Stop treating stuff like vertical tabs as line breaks when dealing wi…

…th unified diffs (#435)

* Add test of vertical tab handling

* Stop treating stuff like vertical tabs as line breaks when dealing with unified diffs

* Add release notes

* Add a consistency test

release-notes.md

@@ -1,5 +1,11 @@
 # Release Notes
 
+## Future breaking 6.0.0 release
+
+[Commits](https://github.com/kpdecker/jsdiff/compare/master...v6.0.0-staging)
+
+- [#435](https://github.com/kpdecker/jsdiff/pull/435) Fix `parsePatch` handling of control characters. `parsePatch` used to interpret various unusual control characters - namely vertical tabs, form feeds, lone carriage returns without a line feed, and EBCDIC NELs - as line breaks when parsing a patch file. This was inconsistent with the behavior of both JsDiff's own `diffLines` method and also the Unix `diff` and `patch` utils, which all simply treat those control characters as ordinary characters. The result of this discrepancy was that some well-formed patches - produced either by `diff` or by JsDiff itself and handled properly by the `patch` util - would be wrongly parsed by `parsePatch`, with the effect that it would disregard the remainder of a hunk after encountering one of these control characters.
+
 ## Development
 
 [Commits](https://github.com/kpdecker/jsdiff/compare/v5.1.0...master)

src/patch/apply.js

@@ -15,8 +15,8 @@ export function applyPatch(source, uniDiff, options = {}) {
   }
 
   // Apply the diff to the input
-  let lines = source.split(/\r\n|[\n\v\f\r\x85]/),
-      delimiters = source.match(/\r\n|[\n\v\f\r\x85]/g) || [],
+  let lines = source.split(/\r?\n/),
+      delimiters = source.match(/\r?\n/g) || [],
       hunks = uniDiff.hunks,
 
       compareLine = options.compareLine || ((lineNumber, line, operation, patchContent) => line === patchContent),

src/patch/parse.js

@@ -1,6 +1,6 @@
 export function parsePatch(uniDiff, options = {}) {
-  let diffstr = uniDiff.split(/\r\n|[\n\v\f\r\x85]/),
-      delimiters = uniDiff.match(/\r\n|[\n\v\f\r\x85]/g) || [],
+  let diffstr = uniDiff.split(/\r?\n/),
+      delimiters = uniDiff.match(/\r?\n/g) || [],
       list = [],
       i = 0;
 

test/patch/parse.js

@@ -1,4 +1,5 @@
 import {parsePatch} from '../../lib/patch/parse';
+import {createPatch} from '../../lib/patch/create';
 
 import {expect} from 'chai';
 
@@ -411,5 +412,91 @@ Index: test2
     it('should handle OOM case', function() {
       parsePatch('Index: \n===================================================================\n--- \n+++ \n@@ -1,1 +1,2 @@\n-1\n\\ No newline at end of file\n+1\n+2\n');
     });
+
+    it('should treat vertical tabs like ordinary characters', function() {
+      // Patch below was generated as follows:
+      // 1. From Node, run:
+      //      fs.writeFileSync("foo", "foo\nbar\vbar\nbaz\nqux")
+      //      fs.writeFileSync("bar", "foo\nbarry\vbarry\nbaz\nqux")
+      // 2. From shell, run:
+      //      diff -u foo bar > diff.txt
+      // 3. From Node, run
+      //      fs.readFileSync("diff.txt")
+      //    and copy the string literal you get.
+      // This patch illustrates how the Unix `diff` and `patch` tools handle
+      // characters like vertical tabs - namely, they simply treat them as
+      // ordinary characters, NOT as line breaks. JsDiff used to treat them as
+      // line breaks but this breaks its parsing of patches like this one; this
+      // test was added to demonstrate the brokenness and prevent us from
+      // reintroducing it.
+      const patch = '--- foo\t2023-12-20 16:11:20.908225554 +0000\n' +
+      '+++ bar\t2023-12-20 16:11:34.391473579 +0000\n' +
+      '@@ -1,4 +1,4 @@\n' +
+      ' foo\n' +
+      '-bar\x0Bbar\n' +
+      '+barry\x0Bbarry\n' +
+      ' baz\n' +
+      ' qux\n' +
+      '\\ No newline at end of file\n';
+
+      expect(parsePatch(patch)).to.eql([
+        {
+          oldFileName: 'foo',
+          oldHeader: '2023-12-20 16:11:20.908225554 +0000',
+          newFileName: 'bar',
+          newHeader: '2023-12-20 16:11:34.391473579 +0000',
+          hunks: [
+            {
+              oldStart: 1,
+              oldLines: 4,
+              newStart: 1,
+              newLines: 4,
+              lines: [
+                ' foo',
+                '-bar\x0Bbar',
+                '+barry\x0Bbarry',
+                ' baz',
+                ' qux',
+                '\\ No newline at end of file'
+              ],
+              linedelimiters: [ '\n', '\n', '\n', '\n', '\n', '\n' ]
+            }
+          ]
+        }
+      ]);
+    });
+
+    it('should treat vertical tabs in a way consistent with createPatch', function() {
+      // This is basically the same as the test above, but this time we create
+      // the patch USING JsDiff instead of testing one created with Unix diff
+      const patch = createPatch('foo', 'foo\nbar\vbar\nbaz\nqux', 'foo\nbarry\vbarry\nbaz\nqux');
+
+      expect(parsePatch(patch)).to.eql([
+        {
+          oldFileName: 'foo',
+          oldHeader: '',
+          newFileName: 'foo',
+          newHeader: '',
+          index: 'foo',
+          hunks: [
+            {
+              oldStart: 1,
+              oldLines: 4,
+              newStart: 1,
+              newLines: 4,
+              lines: [
+                ' foo',
+                '-bar\x0Bbar',
+                '+barry\x0Bbarry',
+                ' baz',
+                ' qux',
+                '\\ No newline at end of file'
+              ],
+              linedelimiters: [ '\n', '\n', '\n', '\n', '\n', '\n' ]
+            }
+          ]
+        }
+      ]);
+    });
   });
 });