_match_patch.prototype.diff_main=function(a,b,c,d){"undefined"==typeof d&&(d=0>=this.Diff_Timeout?Number.MAX_VALUE:(new Date).getTime()+1E3*this.Diff_Timeout);if(null==a||null==b)throw Error("Null input. (diff_main)");if(a==b)return a?[[0,a]]:[];"undefined"==typeof c&&(c=!0);var e=c,f=this.diff_commonPrefix(a,b);c=a.substring(0,f);a=a.substring(f);b=b.substring(f);var f=this.diff_commonSuffix(a,b),g=a.substring(a.length-f);a=a.substring(0,a.length-f);b=b.substring(0,b.length-f);a=this.diff_compute_(a, b,e,d);c&&a.unshift([0,c]);g&&a.push([0,g]);this.diff_cleanupMerge(a);return a}; diff_match_patch.prototype.diff_compute_=function(a,b,c,d){if(!a)return[[1,b]];if(!b)return[[-1,a]];var e=a.length>b.length?a:b,f=a.length>b.length?b:a,g=e.indexOf(f);return-1!=g?(c=[[1,e.substring(0,g)],[0,f],[1,e.substring(g+f.length)]],a.length>b.length&&(c[0][0]=c[2][0]=-1),c):1==f.length?[[-1,a],[1,b]]:(e=this.diff_halfMatch_(a,b))?(f=e[0],a=e[1],g=e[2],b=e[3],e=e[4],f=this.diff_main(f,g,c,d),c=this.diff_main(a,b,c,d),f.concat([[0,e]],c)):c&&100c);v++){for(var n=-v+r;n<=v-t;n+=2){var l=g+n,m;m=n==-v||n!=v&&j[l-1]d)t+=2;else if(s>e)r+=2;else if(q&&(l=g+k-n,0<=l&&l= u)return this.diff_bisectSplit_(a,b,m,s,c)}}for(n=-v+p;n<=v-w;n+=2){l=g+n;u=n==-v||n!=v&&i[l-1]d)w+=2;else if(m>e)p+=2;else if(!q&&(l=g+k-n,0<=l&&(l=u)))return this.diff_bisectSplit_(a,b,m,s,c)}}return[[-1,a],[1,b]]}; diff_match_patch.prototype.diff_bisectSplit_=function(a,b,c,d,e){var f=a.substring(0,c),g=b.substring(0,d);a=a.substring(c);b=b.substring(d);f=this.diff_main(f,g,!1,e);e=this.diff_main(a,b,!1,e);return f.concat(e)}; diff_match_patch.prototype.diff_linesToChars_=function(a,b){function c(a){for(var b="",c=0,f=-1,g=d.length;fd?a=a.substring(c-d):c=a.length?[h,j,n,l,g]:null}if(0>=this.Diff_Timeout)return null; var d=a.length>b.length?a:b,e=a.length>b.length?b:a;if(4>d.length||2*e.lengthd[4].length?g:d:d:g;var j;a.length>b.length?(g=h[0],d=h[1],e=h[2],j=h[3]):(e=h[0],j=h[1],g=h[2],d=h[3]);h=h[4];return[g,d,e,j,h]}; diff_match_patch.prototype.diff_cleanupSemantic=function(a){for(var b=!1,c=[],d=0,e=null,f=0,g=0,h=0,j=0,i=0;f=e){if(d>=b.length/2||d>=c.length/2)a.splice(f,0,[0,c.substring(0,d)]),a[f-1][1]=b.substring(0,b.length-d),a[f+1][1]=c.substring(d),f++}else if(e>=b.length/2||e>=c.length/2)a.splice(f,0,[0,b.substring(0,e)]),a[f-1][0]=1,a[f-1][1]=c.substring(0,c.length-e),a[f+1][0]=-1,a[f+1][1]=b.substring(e),f++;f++}f++}}; diff_match_patch.prototype.diff_cleanupSemanticLossless=function(a){function b(a,b){if(!a||!b)return 6;var c=a.charAt(a.length-1),d=b.charAt(0),e=c.match(diff_match_patch.nonAlphaNumericRegex_),f=d.match(diff_match_patch.nonAlphaNumericRegex_),g=e&&c.match(diff_match_patch.whitespaceRegex_),h=f&&d.match(diff_match_patch.whitespaceRegex_),c=g&&c.match(diff_match_patch.linebreakRegex_),d=h&&d.match(diff_match_patch.linebreakRegex_),i=c&&a.match(diff_match_patch.blanklineEndRegex_),j=d&&b.match(diff_match_patch.blanklineStartRegex_); return i||j?5:c||d?4:e&&!g&&h?3:g||h?2:e||f?1:0}for(var c=1;c=i&&(i=k,g=d,h=e,j=f)}a[c-1][1]!=g&&(g?a[c-1][1]=g:(a.splice(c-1,1),c--),a[c][1]= h,j?a[c+1][1]=j:(a.splice(c+1,1),c--))}c++}};diff_match_patch.nonAlphaNumericRegex_=/[^a-zA-Z0-9]/;diff_match_patch.whitespaceRegex_=/\s/;diff_match_patch.linebreakRegex_=/[\r\n]/;diff_match_patch.blanklineEndRegex_=/\n\r?\n$/;diff_match_patch.blanklineStartRegex_=/^\r?\n\r?\n/; diff_match_patch.prototype.diff_cleanupEfficiency=function(a){for(var b=!1,c=[],d=0,e=null,f=0,g=!1,h=!1,j=!1,i=!1;fb)break;e=c;f=d}return a.length!=g&&-1===a[g][0]?f:f+(b-e)}; diff_match_patch.prototype.diff_prettyHtml=function(a){for(var b=[],c=/&/g,d=//g,f=/\n/g,g=0;g");switch(h){case 1:b[g]=''+j+"";break;case -1:b[g]=''+j+"";break;case 0:b[g]=""+j+""}}return b.join("")}; diff_match_patch.prototype.diff_text1=function(a){for(var b=[],c=0;cthis.Match_MaxBits)throw Error("Pattern too long for this browser.");var e=this.match_alphabet_(b),f=this,g=this.Match_Threshold,h=a.indexOf(b,c);-1!=h&&(g=Math.min(d(0,h),g),h=a.lastIndexOf(b,c+b.length),-1!=h&&(g=Math.min(d(0,h),g)));for(var j=1<=i;p--){var w=e[a.charAt(p-1)];k[p]=0===t?(k[p+1]<<1|1)&w:(k[p+1]<<1|1)&w|((r[p+1]|r[p])<<1|1)|r[p+1];if(k[p]&j&&(w=d(t,p-1),w<=g))if(g=w,h=p-1,h>c)i=Math.max(1,2*c-h);else break}if(d(t+1,c)>g)break;r=k}return h}; diff_match_patch.prototype.match_alphabet_=function(a){for(var b={},c=0;c=2*this.Patch_Margin&& e&&(this.patch_addContext_(a,h),c.push(a),a=new diff_match_patch.patch_obj,e=0,h=d,f=g)}1!==i&&(f+=k.length);-1!==i&&(g+=k.length)}e&&(this.patch_addContext_(a,h),c.push(a));return c};diff_match_patch.prototype.patch_deepCopy=function(a){for(var b=[],c=0;cthis.Match_MaxBits){if(j=this.match_main(b,h.substring(0,this.Match_MaxBits),g),-1!=j&&(i=this.match_main(b,h.substring(h.length-this.Match_MaxBits),g+h.length-this.Match_MaxBits),-1==i||j>=i))j=-1}else j=this.match_main(b,h,g); if(-1==j)e[f]=!1,d-=a[f].length2-a[f].length1;else if(e[f]=!0,d=j-g,g=-1==i?b.substring(j,j+h.length):b.substring(j,i+this.Match_MaxBits),h==g)b=b.substring(0,j)+this.diff_text2(a[f].diffs)+b.substring(j+h.length);else if(g=this.diff_main(h,g,!1),h.length>this.Match_MaxBits&&this.diff_levenshtein(g)/h.length>this.Patch_DeleteThreshold)e[f]=!1;else{this.diff_cleanupSemanticLossless(g);for(var h=0,k,i=0;ie[0][1].length){var f=b-e[0][1].length;e[0][1]=c.substring(e[0][1].length)+e[0][1];d.start1-=f;d.start2-=f;d.length1+=f;d.length2+=f}d=a[a.length-1];e=d.diffs;0==e.length||0!=e[e.length-1][0]?(e.push([0, c]),d.length1+=b,d.length2+=b):b>e[e.length-1][1].length&&(f=b-e[e.length-1][1].length,e[e.length-1][1]+=c.substring(0,f),d.length1+=f,d.length2+=f);return c}; diff_match_patch.prototype.patch_splitMax=function(a){for(var b=this.Match_MaxBits,c=0;c2*b?(h.length1+=i.length,e+=i.length,j=!1,h.diffs.push([g,i]),d.diffs.shift()):(i=i.substring(0,b-h.length1-this.Patch_Margin),h.length1+=i.length,e+=i.length,0===g?(h.length2+=i.length,f+=i.length):j=!1,h.diffs.push([g,i]),i==d.diffs[0][1]?d.diffs.shift():d.diffs[0][1]=d.diffs[0][1].substring(i.length))}g=this.diff_text2(h.diffs);g=g.substring(g.length-this.Patch_Margin);i=this.diff_text1(d.diffs).substring(0,this.Patch_Margin);""!==i&& (h.length1+=i.length,h.length2+=i.length,0!==h.diffs.length&&0===h.diffs[h.diffs.length-1][0]?h.diffs[h.diffs.length-1][1]+=i:h.diffs.push([0,i]));j||a.splice(++c,0,h)}}};diff_match_patch.prototype.patch_toText=function(a){for(var b=[],c=0;cb'], [DIFF_INSERT, 'c&d']]; assertEquals('a¶
<B>b</B>c&d', dmp.diff_prettyHtml(diffs)); } function testDiffText() { // Compute the source and destination texts. var diffs = [[DIFF_EQUAL, 'jump'], [DIFF_DELETE, 's'], [DIFF_INSERT, 'ed'], [DIFF_EQUAL, ' over '], [DIFF_DELETE, 'the'], [DIFF_INSERT, 'a'], [DIFF_EQUAL, ' lazy']]; assertEquals('jumps over the lazy', dmp.diff_text1(diffs)); assertEquals('jumped over a lazy', dmp.diff_text2(diffs)); } function testDiffDelta() { // Convert a diff into delta string. var diffs = [[DIFF_EQUAL, 'jump'], [DIFF_DELETE, 's'], [DIFF_INSERT, 'ed'], [DIFF_EQUAL, ' over '], [DIFF_DELETE, 'the'], [DIFF_INSERT, 'a'], [DIFF_EQUAL, ' lazy'], [DIFF_INSERT, 'old dog']]; var text1 = dmp.diff_text1(diffs); assertEquals('jumps over the lazy', text1); var delta = dmp.diff_toDelta(diffs); assertEquals('=4\t-1\t+ed\t=6\t-3\t+a\t=5\t+old dog', delta); // Convert delta string into a diff. assertEquivalent(diffs, dmp.diff_fromDelta(text1, delta)); // Generates error (19 != 20). try { dmp.diff_fromDelta(text1 + 'x', delta); assertEquals(Error, null); } catch (e) { // Exception expected. } // Generates error (19 != 18). try { dmp.diff_fromDelta(text1.substring(1), delta); assertEquals(Error, null); } catch (e) { // Exception expected. } // Generates error (%c3%xy invalid Unicode). try { dmp.diff_fromDelta('', '+%c3%xy'); assertEquals(Error, null); } catch (e) { // Exception expected. } // Test deltas with special characters. diffs = [[DIFF_EQUAL, '\u0680 \x00 \t %'], [DIFF_DELETE, '\u0681 \x01 \n ^'], [DIFF_INSERT, '\u0682 \x02 \\ |']]; text1 = dmp.diff_text1(diffs); assertEquals('\u0680 \x00 \t %\u0681 \x01 \n ^', text1); delta = dmp.diff_toDelta(diffs); assertEquals('=7\t-7\t+%DA%82 %02 %5C %7C', delta); // Convert delta string into a diff. assertEquivalent(diffs, dmp.diff_fromDelta(text1, delta)); // Verify pool of unchanged characters. diffs = [[DIFF_INSERT, 'A-Z a-z 0-9 - _ . ! ~ * \' ( ) ; / ? : @ & = + $ , # ']]; var text2 = dmp.diff_text2(diffs); assertEquals('A-Z a-z 0-9 - _ . ! ~ * \' ( ) ; / ? : @ & = + $ , # ', text2); delta = dmp.diff_toDelta(diffs); assertEquals('+A-Z a-z 0-9 - _ . ! ~ * \' ( ) ; / ? : @ & = + $ , # ', delta); // Convert delta string into a diff. assertEquivalent(diffs, dmp.diff_fromDelta('', delta)); } function testDiffXIndex() { // Translate a location in text1 to text2. // Translation on equality. assertEquals(5, dmp.diff_xIndex([[DIFF_DELETE, 'a'], [DIFF_INSERT, '1234'], [DIFF_EQUAL, 'xyz']], 2)); // Translation on deletion. assertEquals(1, dmp.diff_xIndex([[DIFF_EQUAL, 'a'], [DIFF_DELETE, '1234'], [DIFF_EQUAL, 'xyz']], 3)); } function testDiffLevenshtein() { // Levenshtein with trailing equality. assertEquals(4, dmp.diff_levenshtein([[DIFF_DELETE, 'abc'], [DIFF_INSERT, '1234'], [DIFF_EQUAL, 'xyz']])); // Levenshtein with leading equality. assertEquals(4, dmp.diff_levenshtein([[DIFF_EQUAL, 'xyz'], [DIFF_DELETE, 'abc'], [DIFF_INSERT, '1234']])); // Levenshtein with middle equality. assertEquals(7, dmp.diff_levenshtein([[DIFF_DELETE, 'abc'], [DIFF_EQUAL, 'xyz'], [DIFF_INSERT, '1234']])); } function testDiffBisect() { // Normal. var a = 'cat'; var b = 'map'; // Since the resulting diff hasn't been normalized, it would be ok if // the insertion and deletion pairs are swapped. // If the order changes, tweak this test as required. assertEquivalent([[DIFF_DELETE, 'c'], [DIFF_INSERT, 'm'], [DIFF_EQUAL, 'a'], [DIFF_DELETE, 't'], [DIFF_INSERT, 'p']], dmp.diff_bisect_(a, b, Number.MAX_VALUE)); // Timeout. assertEquivalent([[DIFF_DELETE, 'cat'], [DIFF_INSERT, 'map']], dmp.diff_bisect_(a, b, 0)); } function testDiffMain() { // Perform a trivial diff. // Null case. assertEquivalent([], dmp.diff_main('', '', false)); // Equality. assertEquivalent([[DIFF_EQUAL, 'abc']], dmp.diff_main('abc', 'abc', false)); // Simple insertion. assertEquivalent([[DIFF_EQUAL, 'ab'], [DIFF_INSERT, '123'], [DIFF_EQUAL, 'c']], dmp.diff_main('abc', 'ab123c', false)); // Simple deletion. assertEquivalent([[DIFF_EQUAL, 'a'], [DIFF_DELETE, '123'], [DIFF_EQUAL, 'bc']], dmp.diff_main('a123bc', 'abc', false)); // Two insertions. assertEquivalent([[DIFF_EQUAL, 'a'], [DIFF_INSERT, '123'], [DIFF_EQUAL, 'b'], [DIFF_INSERT, '456'], [DIFF_EQUAL, 'c']], dmp.diff_main('abc', 'a123b456c', false)); // Two deletions. assertEquivalent([[DIFF_EQUAL, 'a'], [DIFF_DELETE, '123'], [DIFF_EQUAL, 'b'], [DIFF_DELETE, '456'], [DIFF_EQUAL, 'c']], dmp.diff_main('a123b456c', 'abc', false)); // Perform a real diff. // Switch off the timeout. dmp.Diff_Timeout = 0; // Simple cases. assertEquivalent([[DIFF_DELETE, 'a'], [DIFF_INSERT, 'b']], dmp.diff_main('a', 'b', false)); assertEquivalent([[DIFF_DELETE, 'Apple'], [DIFF_INSERT, 'Banana'], [DIFF_EQUAL, 's are a'], [DIFF_INSERT, 'lso'], [DIFF_EQUAL, ' fruit.']], dmp.diff_main('Apples are a fruit.', 'Bananas are also fruit.', false)); assertEquivalent([[DIFF_DELETE, 'a'], [DIFF_INSERT, '\u0680'], [DIFF_EQUAL, 'x'], [DIFF_DELETE, '\t'], [DIFF_INSERT, '\0']], dmp.diff_main('ax\t', '\u0680x\0', false)); // Overlaps. assertEquivalent([[DIFF_DELETE, '1'], [DIFF_EQUAL, 'a'], [DIFF_DELETE, 'y'], [DIFF_EQUAL, 'b'], [DIFF_DELETE, '2'], [DIFF_INSERT, 'xab']], dmp.diff_main('1ayb2', 'abxab', false)); assertEquivalent([[DIFF_INSERT, 'xaxcx'], [DIFF_EQUAL, 'abc'], [DIFF_DELETE, 'y']], dmp.diff_main('abcy', 'xaxcxabc', false)); assertEquivalent([[DIFF_DELETE, 'ABCD'], [DIFF_EQUAL, 'a'], [DIFF_DELETE, '='], [DIFF_INSERT, '-'], [DIFF_EQUAL, 'bcd'], [DIFF_DELETE, '='], [DIFF_INSERT, '-'], [DIFF_EQUAL, 'efghijklmnopqrs'], [DIFF_DELETE, 'EFGHIJKLMNOefg']], dmp.diff_main('ABCDa=bcd=efghijklmnopqrsEFGHIJKLMNOefg', 'a-bcd-efghijklmnopqrs', false)); // Large equality. assertEquivalent([[DIFF_INSERT, ' '], [DIFF_EQUAL, 'a'], [DIFF_INSERT, 'nd'], [DIFF_EQUAL, ' [[Pennsylvania]]'], [DIFF_DELETE, ' and [[New']], dmp.diff_main('a [[Pennsylvania]] and [[New', ' and [[Pennsylvania]]', false)); // Timeout. dmp.Diff_Timeout = 0.1; // 100ms var a = '`Twas brillig, and the slithy toves\nDid gyre and gimble in the wabe:\nAll mimsy were the borogoves,\nAnd the mome raths outgrabe.\n'; var b = 'I am the very model of a modern major general,\nI\'ve information vegetable, animal, and mineral,\nI know the kings of England, and I quote the fights historical,\nFrom Marathon to Waterloo, in order categorical.\n'; // Increase the text lengths by 1024 times to ensure a timeout. for (var x = 0; x < 10; x++) { a = a + a; b = b + b; } var startTime = (new Date()).getTime(); dmp.diff_main(a, b); var endTime = (new Date()).getTime(); // Test that we took at least the timeout period. assertTrue(dmp.Diff_Timeout * 1000 <= endTime - startTime); // Test that we didn't take forever (be forgiving). // Theoretically this test could fail very occasionally if the // OS task swaps or locks up for a second at the wrong moment. // **** // TODO(fraser): For unknown reasons this is taking 500 ms on Google's // internal test system. Whereas browsers take 140 ms. //assertTrue(dmp.Diff_Timeout * 1000 * 2 > endTime - startTime); // **** dmp.Diff_Timeout = 0; // Test the linemode speedup. // Must be long to pass the 100 char cutoff. // Simple line-mode. a = '1234567890\n1234567890\n1234567890\n1234567890\n1234567890\n1234567890\n1234567890\n1234567890\n1234567890\n1234567890\n1234567890\n1234567890\n1234567890\n'; b = 'abcdefghij\nabcdefghij\nabcdefghij\nabcdefghij\nabcdefghij\nabcdefghij\nabcdefghij\nabcdefghij\nabcdefghij\nabcdefghij\nabcdefghij\nabcdefghij\nabcdefghij\n'; assertEquivalent(dmp.diff_main(a, b, false), dmp.diff_main(a, b, true)); // Single line-mode. a = '1234567890123456789012345678901234567890123456789012345678901234567890123456789012345678901234567890123456789012345678901234567890'; b = 'abcdefghijabcdefghijabcdefghijabcdefghijabcdefghijabcdefghijabcdefghijabcdefghijabcdefghijabcdefghijabcdefghijabcdefghijabcdefghij'; assertEquivalent(dmp.diff_main(a, b, false), dmp.diff_main(a, b, true)); // Overlap line-mode. a = '1234567890\n1234567890\n1234567890\n1234567890\n1234567890\n1234567890\n1234567890\n1234567890\n1234567890\n1234567890\n1234567890\n1234567890\n1234567890\n'; b = 'abcdefghij\n1234567890\n1234567890\n1234567890\nabcdefghij\n1234567890\n1234567890\n1234567890\nabcdefghij\n1234567890\n1234567890\n1234567890\nabcdefghij\n'; var texts_linemode = diff_rebuildtexts(dmp.diff_main(a, b, true)); var texts_textmode = diff_rebuildtexts(dmp.diff_main(a, b, false)); assertEquivalent(texts_textmode, texts_linemode); // Test null inputs. try { dmp.diff_main(null, null); assertEquals(Error, null); } catch (e) { // Exception expected. } } // MATCH TEST FUNCTIONS function testMatchAlphabet() { // Initialise the bitmasks for Bitap. // Unique. assertEquivalent({'a':4, 'b':2, 'c':1}, dmp.match_alphabet_('abc')); // Duplicates. assertEquivalent({'a':37, 'b':18, 'c':8}, dmp.match_alphabet_('abcaba')); } function testMatchBitap() { // Bitap algorithm. dmp.Match_Distance = 100; dmp.Match_Threshold = 0.5; // Exact matches. assertEquals(5, dmp.match_bitap_('abcdefghijk', 'fgh', 5)); assertEquals(5, dmp.match_bitap_('abcdefghijk', 'fgh', 0)); // Fuzzy matches. assertEquals(4, dmp.match_bitap_('abcdefghijk', 'efxhi', 0)); assertEquals(2, dmp.match_bitap_('abcdefghijk', 'cdefxyhijk', 5)); assertEquals(-1, dmp.match_bitap_('abcdefghijk', 'bxy', 1)); // Overflow. assertEquals(2, dmp.match_bitap_('123456789xx0', '3456789x0', 2)); // Threshold test. dmp.Match_Threshold = 0.4; assertEquals(4, dmp.match_bitap_('abcdefghijk', 'efxyhi', 1)); dmp.Match_Threshold = 0.3; assertEquals(-1, dmp.match_bitap_('abcdefghijk', 'efxyhi', 1)); dmp.Match_Threshold = 0.0; assertEquals(1, dmp.match_bitap_('abcdefghijk', 'bcdef', 1)); dmp.Match_Threshold = 0.5; // Multiple select. assertEquals(0, dmp.match_bitap_('abcdexyzabcde', 'abccde', 3)); assertEquals(8, dmp.match_bitap_('abcdexyzabcde', 'abccde', 5)); // Distance test. dmp.Match_Distance = 10; // Strict location. assertEquals(-1, dmp.match_bitap_('abcdefghijklmnopqrstuvwxyz', 'abcdefg', 24)); assertEquals(0, dmp.match_bitap_('abcdefghijklmnopqrstuvwxyz', 'abcdxxefg', 1)); dmp.Match_Distance = 1000; // Loose location. assertEquals(0, dmp.match_bitap_('abcdefghijklmnopqrstuvwxyz', 'abcdefg', 24)); } function testMatchMain() { // Full match. // Shortcut matches. assertEquals(0, dmp.match_main('abcdef', 'abcdef', 1000)); assertEquals(-1, dmp.match_main('', 'abcdef', 1)); assertEquals(3, dmp.match_main('abcdef', '', 3)); assertEquals(3, dmp.match_main('abcdef', 'de', 3)); // Beyond end match. assertEquals(3, dmp.match_main("abcdef", "defy", 4)); // Oversized pattern. assertEquals(0, dmp.match_main("abcdef", "abcdefy", 0)); // Complex match. assertEquals(4, dmp.match_main('I am the very model of a modern major general.', ' that berry ', 5)); // Test null inputs. try { dmp.match_main(null, null, 0); assertEquals(Error, null); } catch (e) { // Exception expected. } } // PATCH TEST FUNCTIONS function testPatchObj() { // Patch Object. var p = new diff_match_patch.patch_obj(); p.start1 = 20; p.start2 = 21; p.length1 = 18; p.length2 = 17; p.diffs = [[DIFF_EQUAL, 'jump'], [DIFF_DELETE, 's'], [DIFF_INSERT, 'ed'], [DIFF_EQUAL, ' over '], [DIFF_DELETE, 'the'], [DIFF_INSERT, 'a'], [DIFF_EQUAL, '\nlaz']]; var strp = p.toString(); assertEquals('@@ -21,18 +22,17 @@\n jump\n-s\n+ed\n over \n-the\n+a\n %0Alaz\n', strp); } function testPatchFromText() { assertEquivalent([], dmp.patch_fromText(strp)); var strp = '@@ -21,18 +22,17 @@\n jump\n-s\n+ed\n over \n-the\n+a\n %0Alaz\n'; assertEquals(strp, dmp.patch_fromText(strp)[0].toString()); assertEquals('@@ -1 +1 @@\n-a\n+b\n', dmp.patch_fromText('@@ -1 +1 @@\n-a\n+b\n')[0].toString()); assertEquals('@@ -1,3 +0,0 @@\n-abc\n', dmp.patch_fromText('@@ -1,3 +0,0 @@\n-abc\n')[0].toString()); assertEquals('@@ -0,0 +1,3 @@\n+abc\n', dmp.patch_fromText('@@ -0,0 +1,3 @@\n+abc\n')[0].toString()); // Generates error. try { dmp.patch_fromText('Bad\nPatch\n'); assertEquals(Error, null); } catch (e) { // Exception expected. } } function testPatchToText() { var strp = '@@ -21,18 +22,17 @@\n jump\n-s\n+ed\n over \n-the\n+a\n laz\n'; var p = dmp.patch_fromText(strp); assertEquals(strp, dmp.patch_toText(p)); strp = '@@ -1,9 +1,9 @@\n-f\n+F\n oo+fooba\n@@ -7,9 +7,9 @@\n obar\n-,\n+.\n tes\n'; p = dmp.patch_fromText(strp); assertEquals(strp, dmp.patch_toText(p)); } function testPatchAddContext() { dmp.Patch_Margin = 4; var p = dmp.patch_fromText('@@ -21,4 +21,10 @@\n-jump\n+somersault\n')[0]; dmp.patch_addContext_(p, 'The quick brown fox jumps over the lazy dog.'); assertEquals('@@ -17,12 +17,18 @@\n fox \n-jump\n+somersault\n s ov\n', p.toString()); // Same, but not enough trailing context. p = dmp.patch_fromText('@@ -21,4 +21,10 @@\n-jump\n+somersault\n')[0]; dmp.patch_addContext_(p, 'The quick brown fox jumps.'); assertEquals('@@ -17,10 +17,16 @@\n fox \n-jump\n+somersault\n s.\n', p.toString()); // Same, but not enough leading context. p = dmp.patch_fromText('@@ -3 +3,2 @@\n-e\n+at\n')[0]; dmp.patch_addContext_(p, 'The quick brown fox jumps.'); assertEquals('@@ -1,7 +1,8 @@\n Th\n-e\n+at\n qui\n', p.toString()); // Same, but with ambiguity. p = dmp.patch_fromText('@@ -3 +3,2 @@\n-e\n+at\n')[0]; dmp.patch_addContext_(p, 'The quick brown fox jumps. The quick brown fox crashes.'); assertEquals('@@ -1,27 +1,28 @@\n Th\n-e\n+at\n quick brown fox jumps. \n', p.toString()); } function testPatchMake() { // Null case. var patches = dmp.patch_make('', ''); assertEquals('', dmp.patch_toText(patches)); var text1 = 'The quick brown fox jumps over the lazy dog.'; var text2 = 'That quick brown fox jumped over a lazy dog.'; // Text2+Text1 inputs. var expectedPatch = '@@ -1,8 +1,7 @@\n Th\n-at\n+e\n qui\n@@ -21,17 +21,18 @@\n jump\n-ed\n+s\n over \n-a\n+the\n laz\n'; // The second patch must be "-21,17 +21,18", not "-22,17 +21,18" due to rolling context. patches = dmp.patch_make(text2, text1); assertEquals(expectedPatch, dmp.patch_toText(patches)); // Text1+Text2 inputs. expectedPatch = '@@ -1,11 +1,12 @@\n Th\n-e\n+at\n quick b\n@@ -22,18 +22,17 @@\n jump\n-s\n+ed\n over \n-the\n+a\n laz\n'; patches = dmp.patch_make(text1, text2); assertEquals(expectedPatch, dmp.patch_toText(patches)); // Diff input. var diffs = dmp.diff_main(text1, text2, false); patches = dmp.patch_make(diffs); assertEquals(expectedPatch, dmp.patch_toText(patches)); // Text1+Diff inputs. patches = dmp.patch_make(text1, diffs); assertEquals(expectedPatch, dmp.patch_toText(patches)); // Text1+Text2+Diff inputs (deprecated). patches = dmp.patch_make(text1, text2, diffs); assertEquals(expectedPatch, dmp.patch_toText(patches)); // Character encoding. patches = dmp.patch_make('`1234567890-=[]\\;\',./', '~!@#$%^&*()_+{}|:"<>?'); assertEquals('@@ -1,21 +1,21 @@\n-%601234567890-=%5B%5D%5C;\',./\n+~!@#$%25%5E&*()_+%7B%7D%7C:%22%3C%3E?\n', dmp.patch_toText(patches)); // Character decoding. diffs = [[DIFF_DELETE, '`1234567890-=[]\\;\',./'], [DIFF_INSERT, '~!@#$%^&*()_+{}|:"<>?']]; assertEquivalent(diffs, dmp.patch_fromText('@@ -1,21 +1,21 @@\n-%601234567890-=%5B%5D%5C;\',./\n+~!@#$%25%5E&*()_+%7B%7D%7C:%22%3C%3E?\n')[0].diffs); // Long string with repeats. text1 = ''; for (var x = 0; x < 100; x++) { text1 += 'abcdef'; } text2 = text1 + '123'; expectedPatch = '@@ -573,28 +573,31 @@\n cdefabcdefabcdefabcdefabcdef\n+123\n'; patches = dmp.patch_make(text1, text2); assertEquals(expectedPatch, dmp.patch_toText(patches)); // Test null inputs. try { dmp.patch_make(null); assertEquals(Error, null); } catch (e) { // Exception expected. } } function testPatchSplitMax() { // Assumes that dmp.Match_MaxBits is 32. var patches = dmp.patch_make('abcdefghijklmnopqrstuvwxyz01234567890', 'XabXcdXefXghXijXklXmnXopXqrXstXuvXwxXyzX01X23X45X67X89X0'); dmp.patch_splitMax(patches); assertEquals('@@ -1,32 +1,46 @@\n+X\n ab\n+X\n cd\n+X\n ef\n+X\n gh\n+X\n ij\n+X\n kl\n+X\n mn\n+X\n op\n+X\n qr\n+X\n st\n+X\n uv\n+X\n wx\n+X\n yz\n+X\n 012345\n@@ -25,13 +39,18 @@\n zX01\n+X\n 23\n+X\n 45\n+X\n 67\n+X\n 89\n+X\n 0\n', dmp.patch_toText(patches)); patches = dmp.patch_make('abcdef1234567890123456789012345678901234567890123456789012345678901234567890uvwxyz', 'abcdefuvwxyz'); var oldToText = dmp.patch_toText(patches); dmp.patch_splitMax(patches); assertEquals(oldToText, dmp.patch_toText(patches)); patches = dmp.patch_make('1234567890123456789012345678901234567890123456789012345678901234567890', 'abc'); dmp.patch_splitMax(patches); assertEquals('@@ -1,32 +1,4 @@\n-1234567890123456789012345678\n 9012\n@@ -29,32 +1,4 @@\n-9012345678901234567890123456\n 7890\n@@ -57,14 +1,3 @@\n-78901234567890\n+abc\n', dmp.patch_toText(patches)); patches = dmp.patch_make('abcdefghij , h : 0 , t : 1 abcdefghij , h : 0 , t : 1 abcdefghij , h : 0 , t : 1', 'abcdefghij , h : 1 , t : 1 abcdefghij , h : 1 , t : 1 abcdefghij , h : 0 , t : 1'); dmp.patch_splitMax(patches); assertEquals('@@ -2,32 +2,32 @@\n bcdefghij , h : \n-0\n+1\n , t : 1 abcdef\n@@ -29,32 +29,32 @@\n bcdefghij , h : \n-0\n+1\n , t : 1 abcdef\n', dmp.patch_toText(patches)); } function testPatchAddPadding() { // Both edges full. var patches = dmp.patch_make('', 'test'); assertEquals('@@ -0,0 +1,4 @@\n+test\n', dmp.patch_toText(patches)); dmp.patch_addPadding(patches); assertEquals('@@ -1,8 +1,12 @@\n %01%02%03%04\n+test\n %01%02%03%04\n', dmp.patch_toText(patches)); // Both edges partial. patches = dmp.patch_make('XY', 'XtestY'); assertEquals('@@ -1,2 +1,6 @@\n X\n+test\n Y\n', dmp.patch_toText(patches)); dmp.patch_addPadding(patches); assertEquals('@@ -2,8 +2,12 @@\n %02%03%04X\n+test\n Y%01%02%03\n', dmp.patch_toText(patches)); // Both edges none. patches = dmp.patch_make('XXXXYYYY', 'XXXXtestYYYY'); assertEquals('@@ -1,8 +1,12 @@\n XXXX\n+test\n YYYY\n', dmp.patch_toText(patches)); dmp.patch_addPadding(patches); assertEquals('@@ -5,8 +5,12 @@\n XXXX\n+test\n YYYY\n', dmp.patch_toText(patches)); } function testPatchApply() { dmp.Match_Distance = 1000; dmp.Match_Threshold = 0.5; dmp.Patch_DeleteThreshold = 0.5; // Null case. var patches = dmp.patch_make('', ''); var results = dmp.patch_apply(patches, 'Hello world.'); assertEquivalent(['Hello world.', []], results); // Exact match. patches = dmp.patch_make('The quick brown fox jumps over the lazy dog.', 'That quick brown fox jumped over a lazy dog.'); results = dmp.patch_apply(patches, 'The quick brown fox jumps over the lazy dog.'); assertEquivalent(['That quick brown fox jumped over a lazy dog.', [true, true]], results); // Partial match. results = dmp.patch_apply(patches, 'The quick red rabbit jumps over the tired tiger.'); assertEquivalent(['That quick red rabbit jumped over a tired tiger.', [true, true]], results); // Failed match. results = dmp.patch_apply(patches, 'I am the very model of a modern major general.'); assertEquivalent(['I am the very model of a modern major general.', [false, false]], results); // Big delete, small change. patches = dmp.patch_make('x1234567890123456789012345678901234567890123456789012345678901234567890y', 'xabcy'); results = dmp.patch_apply(patches, 'x123456789012345678901234567890-----++++++++++-----123456789012345678901234567890y'); assertEquivalent(['xabcy', [true, true]], results); // Big delete, big change 1. patches = dmp.patch_make('x1234567890123456789012345678901234567890123456789012345678901234567890y', 'xabcy'); results = dmp.patch_apply(patches, 'x12345678901234567890---------------++++++++++---------------12345678901234567890y'); assertEquivalent(['xabc12345678901234567890---------------++++++++++---------------12345678901234567890y', [false, true]], results); // Big delete, big change 2. dmp.Patch_DeleteThreshold = 0.6; patches = dmp.patch_make('x1234567890123456789012345678901234567890123456789012345678901234567890y', 'xabcy'); results = dmp.patch_apply(patches, 'x12345678901234567890---------------++++++++++---------------12345678901234567890y'); assertEquivalent(['xabcy', [true, true]], results); dmp.Patch_DeleteThreshold = 0.5; // Compensate for failed patch. dmp.Match_Threshold = 0.0; dmp.Match_Distance = 0; patches = dmp.patch_make('abcdefghijklmnopqrstuvwxyz--------------------1234567890', 'abcXXXXXXXXXXdefghijklmnopqrstuvwxyz--------------------1234567YYYYYYYYYY890'); results = dmp.patch_apply(patches, 'ABCDEFGHIJKLMNOPQRSTUVWXYZ--------------------1234567890'); assertEquivalent(['ABCDEFGHIJKLMNOPQRSTUVWXYZ--------------------1234567YYYYYYYYYY890', [false, true]], results); dmp.Match_Threshold = 0.5; dmp.Match_Distance = 1000; // No side effects. patches = dmp.patch_make('', 'test'); var patchstr = dmp.patch_toText(patches); dmp.patch_apply(patches, ''); assertEquals(patchstr, dmp.patch_toText(patches)); // No side effects with major delete. patches = dmp.patch_make('The quick brown fox jumps over the lazy dog.', 'Woof'); patchstr = dmp.patch_toText(patches); dmp.patch_apply(patches, 'The quick brown fox jumps over the lazy dog.'); assertEquals(patchstr, dmp.patch_toText(patches)); // Edge exact match. patches = dmp.patch_make('', 'test'); results = dmp.patch_apply(patches, ''); assertEquivalent(['test', [true]], results); // Near edge exact match. patches = dmp.patch_make('XY', 'XtestY'); results = dmp.patch_apply(patches, 'XY'); assertEquivalent(['XtestY', [true]], results); // Edge partial match. patches = dmp.patch_make('y', 'y123'); results = dmp.patch_apply(patches, 'x'); assertEquivalent(['x123', [true]], results); } package/javascript/diff_match_patch_uncompressed.js000644 001751 000062 0000225315 12066005331021267 0ustar00000000 000000 /** * Diff Match and Patch * * Copyright 2006 Google Inc. * http://code.google.com/p/google-diff-match-patch/ * * Licensed under the Apache License, Version 2.0 (the "License"); * you may not use this file except in compliance with the License. * You may obtain a copy of the License at * * http://www.apache.org/licenses/LICENSE-2.0 * * Unless required by applicable law or agreed to in writing, software * distributed under the License is distributed on an "AS IS" BASIS, * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. * See the License for the specific language governing permissions and * limitations under the License. */ /** * @fileoverview Computes the difference between two texts to create a patch. * Applies the patch onto another text, allowing for errors. * @author fraser@google.com (Neil Fraser) */ /** * Class containing the diff, match and patch methods. * @constructor */ function diff_match_patch() { // Defaults. // Redefine these in your program to override the defaults. // Number of seconds to map a diff before giving up (0 for infinity). this.Diff_Timeout = 1.0; // Cost of an empty edit operation in terms of edit characters. this.Diff_EditCost = 4; // At what point is no match declared (0.0 = perfection, 1.0 = very loose). this.Match_Threshold = 0.5; // How far to search for a match (0 = exact location, 1000+ = broad match). // A match this many characters away from the expected location will add // 1.0 to the score (0.0 is a perfect match). this.Match_Distance = 1000; // When deleting a large block of text (over ~64 characters), how close do // the contents have to be to match the expected contents. (0.0 = perfection, // 1.0 = very loose). Note that Match_Threshold controls how closely the // end points of a delete need to match. this.Patch_DeleteThreshold = 0.5; // Chunk size for context length. this.Patch_Margin = 4; // The number of bits in an int. this.Match_MaxBits = 32; } // DIFF FUNCTIONS /** * The data structure representing a diff is an array of tuples: * [[DIFF_DELETE, 'Hello'], [DIFF_INSERT, 'Goodbye'], [DIFF_EQUAL, ' world.']] * which means: delete 'Hello', add 'Goodbye' and keep ' world.' */ var DIFF_DELETE = -1; var DIFF_INSERT = 1; var DIFF_EQUAL = 0; /** @typedef {{0: number, 1: string}} */ diff_match_patch.Diff; /** * 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 {boolean=} opt_checklines Optional speedup flag. If present and false, * then don't run a line-level diff first to identify the changed areas. * Defaults to true, which does a faster, slightly less optimal diff. * @param {number} opt_deadline Optional time when the diff should be complete * by. Used internally for recursive calls. Users should set DiffTimeout * instead. * @return {!Array.} Array of diff tuples. */ diff_match_patch.prototype.diff_main = function(text1, text2, opt_checklines, opt_deadline) { // Set a deadline by which time the diff must be complete. if (typeof opt_deadline == 'undefined') { if (this.Diff_Timeout <= 0) { opt_deadline = Number.MAX_VALUE; } else { opt_deadline = (new Date).getTime() + this.Diff_Timeout * 1000; } } var deadline = opt_deadline; // Check for null inputs. if (text1 == null || text2 == null) { throw new Error('Null input. (diff_main)'); } // Check for equality (speedup). if (text1 == text2) { if (text1) { return [[DIFF_EQUAL, text1]]; } return []; } if (typeof opt_checklines == 'undefined') { opt_checklines = true; } var checklines = opt_checklines; // Trim off common prefix (speedup). var commonlength = this.diff_commonPrefix(text1, text2); var commonprefix = text1.substring(0, commonlength); text1 = text1.substring(commonlength); text2 = text2.substring(commonlength); // Trim off common suffix (speedup). commonlength = this.diff_commonSuffix(text1, text2); var commonsuffix = text1.substring(text1.length - commonlength); text1 = text1.substring(0, text1.length - commonlength); text2 = text2.substring(0, text2.length - commonlength); // Compute the diff on the middle block. var diffs = this.diff_compute_(text1, text2, checklines, deadline); // Restore the prefix and suffix. if (commonprefix) { diffs.unshift([DIFF_EQUAL, commonprefix]); } if (commonsuffix) { diffs.push([DIFF_EQUAL, commonsuffix]); } this.diff_cleanupMerge(diffs); return diffs; }; /** * Find the differences between two texts. Assumes that the texts do not * have any common prefix or suffix. * @param {string} text1 Old string to be diffed. * @param {string} text2 New string to be diffed. * @param {boolean} checklines Speedup flag. If false, then don't run a * line-level diff first to identify the changed areas. * If true, then run a faster, slightly less optimal diff. * @param {number} deadline Time when the diff should be complete by. * @return {!Array.} Array of diff tuples. * @private */ diff_match_patch.prototype.diff_compute_ = function(text1, text2, checklines, deadline) { var diffs; if (!text1) { // Just add some text (speedup). return [[DIFF_INSERT, text2]]; } if (!text2) { // Just delete some text (speedup). return [[DIFF_DELETE, text1]]; } var longtext = text1.length > text2.length ? text1 : text2; var shorttext = text1.length > text2.length ? text2 : text1; var i = longtext.indexOf(shorttext); if (i != -1) { // Shorter text is inside the longer text (speedup). diffs = [[DIFF_INSERT, longtext.substring(0, i)], [DIFF_EQUAL, shorttext], [DIFF_INSERT, longtext.substring(i + shorttext.length)]]; // Swap insertions for deletions if diff is reversed. if (text1.length > text2.length) { diffs[0][0] = diffs[2][0] = DIFF_DELETE; } return diffs; } 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]]; } // Check to see if the problem can be split in two. var hm = this.diff_halfMatch_(text1, text2); if (hm) { // A half-match was found, sort out the return data. var text1_a = hm[0]; var text1_b = hm[1]; var text2_a = hm[2]; var text2_b = hm[3]; var mid_common = hm[4]; // Send both pairs off for separate processing. var diffs_a = this.diff_main(text1_a, text2_a, checklines, deadline); var diffs_b = this.diff_main(text1_b, text2_b, checklines, deadline); // Merge the results. return diffs_a.concat([[DIFF_EQUAL, mid_common]], diffs_b); } if (checklines && text1.length > 100 && text2.length > 100) { return this.diff_lineMode_(text1, text2, deadline); } return this.diff_bisect_(text1, text2, deadline); }; /** * Do a quick line-level diff on both strings, then rediff the parts for * greater accuracy. * This speedup can produce non-minimal diffs. * @param {string} text1 Old string to be diffed. * @param {string} text2 New string to be diffed. * @param {number} deadline Time when the diff should be complete by. * @return {!Array.} Array of diff tuples. * @private */ diff_match_patch.prototype.diff_lineMode_ = function(text1, text2, deadline) { // Scan the text on a line-by-line basis first. var a = this.diff_linesToChars_(text1, text2); text1 = a.chars1; text2 = a.chars2; var linearray = a.lineArray; var diffs = this.diff_main(text1, text2, false, deadline); // Convert the diff back to original text. this.diff_charsToLines_(diffs, linearray); // Eliminate freak matches (e.g. blank lines) this.diff_cleanupSemantic(diffs); // Rediff any replacement blocks, this time character-by-character. // Add a dummy entry at the end. diffs.push([DIFF_EQUAL, '']); var pointer = 0; var count_delete = 0; var count_insert = 0; var text_delete = ''; var text_insert = ''; while (pointer < diffs.length) { switch (diffs[pointer][0]) { case DIFF_INSERT: count_insert++; text_insert += diffs[pointer][1]; break; case DIFF_DELETE: count_delete++; text_delete += diffs[pointer][1]; break; case DIFF_EQUAL: // Upon reaching an equality, check for prior redundancies. if (count_delete >= 1 && count_insert >= 1) { // Delete the offending records and add the merged ones. diffs.splice(pointer - count_delete - count_insert, count_delete + count_insert); pointer = pointer - count_delete - count_insert; var a = this.diff_main(text_delete, text_insert, false, deadline); for (var j = a.length - 1; j >= 0; j--) { diffs.splice(pointer, 0, a[j]); } pointer = pointer + a.length; } count_insert = 0; count_delete = 0; text_delete = ''; text_insert = ''; break; } pointer++; } diffs.pop(); // Remove the dummy entry at the end. return diffs; }; /** * Find the 'middle snake' of a diff, split the problem in two * and return the recursively constructed diff. * See Myers 1986 paper: An O(ND) Difference Algorithm and Its Variations. * @param {string} text1 Old string to be diffed. * @param {string} text2 New string to be diffed. * @param {number} deadline Time at which to bail if not yet complete. * @return {!Array.} Array of diff tuples. * @private */ diff_match_patch.prototype.diff_bisect_ = function(text1, text2, deadline) { // Cache the text lengths to prevent multiple calls. var text1_length = text1.length; var text2_length = text2.length; var max_d = Math.ceil((text1_length + text2_length) / 2); var v_offset = max_d; var v_length = 2 * max_d; var v1 = new Array(v_length); var v2 = new Array(v_length); // Setting all elements to -1 is faster in Chrome & Firefox than mixing // integers and undefined. for (var x = 0; x < v_length; x++) { v1[x] = -1; v2[x] = -1; } v1[v_offset + 1] = 0; v2[v_offset + 1] = 0; 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); // Offsets for start and end of k loop. // Prevents mapping of space beyond the grid. var k1start = 0; var k1end = 0; var k2start = 0; var k2end = 0; for (var d = 0; d < max_d; d++) { // Bail out if deadline is reached. if ((new Date()).getTime() > deadline) { break; } // Walk the front path one step. for (var k1 = -d + k1start; k1 <= d - k1end; k1 += 2) { var k1_offset = v_offset + k1; var x1; if (k1 == -d || (k1 != d && v1[k1_offset - 1] < v1[k1_offset + 1])) { x1 = v1[k1_offset + 1]; } else { x1 = v1[k1_offset - 1] + 1; } var y1 = x1 - k1; while (x1 < text1_length && y1 < text2_length && text1.charAt(x1) == text2.charAt(y1)) { x1++; y1++; } v1[k1_offset] = x1; if (x1 > text1_length) { // Ran off the right of the graph. k1end += 2; } else if (y1 > text2_length) { // Ran off the bottom of the graph. k1start += 2; } else if (front) { var k2_offset = v_offset + delta - k1; if (k2_offset >= 0 && k2_offset < v_length && v2[k2_offset] != -1) { // Mirror x2 onto top-left coordinate system. var x2 = text1_length - v2[k2_offset]; if (x1 >= x2) { // Overlap detected. return this.diff_bisectSplit_(text1, text2, x1, y1, deadline); } } } } // Walk the reverse path one step. for (var k2 = -d + k2start; k2 <= d - k2end; k2 += 2) { var k2_offset = v_offset + k2; var x2; if (k2 == -d || (k2 != d && v2[k2_offset - 1] < v2[k2_offset + 1])) { x2 = v2[k2_offset + 1]; } else { x2 = v2[k2_offset - 1] + 1; } var y2 = x2 - k2; while (x2 < text1_length && y2 < text2_length && text1.charAt(text1_length - x2 - 1) == text2.charAt(text2_length - y2 - 1)) { x2++; y2++; } v2[k2_offset] = x2; if (x2 > text1_length) { // Ran off the left of the graph. k2end += 2; } else if (y2 > text2_length) { // Ran off the top of the graph. k2start += 2; } else if (!front) { var k1_offset = v_offset + delta - k2; if (k1_offset >= 0 && k1_offset < v_length && v1[k1_offset] != -1) { var x1 = v1[k1_offset]; var y1 = v_offset + x1 - k1_offset; // Mirror x2 onto top-left coordinate system. x2 = text1_length - x2; if (x1 >= x2) { // Overlap detected. return this.diff_bisectSplit_(text1, text2, x1, y1, deadline); } } } } } // 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]]; }; /** * Given the location of the 'middle snake', split the diff in two parts * and recurse. * @param {string} text1 Old string to be diffed. * @param {string} text2 New string to be diffed. * @param {number} x Index of split point in text1. * @param {number} y Index of split point in text2. * @param {number} deadline Time at which to bail if not yet complete. * @return {!Array.} Array of diff tuples. * @private */ diff_match_patch.prototype.diff_bisectSplit_ = function(text1, text2, x, y, deadline) { var text1a = text1.substring(0, x); var text2a = text2.substring(0, y); var text1b = text1.substring(x); var text2b = text2.substring(y); // Compute both diffs serially. var diffs = this.diff_main(text1a, text2a, false, deadline); var diffsb = this.diff_main(text1b, text2b, false, deadline); return diffs.concat(diffsb); }; /** * Split two texts into an array of strings. Reduce the texts to a string of * hashes where each Unicode character represents one line. * @param {string} text1 First string. * @param {string} text2 Second string. * @return {{chars1: string, chars2: string, lineArray: !Array.}} * An object containing the encoded text1, the encoded text2 and * the array of unique strings. * The zeroth element of the array of unique strings is intentionally blank. * @private */ diff_match_patch.prototype.diff_linesToChars_ = function(text1, text2) { var lineArray = []; // e.g. lineArray[4] == 'Hello\n' var lineHash = {}; // e.g. lineHash['Hello\n'] == 4 // '\x00' is a valid character, but various debuggers don't like it. // So we'll insert a junk entry to avoid generating a null character. lineArray[0] = ''; /** * Split a text into an array of strings. Reduce the texts to a string of * hashes where each Unicode character represents one line. * Modifies linearray and linehash through being a closure. * @param {string} text String to encode. * @return {string} Encoded string. * @private */ function diff_linesToCharsMunge_(text) { var chars = ''; // Walk the text, pulling out a substring for each line. // text.split('\n') would would temporarily double our memory footprint. // Modifying text would create many large strings to garbage collect. var lineStart = 0; var lineEnd = -1; // Keeping our own length variable is faster than looking it up. var lineArrayLength = lineArray.length; while (lineEnd < text.length - 1) { lineEnd = text.indexOf('\n', lineStart); if (lineEnd == -1) { lineEnd = text.length - 1; } var line = text.substring(lineStart, lineEnd + 1); lineStart = lineEnd + 1; if (lineHash.hasOwnProperty ? lineHash.hasOwnProperty(line) : (lineHash[line] !== undefined)) { chars += String.fromCharCode(lineHash[line]); } else { chars += String.fromCharCode(lineArrayLength); lineHash[line] = lineArrayLength; lineArray[lineArrayLength++] = line; } } return chars; } var chars1 = diff_linesToCharsMunge_(text1); var chars2 = diff_linesToCharsMunge_(text2); return {chars1: chars1, chars2: chars2, lineArray: lineArray}; }; /** * Rehydrate the text in a diff from a string of line hashes to real lines of * text. * @param {!Array.} diffs Array of diff tuples. * @param {!Array.} lineArray Array of unique strings. * @private */ diff_match_patch.prototype.diff_charsToLines_ = function(diffs, lineArray) { for (var x = 0; x < diffs.length; x++) { var chars = diffs[x][1]; var text = []; for (var y = 0; y < chars.length; y++) { text[y] = lineArray[chars.charCodeAt(y)]; } diffs[x][1] = text.join(''); } }; /** * Determine the common prefix of two strings. * @param {string} text1 First string. * @param {string} text2 Second string. * @return {number} The number of characters common to the start of each * string. */ diff_match_patch.prototype.diff_commonPrefix = function(text1, text2) { // Quick check for common null cases. if (!text1 || !text2 || text1.charAt(0) != text2.charAt(0)) { return 0; } // Binary search. // Performance analysis: http://neil.fraser.name/news/2007/10/09/ var pointermin = 0; var pointermax = Math.min(text1.length, text2.length); var pointermid = pointermax; var pointerstart = 0; while (pointermin < pointermid) { if (text1.substring(pointerstart, pointermid) == text2.substring(pointerstart, pointermid)) { pointermin = pointermid; pointerstart = pointermin; } else { pointermax = pointermid; } pointermid = Math.floor((pointermax - pointermin) / 2 + pointermin); } return pointermid; }; /** * Determine the common suffix of two strings. * @param {string} text1 First string. * @param {string} text2 Second string. * @return {number} The number of characters common to the end of each string. */ diff_match_patch.prototype.diff_commonSuffix = function(text1, text2) { // Quick check for common null cases. if (!text1 || !text2 || text1.charAt(text1.length - 1) != text2.charAt(text2.length - 1)) { return 0; } // Binary search. // Performance analysis: http://neil.fraser.name/news/2007/10/09/ var pointermin = 0; var pointermax = Math.min(text1.length, text2.length); var pointermid = pointermax; var pointerend = 0; while (pointermin < pointermid) { if (text1.substring(text1.length - pointermid, text1.length - pointerend) == text2.substring(text2.length - pointermid, text2.length - pointerend)) { pointermin = pointermid; pointerend = pointermin; } else { pointermax = pointermid; } pointermid = Math.floor((pointermax - pointermin) / 2 + pointermin); } 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 */ diff_match_patch.prototype.diff_commonOverlap_ = function(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++; } } }; /** * Do the two texts share a substring which is at least half the length of the * longer text? * This speedup can produce non-minimal diffs. * @param {string} text1 First string. * @param {string} text2 Second string. * @return {Array.} Five element Array, containing the prefix of * text1, the suffix of text1, the prefix of text2, the suffix of * text2 and the common middle. Or null if there was no match. * @private */ diff_match_patch.prototype.diff_halfMatch_ = function(text1, text2) { if (this.Diff_Timeout <= 0) { // Don't risk returning a non-optimal diff if we have unlimited time. return null; } 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. } var dmp = this; // 'this' becomes 'window' in a closure. /** * Does a substring of shorttext exist within longtext such that the substring * is at least half the length of longtext? * Closure, but does not reference any external variables. * @param {string} longtext Longer string. * @param {string} shorttext Shorter string. * @param {number} i Start index of quarter length substring within longtext. * @return {Array.} Five element Array, containing the prefix of * longtext, the suffix of longtext, the prefix of shorttext, the suffix * of shorttext and the common middle. Or null if there was no match. * @private */ function diff_halfMatchI_(longtext, shorttext, i) { // 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_longtext_a, best_longtext_b, best_shorttext_a, best_shorttext_b; while ((j = shorttext.indexOf(seed, j + 1)) != -1) { var prefixLength = dmp.diff_commonPrefix(longtext.substring(i), shorttext.substring(j)); var suffixLength = dmp.diff_commonSuffix(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_longtext_a = longtext.substring(0, i - suffixLength); best_longtext_b = longtext.substring(i + prefixLength); best_shorttext_a = shorttext.substring(0, j - suffixLength); best_shorttext_b = shorttext.substring(j + prefixLength); } } if (best_common.length * 2 >= longtext.length) { return [best_longtext_a, best_longtext_b, best_shorttext_a, best_shorttext_b, best_common]; } else { return null; } } // First check if the second quarter is the seed for a half-match. 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 hm; if (!hm1 && !hm2) { return null; } else if (!hm2) { hm = hm1; } else if (!hm1) { hm = hm2; } else { // Both matched. Select the longest. hm = hm1[4].length > hm2[4].length ? hm1 : hm2; } // A half-match was found, sort out the return data. var text1_a, text1_b, text2_a, text2_b; if (text1.length > text2.length) { text1_a = hm[0]; text1_b = hm[1]; text2_a = hm[2]; text2_b = hm[3]; } else { text2_a = hm[0]; text2_b = hm[1]; text1_a = hm[2]; text1_b = hm[3]; } 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.} diffs Array of diff tuples. */ diff_match_patch.prototype.diff_cleanupSemantic = function(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) { this.diff_cleanupMerge(diffs); } this.diff_cleanupSemanticLossless(diffs); // Find any overlaps between deletions and insertions. // e.g: abcxxxxxxdef // -> abcxxxdef // e.g: xxxabcdefxxx // -> defxxxabc // 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 = this.diff_commonOverlap_(deletion, insertion); var overlap_length2 = this.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++; } }; /** * 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 cat came. -> The cat came. * @param {!Array.} diffs Array of diff tuples. */ diff_match_patch.prototype.diff_cleanupSemanticLossless = function(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(diff_match_patch.nonAlphaNumericRegex_); var nonAlphaNumeric2 = char2.match(diff_match_patch.nonAlphaNumericRegex_); var whitespace1 = nonAlphaNumeric1 && char1.match(diff_match_patch.whitespaceRegex_); var whitespace2 = nonAlphaNumeric2 && char2.match(diff_match_patch.whitespaceRegex_); var lineBreak1 = whitespace1 && char1.match(diff_match_patch.linebreakRegex_); var lineBreak2 = whitespace2 && char2.match(diff_match_patch.linebreakRegex_); var blankLine1 = lineBreak1 && one.match(diff_match_patch.blanklineEndRegex_); var blankLine2 = lineBreak2 && two.match(diff_match_patch.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 = this.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++; } }; // Define some regex patterns for matching boundaries. diff_match_patch.nonAlphaNumericRegex_ = /[^a-zA-Z0-9]/; diff_match_patch.whitespaceRegex_ = /\s/; diff_match_patch.linebreakRegex_ = /[\r\n]/; diff_match_patch.blanklineEndRegex_ = /\n\r?\n$/; diff_match_patch.blanklineStartRegex_ = /^\r?\n\r?\n/; /** * Reduce the number of edits by eliminating operationally trivial equalities. * @param {!Array.} diffs Array of diff tuples. */ diff_match_patch.prototype.diff_cleanupEfficiency = function(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. // Is there an insertion operation before the last equality. var pre_ins = false; // Is there a deletion operation before the last equality. var pre_del = false; // Is there an insertion operation after the last equality. var post_ins = false; // Is there a deletion operation after the last equality. var post_del = false; while (pointer < diffs.length) { if (diffs[pointer][0] == DIFF_EQUAL) { // Equality found. if (diffs[pointer][1].length < this.Diff_EditCost && (post_ins || post_del)) { // Candidate found. equalities[equalitiesLength++] = pointer; pre_ins = post_ins; pre_del = post_del; lastequality = diffs[pointer][1]; } else { // Not a candidate, and can never become one. equalitiesLength = 0; lastequality = null; } post_ins = post_del = false; } else { // An insertion or deletion. if (diffs[pointer][0] == DIFF_DELETE) { post_del = true; } else { post_ins = true; } /* * Five types to be split: * ABXYCD * AXCD * ABXC * AXCD * ABXC */ if (lastequality && ((pre_ins && pre_del && post_ins && post_del) || ((lastequality.length < this.Diff_EditCost / 2) && (pre_ins + pre_del + post_ins + post_del) == 3))) { // Duplicate record. diffs.splice(equalities[equalitiesLength - 1], 0, [DIFF_DELETE, lastequality]); // Change second copy to insert. diffs[equalities[equalitiesLength - 1] + 1][0] = DIFF_INSERT; equalitiesLength--; // Throw away the equality we just deleted; lastequality = null; if (pre_ins && pre_del) { // No changes made which could affect previous entry, keep going. post_ins = post_del = true; equalitiesLength = 0; } else { equalitiesLength--; // Throw away the previous equality. pointer = equalitiesLength > 0 ? equalities[equalitiesLength - 1] : -1; post_ins = post_del = false; } changes = true; } } pointer++; } if (changes) { this.diff_cleanupMerge(diffs); } }; /** * Reorder and merge like edit sections. Merge equalities. * Any edit section can move as long as it doesn't cross an equality. * @param {!Array.} diffs Array of diff tuples. */ diff_match_patch.prototype.diff_cleanupMerge = function(diffs) { 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 commonlength; while (pointer < diffs.length) { switch (diffs[pointer][0]) { case DIFF_INSERT: count_insert++; text_insert += diffs[pointer][1]; pointer++; break; case DIFF_DELETE: count_delete++; text_delete += diffs[pointer][1]; pointer++; break; case DIFF_EQUAL: // Upon reaching an equality, check for prior redundancies. if (count_delete + count_insert > 1) { if (count_delete !== 0 && count_insert !== 0) { // Factor out any common prefixies. commonlength = this.diff_commonPrefix(text_insert, text_delete); if (commonlength !== 0) { if ((pointer - count_delete - count_insert) > 0 && diffs[pointer - count_delete - count_insert - 1][0] == DIFF_EQUAL) { diffs[pointer - count_delete - count_insert - 1][1] += text_insert.substring(0, commonlength); } else { diffs.splice(0, 0, [DIFF_EQUAL, text_insert.substring(0, commonlength)]); pointer++; } text_insert = text_insert.substring(commonlength); text_delete = text_delete.substring(commonlength); } // Factor out any common suffixies. commonlength = this.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); } } // Delete the offending records and add the merged ones. if (count_delete === 0) { diffs.splice(pointer - count_insert, count_delete + count_insert, [DIFF_INSERT, text_insert]); } else if (count_insert === 0) { diffs.splice(pointer - count_delete, count_delete + count_insert, [DIFF_DELETE, text_delete]); } else { diffs.splice(pointer - count_delete - count_insert, count_delete + count_insert, [DIFF_DELETE, text_delete], [DIFF_INSERT, text_insert]); } pointer = pointer - count_delete - count_insert + (count_delete ? 1 : 0) + (count_insert ? 1 : 0) + 1; } else if (pointer !== 0 && diffs[pointer - 1][0] == DIFF_EQUAL) { // Merge this equality with the previous one. diffs[pointer - 1][1] += diffs[pointer][1]; diffs.splice(pointer, 1); } else { pointer++; } count_insert = 0; count_delete = 0; text_delete = ''; text_insert = ''; break; } } 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 // which can be shifted sideways to eliminate an equality. // e.g: ABAC -> ABAC var changes = false; 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. 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][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]) { // Shift the edit over the next equality. diffs[pointer - 1][1] += diffs[pointer + 1][1]; diffs[pointer][1] = diffs[pointer][1].substring(diffs[pointer + 1][1].length) + diffs[pointer + 1][1]; diffs.splice(pointer + 1, 1); changes = true; } } pointer++; } // If shifts were made, the diff needs reordering and another shift sweep. if (changes) { this.diff_cleanupMerge(diffs); } }; /** * loc is a location in text1, compute and return the equivalent location in * text2. * e.g. 'The cat' vs 'The big cat', 1->1, 5->8 * @param {!Array.} diffs Array of diff tuples. * @param {number} loc Location within text1. * @return {number} Location within text2. */ diff_match_patch.prototype.diff_xIndex = function(diffs, loc) { var chars1 = 0; var chars2 = 0; var last_chars1 = 0; var last_chars2 = 0; var x; for (x = 0; x < diffs.length; x++) { if (diffs[x][0] !== DIFF_INSERT) { // Equality or deletion. chars1 += diffs[x][1].length; } if (diffs[x][0] !== DIFF_DELETE) { // Equality or insertion. chars2 += diffs[x][1].length; } if (chars1 > loc) { // Overshot the location. break; } last_chars1 = chars1; last_chars2 = chars2; } // Was the location was deleted? if (diffs.length != x && diffs[x][0] === DIFF_DELETE) { return last_chars2; } // Add the remaining character length. return last_chars2 + (loc - last_chars1); }; /** * Convert a diff array into a pretty HTML report. * @param {!Array.} diffs Array of diff tuples. * @return {string} HTML representation. */ diff_match_patch.prototype.diff_prettyHtml = function(diffs) { var html = []; var pattern_amp = /&/g; var pattern_lt = //g; var pattern_para = /\n/g; for (var x = 0; x < diffs.length; x++) { var op = diffs[x][0]; // Operation (insert, delete, equal) var data = diffs[x][1]; // Text of change. var text = data.replace(pattern_amp, '&').replace(pattern_lt, '<') .replace(pattern_gt, '>').replace(pattern_para, '¶
'); switch (op) { case DIFF_INSERT: html[x] = '' + text + ''; break; case DIFF_DELETE: html[x] = '' + text + ''; break; case DIFF_EQUAL: html[x] = '' + text + ''; break; } } return html.join(''); }; /** * Compute and return the source text (all equalities and deletions). * @param {!Array.} diffs Array of diff tuples. * @return {string} Source text. */ diff_match_patch.prototype.diff_text1 = function(diffs) { var text = []; for (var x = 0; x < diffs.length; x++) { if (diffs[x][0] !== DIFF_INSERT) { text[x] = diffs[x][1]; } } return text.join(''); }; /** * Compute and return the destination text (all equalities and insertions). * @param {!Array.} diffs Array of diff tuples. * @return {string} Destination text. */ diff_match_patch.prototype.diff_text2 = function(diffs) { var text = []; for (var x = 0; x < diffs.length; x++) { if (diffs[x][0] !== DIFF_DELETE) { text[x] = diffs[x][1]; } } return text.join(''); }; /** * Compute the Levenshtein distance; the number of inserted, deleted or * substituted characters. * @param {!Array.} diffs Array of diff tuples. * @return {number} Number of changes. */ diff_match_patch.prototype.diff_levenshtein = function(diffs) { var levenshtein = 0; var insertions = 0; var deletions = 0; for (var x = 0; x < diffs.length; x++) { var op = diffs[x][0]; var data = diffs[x][1]; switch (op) { case DIFF_INSERT: insertions += data.length; break; case DIFF_DELETE: deletions += data.length; break; case DIFF_EQUAL: // A deletion and an insertion is one substitution. levenshtein += Math.max(insertions, deletions); insertions = 0; deletions = 0; break; } } levenshtein += Math.max(insertions, deletions); return levenshtein; }; /** * Crush the diff into an encoded string which describes the operations * required to transform text1 into text2. * E.g. =3\t-2\t+ing -> Keep 3 chars, delete 2 chars, insert 'ing'. * Operations are tab-separated. Inserted text is escaped using %xx notation. * @param {!Array.} diffs Array of diff tuples. * @return {string} Delta text. */ diff_match_patch.prototype.diff_toDelta = function(diffs) { var text = []; for (var x = 0; x < diffs.length; x++) { switch (diffs[x][0]) { case DIFF_INSERT: text[x] = '+' + encodeURI(diffs[x][1]); break; case DIFF_DELETE: text[x] = '-' + diffs[x][1].length; break; case DIFF_EQUAL: text[x] = '=' + diffs[x][1].length; break; } } return text.join('\t').replace(/%20/g, ' '); }; /** * Given the original text1, and an encoded string which describes the * operations required to transform text1 into text2, compute the full diff. * @param {string} text1 Source string for the diff. * @param {string} delta Delta text. * @return {!Array.} Array of diff tuples. * @throws {!Error} If invalid input. */ diff_match_patch.prototype.diff_fromDelta = function(text1, delta) { var diffs = []; var diffsLength = 0; // Keeping our own length var is faster in JS. var pointer = 0; // Cursor in text1 var tokens = delta.split(/\t/g); for (var x = 0; x < tokens.length; x++) { // Each token begins with a one character parameter which specifies the // operation of this token (delete, insert, equality). var param = tokens[x].substring(1); switch (tokens[x].charAt(0)) { case '+': try { diffs[diffsLength++] = [DIFF_INSERT, decodeURI(param)]; } catch (ex) { // Malformed URI sequence. throw new Error('Illegal escape in diff_fromDelta: ' + param); } break; case '-': // Fall through. case '=': var n = parseInt(param, 10); if (isNaN(n) || n < 0) { throw new Error('Invalid number in diff_fromDelta: ' + param); } var text = text1.substring(pointer, pointer += n); if (tokens[x].charAt(0) == '=') { diffs[diffsLength++] = [DIFF_EQUAL, text]; } else { diffs[diffsLength++] = [DIFF_DELETE, text]; } break; default: // Blank tokens are ok (from a trailing \t). // Anything else is an error. if (tokens[x]) { throw new Error('Invalid diff operation in diff_fromDelta: ' + tokens[x]); } } } if (pointer != text1.length) { throw new Error('Delta length (' + pointer + ') does not equal source text length (' + text1.length + ').'); } return diffs; }; // MATCH FUNCTIONS /** * Locate the best instance of 'pattern' in 'text' near 'loc'. * @param {string} text The text to search. * @param {string} pattern The pattern to search for. * @param {number} loc The location to search around. * @return {number} Best match index or -1. */ diff_match_patch.prototype.match_main = function(text, pattern, loc) { // Check for null inputs. if (text == null || pattern == null || loc == null) { throw new Error('Null input. (match_main)'); } loc = Math.max(0, Math.min(loc, text.length)); if (text == pattern) { // Shortcut (potentially not guaranteed by the algorithm) return 0; } else if (!text.length) { // Nothing to match. return -1; } else if (text.substring(loc, loc + pattern.length) == pattern) { // Perfect match at the perfect spot! (Includes case of null pattern) return loc; } else { // Do a fuzzy compare. return this.match_bitap_(text, pattern, loc); } }; /** * Locate the best instance of 'pattern' in 'text' near 'loc' using the * Bitap algorithm. * @param {string} text The text to search. * @param {string} pattern The pattern to search for. * @param {number} loc The location to search around. * @return {number} Best match index or -1. * @private */ diff_match_patch.prototype.match_bitap_ = function(text, pattern, loc) { if (pattern.length > this.Match_MaxBits) { throw new Error('Pattern too long for this browser.'); } // Initialise the alphabet. var s = this.match_alphabet_(pattern); var dmp = this; // 'this' becomes 'window' in a closure. /** * Compute and return the score for a match with e errors and x location. * Accesses loc and pattern through being a closure. * @param {number} e Number of errors in match. * @param {number} x Location of match. * @return {number} Overall score for match (0.0 = good, 1.0 = bad). * @private */ function match_bitapScore_(e, x) { var accuracy = e / pattern.length; var proximity = Math.abs(loc - x); if (!dmp.Match_Distance) { // Dodge divide by zero error. return proximity ? 1.0 : accuracy; } return accuracy + (proximity / dmp.Match_Distance); } // Highest score beyond which we give up. var score_threshold = this.Match_Threshold; // Is there a nearby exact match? (speedup) var best_loc = text.indexOf(pattern, loc); if (best_loc != -1) { score_threshold = Math.min(match_bitapScore_(0, best_loc), score_threshold); // What about in the other direction? (speedup) best_loc = text.lastIndexOf(pattern, loc + pattern.length); if (best_loc != -1) { score_threshold = Math.min(match_bitapScore_(0, best_loc), score_threshold); } } // Initialise the bit arrays. var matchmask = 1 << (pattern.length - 1); best_loc = -1; var bin_min, bin_mid; var bin_max = pattern.length + text.length; var last_rd; for (var d = 0; d < pattern.length; d++) { // Scan for the best match; each iteration allows for one more error. // Run a binary search to determine how far from 'loc' we can stray at this // error level. bin_min = 0; bin_mid = bin_max; while (bin_min < bin_mid) { if (match_bitapScore_(d, loc + bin_mid) <= score_threshold) { bin_min = bin_mid; } else { bin_max = bin_mid; } bin_mid = Math.floor((bin_max - bin_min) / 2 + bin_min); } // Use the result from this iteration as the maximum for the next. bin_max = bin_mid; var start = Math.max(1, loc - bin_mid + 1); var finish = Math.min(loc + bin_mid, text.length) + pattern.length; var rd = Array(finish + 2); rd[finish + 1] = (1 << d) - 1; for (var j = finish; j >= start; j--) { // The alphabet (s) is a sparse hash, so the following line generates // warnings. var charMatch = s[text.charAt(j - 1)]; if (d === 0) { // First pass: exact match. rd[j] = ((rd[j + 1] << 1) | 1) & charMatch; } else { // Subsequent passes: fuzzy match. rd[j] = (((rd[j + 1] << 1) | 1) & charMatch) | (((last_rd[j + 1] | last_rd[j]) << 1) | 1) | last_rd[j + 1]; } if (rd[j] & matchmask) { var score = match_bitapScore_(d, j - 1); // This match will almost certainly be better than any existing match. // But check anyway. if (score <= score_threshold) { // Told you so. score_threshold = score; best_loc = j - 1; if (best_loc > loc) { // When passing loc, don't exceed our current distance from loc. start = Math.max(1, 2 * loc - best_loc); } else { // Already passed loc, downhill from here on in. break; } } } } // No hope for a (better) match at greater error levels. if (match_bitapScore_(d + 1, loc) > score_threshold) { break; } last_rd = rd; } return best_loc; }; /** * Initialise the alphabet for the Bitap algorithm. * @param {string} pattern The text to encode. * @return {!Object} Hash of character locations. * @private */ diff_match_patch.prototype.match_alphabet_ = function(pattern) { var s = {}; for (var i = 0; i < pattern.length; i++) { s[pattern.charAt(i)] = 0; } for (var i = 0; i < pattern.length; i++) { s[pattern.charAt(i)] |= 1 << (pattern.length - i - 1); } return s; }; // PATCH FUNCTIONS /** * Increase the context until it is unique, * but don't let the pattern expand beyond Match_MaxBits. * @param {!diff_match_patch.patch_obj} patch The patch to grow. * @param {string} text Source text. * @private */ diff_match_patch.prototype.patch_addContext_ = function(patch, text) { if (text.length == 0) { return; } var pattern = text.substring(patch.start2, patch.start2 + patch.length1); var padding = 0; // Look for the first and last matches of pattern in text. If two different // matches are found, increase the pattern length. while (text.indexOf(pattern) != text.lastIndexOf(pattern) && pattern.length < this.Match_MaxBits - this.Patch_Margin - this.Patch_Margin) { padding += this.Patch_Margin; pattern = text.substring(patch.start2 - padding, patch.start2 + patch.length1 + padding); } // Add one chunk for good luck. padding += this.Patch_Margin; // Add the prefix. var prefix = text.substring(patch.start2 - padding, patch.start2); if (prefix) { patch.diffs.unshift([DIFF_EQUAL, prefix]); } // Add the suffix. var suffix = text.substring(patch.start2 + patch.length1, patch.start2 + patch.length1 + padding); if (suffix) { patch.diffs.push([DIFF_EQUAL, suffix]); } // Roll back the start points. patch.start1 -= prefix.length; patch.start2 -= prefix.length; // Extend the lengths. patch.length1 += prefix.length + suffix.length; patch.length2 += prefix.length + suffix.length; }; /** * Compute a list of patches to turn text1 into text2. * Use diffs if provided, otherwise compute it ourselves. * There are four ways to call this function, depending on what data is * available to the caller: * Method 1: * a = text1, b = text2 * Method 2: * a = diffs * Method 3 (optimal): * a = text1, b = diffs * Method 4 (deprecated, use method 3): * a = text1, b = text2, c = diffs * * @param {string|!Array.} a text1 (methods 1,3,4) or * Array of diff tuples for text1 to text2 (method 2). * @param {string|!Array.} opt_b text2 (methods 1,4) or * Array of diff tuples for text1 to text2 (method 3) or undefined (method 2). * @param {string|!Array.} opt_c Array of diff tuples * for text1 to text2 (method 4) or undefined (methods 1,2,3). * @return {!Array.} Array of Patch objects. */ diff_match_patch.prototype.patch_make = function(a, opt_b, opt_c) { var text1, diffs; if (typeof a == 'string' && typeof opt_b == 'string' && typeof opt_c == 'undefined') { // Method 1: text1, text2 // Compute diffs from text1 and text2. text1 = /** @type {string} */(a); diffs = this.diff_main(text1, /** @type {string} */(opt_b), true); if (diffs.length > 2) { this.diff_cleanupSemantic(diffs); this.diff_cleanupEfficiency(diffs); } } else if (a && typeof a == 'object' && typeof opt_b == 'undefined' && typeof opt_c == 'undefined') { // Method 2: diffs // Compute text1 from diffs. diffs = /** @type {!Array.} */(a); text1 = this.diff_text1(diffs); } else if (typeof a == 'string' && opt_b && typeof opt_b == 'object' && typeof opt_c == 'undefined') { // Method 3: text1, diffs text1 = /** @type {string} */(a); diffs = /** @type {!Array.} */(opt_b); } else if (typeof a == 'string' && typeof opt_b == 'string' && opt_c && typeof opt_c == 'object') { // Method 4: text1, text2, diffs // text2 is not used. text1 = /** @type {string} */(a); diffs = /** @type {!Array.} */(opt_c); } else { throw new Error('Unknown call format to patch_make.'); } if (diffs.length === 0) { return []; // Get rid of the null case. } var patches = []; var patch = new diff_match_patch.patch_obj(); var patchDiffLength = 0; // Keeping our own length var is faster in JS. var char_count1 = 0; // Number of characters into the text1 string. var char_count2 = 0; // Number of characters into the text2 string. // Start with text1 (prepatch_text) and apply the diffs until we arrive at // text2 (postpatch_text). We recreate the patches one by one to determine // context info. var prepatch_text = text1; var postpatch_text = text1; for (var x = 0; x < diffs.length; x++) { var diff_type = diffs[x][0]; var diff_text = diffs[x][1]; if (!patchDiffLength && diff_type !== DIFF_EQUAL) { // A new patch starts here. patch.start1 = char_count1; patch.start2 = char_count2; } switch (diff_type) { case DIFF_INSERT: patch.diffs[patchDiffLength++] = diffs[x]; patch.length2 += diff_text.length; postpatch_text = postpatch_text.substring(0, char_count2) + diff_text + postpatch_text.substring(char_count2); break; case DIFF_DELETE: patch.length1 += diff_text.length; patch.diffs[patchDiffLength++] = diffs[x]; postpatch_text = postpatch_text.substring(0, char_count2) + postpatch_text.substring(char_count2 + diff_text.length); break; case DIFF_EQUAL: if (diff_text.length <= 2 * this.Patch_Margin && patchDiffLength && diffs.length != x + 1) { // Small equality inside a patch. patch.diffs[patchDiffLength++] = diffs[x]; patch.length1 += diff_text.length; patch.length2 += diff_text.length; } else if (diff_text.length >= 2 * this.Patch_Margin) { // Time for a new patch. if (patchDiffLength) { this.patch_addContext_(patch, prepatch_text); patches.push(patch); patch = new diff_match_patch.patch_obj(); patchDiffLength = 0; // Unlike Unidiff, our patch lists have a rolling context. // http://code.google.com/p/google-diff-match-patch/wiki/Unidiff // Update prepatch text & pos to reflect the application of the // just completed patch. prepatch_text = postpatch_text; char_count1 = char_count2; } } break; } // Update the current character count. if (diff_type !== DIFF_INSERT) { char_count1 += diff_text.length; } if (diff_type !== DIFF_DELETE) { char_count2 += diff_text.length; } } // Pick up the leftover patch if not empty. if (patchDiffLength) { this.patch_addContext_(patch, prepatch_text); patches.push(patch); } return patches; }; /** * Given an array of patches, return another array that is identical. * @param {!Array.} patches Array of Patch objects. * @return {!Array.} Array of Patch objects. */ diff_match_patch.prototype.patch_deepCopy = function(patches) { // Making deep copies is hard in JavaScript. var patchesCopy = []; for (var x = 0; x < patches.length; x++) { var patch = patches[x]; var patchCopy = new diff_match_patch.patch_obj(); patchCopy.diffs = []; for (var y = 0; y < patch.diffs.length; y++) { patchCopy.diffs[y] = patch.diffs[y].slice(); } patchCopy.start1 = patch.start1; patchCopy.start2 = patch.start2; patchCopy.length1 = patch.length1; patchCopy.length2 = patch.length2; patchesCopy[x] = patchCopy; } return patchesCopy; }; /** * Merge a set of patches onto the text. Return a patched text, as well * as a list of true/false values indicating which patches were applied. * @param {!Array.} patches Array of Patch objects. * @param {string} text Old text. * @return {!Array.>} Two element Array, containing the * new text and an array of boolean values. */ diff_match_patch.prototype.patch_apply = function(patches, text) { if (patches.length == 0) { return [text, []]; } // Deep copy the patches so that no changes are made to originals. patches = this.patch_deepCopy(patches); var nullPadding = this.patch_addPadding(patches); text = nullPadding + text + nullPadding; this.patch_splitMax(patches); // delta keeps track of the offset between the expected and actual location // of the previous patch. If there are patches expected at positions 10 and // 20, but the first patch was found at 12, delta is 2 and the second patch // has an effective expected position of 22. var delta = 0; var results = []; for (var x = 0; x < patches.length; x++) { var expected_loc = patches[x].start2 + delta; var text1 = this.diff_text1(patches[x].diffs); var start_loc; var end_loc = -1; if (text1.length > this.Match_MaxBits) { // patch_splitMax will only provide an oversized pattern in the case of // a monster delete. start_loc = this.match_main(text, text1.substring(0, this.Match_MaxBits), expected_loc); if (start_loc != -1) { end_loc = this.match_main(text, text1.substring(text1.length - this.Match_MaxBits), expected_loc + text1.length - this.Match_MaxBits); if (end_loc == -1 || start_loc >= end_loc) { // Can't find valid trailing context. Drop this patch. start_loc = -1; } } } else { start_loc = this.match_main(text, text1, expected_loc); } if (start_loc == -1) { // No match found. :( results[x] = false; // Subtract the delta for this failed patch from subsequent patches. delta -= patches[x].length2 - patches[x].length1; } else { // Found a match. :) results[x] = true; delta = start_loc - expected_loc; var text2; if (end_loc == -1) { text2 = text.substring(start_loc, start_loc + text1.length); } else { text2 = text.substring(start_loc, end_loc + this.Match_MaxBits); } if (text1 == text2) { // Perfect match, just shove the replacement text in. text = text.substring(0, start_loc) + this.diff_text2(patches[x].diffs) + text.substring(start_loc + text1.length); } else { // Imperfect match. Run a diff to get a framework of equivalent // indices. var diffs = this.diff_main(text1, text2, false); if (text1.length > this.Match_MaxBits && this.diff_levenshtein(diffs) / text1.length > this.Patch_DeleteThreshold) { // The end points match, but the content is unacceptably bad. results[x] = false; } else { this.diff_cleanupSemanticLossless(diffs); var index1 = 0; var index2; for (var y = 0; y < patches[x].diffs.length; y++) { var mod = patches[x].diffs[y]; if (mod[0] !== DIFF_EQUAL) { index2 = this.diff_xIndex(diffs, index1); } if (mod[0] === DIFF_INSERT) { // Insertion text = text.substring(0, start_loc + index2) + mod[1] + text.substring(start_loc + index2); } else if (mod[0] === DIFF_DELETE) { // Deletion text = text.substring(0, start_loc + index2) + text.substring(start_loc + this.diff_xIndex(diffs, index1 + mod[1].length)); } if (mod[0] !== DIFF_DELETE) { index1 += mod[1].length; } } } } } } // Strip the padding off. text = text.substring(nullPadding.length, text.length - nullPadding.length); return [text, results]; }; /** * Add some padding on text start and end so that edges can match something. * Intended to be called only from within patch_apply. * @param {!Array.} patches Array of Patch objects. * @return {string} The padding string added to each side. */ diff_match_patch.prototype.patch_addPadding = function(patches) { var paddingLength = this.Patch_Margin; var nullPadding = ''; for (var x = 1; x <= paddingLength; x++) { nullPadding += String.fromCharCode(x); } // Bump all the patches forward. for (var x = 0; x < patches.length; x++) { patches[x].start1 += paddingLength; patches[x].start2 += paddingLength; } // Add some padding on start of first diff. var patch = patches[0]; var diffs = patch.diffs; if (diffs.length == 0 || diffs[0][0] != DIFF_EQUAL) { // Add nullPadding equality. diffs.unshift([DIFF_EQUAL, nullPadding]); patch.start1 -= paddingLength; // Should be 0. patch.start2 -= paddingLength; // Should be 0. patch.length1 += paddingLength; patch.length2 += paddingLength; } else if (paddingLength > diffs[0][1].length) { // Grow first equality. var extraLength = paddingLength - diffs[0][1].length; diffs[0][1] = nullPadding.substring(diffs[0][1].length) + diffs[0][1]; patch.start1 -= extraLength; patch.start2 -= extraLength; patch.length1 += extraLength; patch.length2 += extraLength; } // Add some padding on end of last diff. patch = patches[patches.length - 1]; diffs = patch.diffs; if (diffs.length == 0 || diffs[diffs.length - 1][0] != DIFF_EQUAL) { // Add nullPadding equality. diffs.push([DIFF_EQUAL, nullPadding]); patch.length1 += paddingLength; patch.length2 += paddingLength; } else if (paddingLength > diffs[diffs.length - 1][1].length) { // Grow last equality. var extraLength = paddingLength - diffs[diffs.length - 1][1].length; diffs[diffs.length - 1][1] += nullPadding.substring(0, extraLength); patch.length1 += extraLength; patch.length2 += extraLength; } return nullPadding; }; /** * Look through the patches and break up any which are longer than the maximum * limit of the match algorithm. * Intended to be called only from within patch_apply. * @param {!Array.} patches Array of Patch objects. */ diff_match_patch.prototype.patch_splitMax = function(patches) { var patch_size = this.Match_MaxBits; for (var x = 0; x < patches.length; x++) { if (patches[x].length1 <= patch_size) { continue; } var bigpatch = patches[x]; // Remove the big old patch. patches.splice(x--, 1); var start1 = bigpatch.start1; var start2 = bigpatch.start2; var precontext = ''; while (bigpatch.diffs.length !== 0) { // Create one of several smaller patches. var patch = new diff_match_patch.patch_obj(); var empty = true; patch.start1 = start1 - precontext.length; patch.start2 = start2 - precontext.length; if (precontext !== '') { patch.length1 = patch.length2 = precontext.length; patch.diffs.push([DIFF_EQUAL, precontext]); } while (bigpatch.diffs.length !== 0 && patch.length1 < patch_size - this.Patch_Margin) { var diff_type = bigpatch.diffs[0][0]; var diff_text = bigpatch.diffs[0][1]; if (diff_type === DIFF_INSERT) { // Insertions are harmless. patch.length2 += diff_text.length; start2 += diff_text.length; patch.diffs.push(bigpatch.diffs.shift()); empty = false; } else if (diff_type === DIFF_DELETE && patch.diffs.length == 1 && patch.diffs[0][0] == DIFF_EQUAL && diff_text.length > 2 * patch_size) { // This is a large deletion. Let it pass in one chunk. patch.length1 += diff_text.length; start1 += diff_text.length; empty = false; patch.diffs.push([diff_type, diff_text]); bigpatch.diffs.shift(); } else { // Deletion or equality. Only take as much as we can stomach. diff_text = diff_text.substring(0, patch_size - patch.length1 - this.Patch_Margin); patch.length1 += diff_text.length; start1 += diff_text.length; if (diff_type === DIFF_EQUAL) { patch.length2 += diff_text.length; start2 += diff_text.length; } else { empty = false; } patch.diffs.push([diff_type, diff_text]); if (diff_text == bigpatch.diffs[0][1]) { bigpatch.diffs.shift(); } else { bigpatch.diffs[0][1] = bigpatch.diffs[0][1].substring(diff_text.length); } } } // Compute the head context for the next patch. precontext = this.diff_text2(patch.diffs); precontext = precontext.substring(precontext.length - this.Patch_Margin); // Append the end context for this patch. var postcontext = this.diff_text1(bigpatch.diffs) .substring(0, this.Patch_Margin); if (postcontext !== '') { patch.length1 += postcontext.length; patch.length2 += postcontext.length; if (patch.diffs.length !== 0 && patch.diffs[patch.diffs.length - 1][0] === DIFF_EQUAL) { patch.diffs[patch.diffs.length - 1][1] += postcontext; } else { patch.diffs.push([DIFF_EQUAL, postcontext]); } } if (!empty) { patches.splice(++x, 0, patch); } } } }; /** * Take a list of patches and return a textual representation. * @param {!Array.} patches Array of Patch objects. * @return {string} Text representation of patches. */ diff_match_patch.prototype.patch_toText = function(patches) { var text = []; for (var x = 0; x < patches.length; x++) { text[x] = patches[x]; } return text.join(''); }; /** * Parse a textual representation of patches and return a list of Patch objects. * @param {string} textline Text representation of patches. * @return {!Array.} Array of Patch objects. * @throws {!Error} If invalid input. */ diff_match_patch.prototype.patch_fromText = function(textline) { var patches = []; if (!textline) { return patches; } var text = textline.split('\n'); var textPointer = 0; var patchHeader = /^@@ -(\d+),?(\d*) \+(\d+),?(\d*) @@$/; while (textPointer < text.length) { var m = text[textPointer].match(patchHeader); if (!m) { throw new Error('Invalid patch string: ' + text[textPointer]); } var patch = new diff_match_patch.patch_obj(); patches.push(patch); patch.start1 = parseInt(m[1], 10); if (m[2] === '') { patch.start1--; patch.length1 = 1; } else if (m[2] == '0') { patch.length1 = 0; } else { patch.start1--; patch.length1 = parseInt(m[2], 10); } patch.start2 = parseInt(m[3], 10); if (m[4] === '') { patch.start2--; patch.length2 = 1; } else if (m[4] == '0') { patch.length2 = 0; } else { patch.start2--; patch.length2 = parseInt(m[4], 10); } textPointer++; while (textPointer < text.length) { var sign = text[textPointer].charAt(0); try { var line = decodeURI(text[textPointer].substring(1)); } catch (ex) { // Malformed URI sequence. throw new Error('Illegal escape in patch_fromText: ' + line); } if (sign == '-') { // Deletion. patch.diffs.push([DIFF_DELETE, line]); } else if (sign == '+') { // Insertion. patch.diffs.push([DIFF_INSERT, line]); } else if (sign == ' ') { // Minor equality. patch.diffs.push([DIFF_EQUAL, line]); } else if (sign == '@') { // Start of next patch. break; } else if (sign === '') { // Blank line? Whatever. } else { // WTF? throw new Error('Invalid patch mode "' + sign + '" in: ' + line); } textPointer++; } } return patches; }; /** * Class representing one patch operation. * @constructor */ diff_match_patch.patch_obj = function() { /** @type {!Array.} */ this.diffs = []; /** @type {?number} */ this.start1 = null; /** @type {?number} */ this.start2 = null; /** @type {number} */ this.length1 = 0; /** @type {number} */ this.length2 = 0; }; /** * Emmulate GNU diff's format. * Header: @@ -382,8 +481,9 @@ * Indicies are printed as 1-based, not 0-based. * @return {string} The GNU diff string. */ diff_match_patch.patch_obj.prototype.toString = function() { var coords1, coords2; if (this.length1 === 0) { coords1 = this.start1 + ',0'; } else if (this.length1 == 1) { coords1 = this.start1 + 1; } else { coords1 = (this.start1 + 1) + ',' + this.length1; } if (this.length2 === 0) { coords2 = this.start2 + ',0'; } else if (this.length2 == 1) { coords2 = this.start2 + 1; } else { coords2 = (this.start2 + 1) + ',' + this.length2; } var text = ['@@ -' + coords1 + ' +' + coords2 + ' @@\n']; var op; // Escape the body of the patch with %xx notation. for (var x = 0; x < this.diffs.length; x++) { switch (this.diffs[x][0]) { case DIFF_INSERT: op = '+'; break; case DIFF_DELETE: op = '-'; break; case DIFF_EQUAL: op = ' '; break; } text[x + 1] = op + encodeURI(this.diffs[x][1]) + '\n'; } return text.join('').replace(/%20/g, ' '); }; // Export these global variables so that they survive Google's JS compiler. // In a browser, 'this' will be 'window'. // Users of node.js should 'require' the uncompressed version since Google's // JS compiler may break the following exports for non-browser environments. this['diff_match_patch'] = diff_match_patch; this['DIFF_DELETE'] = DIFF_DELETE; this['DIFF_INSERT'] = DIFF_INSERT; this['DIFF_EQUAL'] = DIFF_EQUAL; package/javascript/diff_match_patch_test.html000644 001751 000062 0000011563 12066005331020065 0ustar00000000 000000

If debugging errors, start with the first reported error, subsequent tests often rely on earlier ones.

package/demos/demo_match.html000644 001751 000062 0000010050 12066005331014612 0ustar00000000 000000 Diff, Match and Patch: Demo of Match

Diff, Match and Patch

Demo of Match

Match looks for a pattern within a larger text. This implementation of match is fuzzy, meaning it can find a match even if the pattern contains errors and doesn't exactly match what is found in the text. This implementation also accepts an expected location, near which the match should be found. The candidate matches are scored based on a) the number of spelling differences between the pattern and the text and b) the distance between the candidate match and the expected location. The match distance parameter sets the relative importance of these two metrics.

Text:

Fuzzy pattern:


Aproximate pattern to search for in the text. Due to limitations of the Bitap algorithm, the pattern has a limited length.

Fuzzy location:


Aproximately where in the text is the pattern expected to be found?

Match distance:


Determines how close the match must be to the fuzzy location (specified above). An exact letter match which is 'distance' characters away from the fuzzy location would score as a complete mismatch. A distance of '0' requires the match be at the exact location specified, a threshold of '1000' would require a perfect match to be within 800 characters of the fuzzy location to be found using a 0.8 threshold.

Match threshold:


At what point does the match algorithm give up. A threshold of '0.0' requires a perfect match (of both letters and location), a threshold of '1.0' would match anything.

Back to Diff, Match and Patch package/demos/demo_diff.html000644 001751 000062 0000006653 12066005331014444 0ustar00000000 000000 Diff, Match and Patch: Demo of Diff

Diff, Match and Patch

Demo of Diff

Diff takes two texts and finds the differences. This implementation works on a character by character basis. The result of any diff may contain 'chaff', irrelevant small commonalities which complicate the output. A post-diff cleanup algorithm factors out these trivial commonalities.

Text Version 1:

Text Version 2:

Diff timeout:

seconds
If the mapping phase of the diff computation takes longer than this, then the computation is truncated and the best solution to date is returned. While guaranteed to be correct, it may not be optimal. A timeout of '0' allows for unlimited computation.

Post-diff cleanup:

Increase human readability by factoring out commonalities which are likely to be coincidental.
, edit cost:
Increase computational efficiency by factoring out short commonalities which are not worth the overhead. The larger the edit cost, the more agressive the cleanup.
Raw output.

Back to Diff, Match and Patch package/demos/demo_patch.html000644 001751 000062 0000010747 12066005331014632 0ustar00000000 000000 Diff, Match and Patch: Demo of Patch

Diff, Match and Patch

Demo of Patch

Two texts can be diffed against each other, generating a list of patches. These patches can then be applied against a third text. If the third text has edits of its own, this version of patch will apply its changes on a best-effort basis, reporting which patches succeeded and which failed.

In this scenario Shakespeare wrote Hamlet in old English, the source document. Then two derivative works were created. One is Hamlet updated to modern English. The other is a Star Trek parody in old English. This demonstrantion creates a list of patches between the source and the modern English version. Then it applies those patches onto the Star Trek parody, thus creating a Star Trek parody in modern English.

Shakespeare's copy:

Old Version:
 New Version:

Trekkie's copy:

Old Version:
 New Version:
    Back to Diff, Match and Patch