diffstat (Stat method on Hunk and FileDiff)

Author: sqs

diff/diff.go

@@ -1,6 +1,9 @@
 package diff
 
-import "time"
+import (
+	"bytes"
+	"time"
+)
 
 // A FileDiff represents a unified diff for a single file.
 //
@@ -20,6 +23,16 @@ type FileDiff struct {
 	Hunks []*Hunk // hunks that were changed from orig to new
 }
 
+// Stat computes the number of lines added/changed/deleted in all
+// hunks in this file's diff.
+func (d *FileDiff) Stat() Stat {
+	total := Stat{}
+	for _, h := range d.Hunks {
+		total.add(h.Stat())
+	}
+	return total
+}
+
 // A Hunk represents a series of changes (additions or deletions) in a
 // file's unified diff.
 type Hunk struct {
@@ -34,6 +47,43 @@ type Hunk struct {
 	Body []byte // hunk body (lines prefixed with '-', '+', or ' ')
 }
 
+// Stat computes the number of lines added/changed/deleted in this
+// hunk.
+func (h *Hunk) Stat() Stat {
+	lines := bytes.Split(h.Body, []byte{'\n'})
+	var last byte
+	st := Stat{}
+	for _, line := range lines {
+		if len(line) == 0 {
+			last = 0
+			continue
+		}
+		switch line[0] {
+		case '-':
+			if last == '+' {
+				st.Added--
+				st.Changed++
+				last = 0 // next line can't change this one since this is already a change
+			} else {
+				st.Deleted++
+				last = line[0]
+			}
+		case '+':
+			if last == '-' {
+				st.Deleted--
+				st.Changed++
+				last = 0 // next line can't change this one since this is already a change
+			} else {
+				st.Added++
+				last = line[0]
+			}
+		default:
+			last = 0
+		}
+	}
+	return st
+}
+
 var (
 	hunkPrefix = []byte("@@ ")
 )
@@ -44,3 +94,15 @@ const hunkHeader = "@@ -%d,%d +%d,%d @@"
 // header timestamps. See
 // http://www.gnu.org/software/diffutils/manual/html_node/Detailed-Unified.html.
 const diffTimeFormat = "2006-01-02 15:04:05.000000000 -0700"
+
+// A Stat is a diff stat that represents the number of lines
+// added/changed/deleted.
+type Stat struct {
+	Added, Changed, Deleted int // numbers of lines
+}
+
+func (s *Stat) add(o Stat) {
+	s.Added += o.Added
+	s.Changed += o.Changed
+	s.Deleted += o.Deleted
+}

diff/diff_test.go

@@ -189,3 +189,63 @@ func TestNoNewlineAtEnd(t *testing.T) {
 		}
 	}
 }
+
+func TestFileDiff_Stat(t *testing.T) {
+	tests := map[string]struct {
+		hunks []*Hunk
+		want  Stat
+	}{
+		"no change": {
+			hunks: []*Hunk{
+				{Body: []byte(`@@ -0,0 +0,0
+ a
+ b
+`)},
+			},
+			want: Stat{},
+		},
+		"added/deleted": {
+			hunks: []*Hunk{
+				{Body: []byte(`@@ -0,0 +0,0
++a
+ b
+-c
+ d
+`)},
+			},
+			want: Stat{Added: 1, Deleted: 1},
+		},
+		"changed": {
+			hunks: []*Hunk{
+				{Body: []byte(`@@ -0,0 +0,0
++a
++b
+-c
+-d
+ e
+`)},
+			},
+			want: Stat{Added: 1, Changed: 1, Deleted: 1},
+		},
+		"many changes": {
+			hunks: []*Hunk{
+				{Body: []byte(`@@ -0,0 +0,0
++a
+-b
++c
+-d
+ e
+`)},
+			},
+			want: Stat{Added: 0, Changed: 2, Deleted: 0},
+		},
+	}
+	for label, test := range tests {
+		fdiff := &FileDiff{Hunks: test.hunks}
+		stat := fdiff.Stat()
+		if !reflect.DeepEqual(stat, test.want) {
+			t.Errorf("%s: got diff stat %+v, want %+v", label, stat, test.want)
+			continue
+		}
+	}
+}