Patches

UnifiedDiff reads and applies diffs in the unified format, the format of diff -u and git diff. To write one, see Write a unified diff.

Apply a patch

UnifiedDiff.Apply applies a patch of one file to a text and returns the patched text:

var patch = UnifiedDiff.Create("one\ntwo\nthree\n", "one\n2\nthree\n").ToString();

// A line was added at the start since the patch was made: the hunk still applies, one line further down
var patched = UnifiedDiff.Apply("zero\none\ntwo\nthree\n", patch);

Assert.Equal("zero\none\n2\nthree\n", patched);

• The context and deleted lines of a hunk must match the text exactly, line terminators included.
• When a hunk does not match at the line its header gives, it is applied at the closest line where it matches, as patch does when lines were added or removed elsewhere. The hunks apply in order, each after the previous one.
• \ No newline at end of file is honored both ways: the hunk matches only a text whose last line has no line feed, and writes one without it.

When a hunk matches nowhere, Apply throws a UnifiedDiffApplyException whose HunkIndex is the failing hunk; the text is not changed:

var patch = UnifiedDiff.Create("one\ntwo\nthree\n", "one\n2\nthree\n").ToString();

// The line the hunk deletes changed: the hunk no longer matches
var exception = Assert.Throws<UnifiedDiffApplyException>(() => UnifiedDiff.Apply("one\nTWO\nthree\n", patch));

Assert.Equal(0, exception.HunkIndex);

Read a patch

UnifiedDiff.Parse reads every file of a patch. The lines outside the files, such as the diff --git and index lines of git, are skipped:

var patch =
    "diff --git a/greeting.txt b/greeting.txt\n" +
    "index 3b18e51..a042389 100644\n" +
    "--- a/greeting.txt\n" +
    "+++ b/greeting.txt\n" +
    "@@ -1,2 +1,2 @@ Greetings\n" +
    " Hello\n" +
    "-World\n" +
    "+There\n";

var files = UnifiedDiff.Parse(patch);

var file = files[0];
Assert.Equal("a/greeting.txt", file.OldFileName);
Assert.Equal("b/greeting.txt", file.NewFileName);

var hunk = file.Hunks[0];
Assert.Equal(1, hunk.OldStart);
Assert.Equal(2, hunk.OldLineCount);
Assert.Equal("Greetings", hunk.SectionHeading);
Assert.Equal([TextDiffOperation.Equal, TextDiffOperation.Delete, TextDiffOperation.Insert], hunk.Lines.Select(line => line.Operation));
Assert.Equal(["Hello\n", "World\n", "There\n"], hunk.Lines.Select(line => line.Text));

Assert.Equal("Hello\nThere\n", file.Apply("Hello\nWorld\n"));

• A file starts at a --- line followed by a +++ line. The names exclude the timestamp diff writes after a tab. Match them against your files yourself: the a/ and b/ prefixes of git are kept.
• UnifiedDiffFile.Apply applies one file of a multi-file patch; UnifiedDiff.Apply throws an ArgumentException when the patch does not hold exactly one file.
• OldStart and NewStart are 1-based. For an empty range, they give the line before it, as in the @@ line.
• Each UnifiedDiffLine has an Operation (Equal for context) and a Text that ends with \n, unless it is marked \ No newline at end of file.
• ToString writes a file or a hunk back in the unified format.
• A malformed hunk throws a FormatException that gives the line number: an invalid @@ line, a line with an unknown prefix, or fewer lines than the header gives. A patch with CRLF line terminators is read: the \r belongs to the lines of the hunks, as in the files it patches.