internal/lsp: add unified diff

This adds unified diff handling to the diff package.
It also adds a lot more testing and also verifies the unified
diff tests against the diff program if you run the tests with
the flag -verify-diff
This functionality is needed for some of the command verbs.

Change-Id: I817438fd25c0b16f3f31578f51a886944e74a948
Reviewed-on: https://go-review.googlesource.com/c/tools/+/171024
Reviewed-by: Rebecca Stambler <rstambler@golang.org>

internal/lsp/diff/diff_test.go

@@ -5,18 +5,62 @@
 package diff
 
 import (
+	"flag"
+	"fmt"
+	"io/ioutil"
+	"os"
+	"os/exec"
 	"reflect"
 	"strings"
 	"testing"
 )
 
+const (
+	fileA         = "a/a.go"
+	fileB         = "b/b.go"
+	unifiedPrefix = "--- " + fileA + "\n+++ " + fileB + "\n"
+)
+
+var verifyDiff = flag.Bool("verify-diff", false, "Check that the unified diff output matches `diff -u`")
+
 func TestDiff(t *testing.T) {
 	for _, test := range []struct {
 		a, b       string
 		lines      []*Op
 		operations []*Op
+		unified    string
+		nodiff     bool
 	}{
 		{
+			a:          "A\nB\nC\n",
+			b:          "A\nB\nC\n",
+			operations: []*Op{},
+			unified: `
+`[1:]}, {
+			a: "A\n",
+			b: "B\n",
+			operations: []*Op{
+				&Op{Kind: Delete, I1: 0, I2: 1, J1: 0},
+				&Op{Kind: Insert, Content: []string{"B\n"}, I1: 1, I2: 1, J1: 0},
+			},
+			unified: `
+@@ -1 +1 @@
+-A
++B
+`[1:]}, {
+			a: "A",
+			b: "B",
+			operations: []*Op{
+				&Op{Kind: Delete, I1: 0, I2: 1, J1: 0},
+				&Op{Kind: Insert, Content: []string{"B"}, I1: 1, I2: 1, J1: 0},
+			},
+			unified: `
+@@ -1 +1 @@
+-A
+\ No newline at end of file
++B
+\ No newline at end of file
+`[1:]}, {
 			a: "A\nB\nC\nA\nB\nB\nA\n",
 			b: "C\nB\nA\nB\nA\nC\n",
 			operations: []*Op{
@@ -26,6 +70,19 @@ func TestDiff(t *testing.T) {
 				&Op{Kind: Delete, I1: 5, I2: 6, J1: 4},
 				&Op{Kind: Insert, Content: []string{"C\n"}, I1: 7, I2: 7, J1: 5},
 			},
+			unified: `
+@@ -1,7 +1,6 @@
+-A
+-B
+ C
++B
+ A
+ B
+-B
+ A
++C
+`[1:],
+			nodiff: true, // diff algorithm produces different delete/insert pattern
 		},
 		{
 			a: "A\nB\n",
@@ -35,26 +92,111 @@ func TestDiff(t *testing.T) {
 				&Op{Kind: Insert, Content: []string{"C\n"}, I1: 2, I2: 2, J1: 1},
 				&Op{Kind: Insert, Content: []string{"\n"}, I1: 2, I2: 2, J1: 2},
 			},
+			unified: `
+@@ -1,2 +1,3 @@
+ A
+-B
++C
++
+`[1:],
 		},
+		{
+			a: "A\nB\nC\nD\nE\nF\nG\n",
+			b: "A\nH\nI\nJ\nE\nF\nK\n",
+			unified: `
+@@ -1,7 +1,7 @@
+ A
+-B
+-C
+-D
++H
++I
++J
+ E
+ F
+-G
++K
+`[1:]},
 	} {
 		a := strings.SplitAfter(test.a, "\n")
 		b := strings.SplitAfter(test.b, "\n")
 		ops := Operations(a, b)
-		if len(ops) != len(test.operations) {
-			t.Fatalf("expected %v operations, got %v", len(test.operations), len(ops))
-		}
-		for i, got := range ops {
-			want := test.operations[i]
-			if !reflect.DeepEqual(want, got) {
-				t.Errorf("expected %v, got %v", want, got)
+		if test.operations != nil {
+			if len(ops) != len(test.operations) {
+				t.Fatalf("expected %v operations, got %v", len(test.operations), len(ops))
+			}
+			for i, got := range ops {
+				want := test.operations[i]
+				if !reflect.DeepEqual(want, got) {
+					t.Errorf("expected %v, got %v", want, got)
+				}
 			}
 		}
-		applied := ApplyEdits(a, test.operations)
+		applied := ApplyEdits(a, ops)
 		for i, want := range applied {
 			got := b[i]
 			if got != want {
 				t.Errorf("expected %v got %v", want, got)
 			}
 		}
+		if test.unified != "" {
+			diff := ToUnified(fileA, fileB, a, ops)
+			got := fmt.Sprint(diff)
+			if !strings.HasPrefix(got, unifiedPrefix) {
+				t.Errorf("expected prefix:\n%s\ngot:\n%s", unifiedPrefix, got)
+				continue
+			}
+			got = got[len(unifiedPrefix):]
+			if test.unified != got {
+				t.Errorf("expected:\n%q\ngot:\n%q", test.unified, got)
+			}
+		}
+		if *verifyDiff && test.unified != "" && !test.nodiff {
+			diff, err := getDiffOutput(test.a, test.b)
+			if err != nil {
+				t.Fatal(err)
+			}
+			if diff != test.unified {
+				t.Errorf("unified:\n%q\ndiff -u:\n%q", test.unified, diff)
+			}
+		}
 	}
 }
+
+func getDiffOutput(a, b string) (string, error) {
+	fileA, err := ioutil.TempFile("", "diff.in")
+	if err != nil {
+		return "", err
+	}
+	defer os.Remove(fileA.Name())
+	if _, err := fileA.Write([]byte(a)); err != nil {
+		return "", err
+	}
+	if err := fileA.Close(); err != nil {
+		return "", err
+	}
+	fileB, err := ioutil.TempFile("", "diff.in")
+	if err != nil {
+		return "", err
+	}
+	defer os.Remove(fileB.Name())
+	if _, err := fileB.Write([]byte(b)); err != nil {
+		return "", err
+	}
+	if err := fileB.Close(); err != nil {
+		return "", err
+	}
+	cmd := exec.Command("diff", "-u", fileA.Name(), fileB.Name())
+	out, err := cmd.CombinedOutput()
+	if err != nil {
+		if _, ok := err.(*exec.ExitError); !ok {
+			return "", fmt.Errorf("failed to run diff -u %v %v: %v\n%v", fileA.Name(), fileB.Name(), err, string(out))
+		}
+	}
+	diff := string(out)
+	bits := strings.SplitN(diff, "\n", 3)
+	if len(bits) != 3 {
+		return "", fmt.Errorf("diff output did not have file prefix:\n%s", diff)
+	}
+	return bits[2], nil
+}

internal/lsp/diff/unified.go

@@ -0,0 +1,151 @@
+// Copyright 2019 The Go Authors. All rights reserved.
+// Use of this source code is governed by a BSD-style
+// license that can be found in the LICENSE file.
+
+package diff
+
+import (
+	"fmt"
+	"strings"
+)
+
+type Unified struct {
+	From, To string
+	Hunks    []*Hunk
+}
+
+type Hunk struct {
+	FromLine int
+	ToLine   int
+	Lines    []Line
+}
+
+type Line struct {
+	Kind    OpKind
+	Content string
+}
+
+const (
+	edge = 3
+	gap  = edge * 2
+)
+
+func ToUnified(from, to string, lines []string, ops []*Op) Unified {
+	u := Unified{
+		From: from,
+		To:   to,
+	}
+	if len(ops) == 0 {
+		return u
+	}
+	if lines[len(lines)-1] == "" {
+		lines = lines[:len(lines)-1]
+	}
+	var h *Hunk
+	last := -(gap + 2)
+	for _, op := range ops {
+		switch {
+		case op.I1 < last:
+			panic("cannot convert unsorted operations to unified diff")
+		case op.I1 == last:
+			//direct extension
+		case op.I1 <= last+gap:
+			//within range of previous lines, add the joiners
+			addEqualLines(h, lines, last, op.I1)
+		default:
+			//need to start a new hunk
+			if h != nil {
+				// add the edge to the previous hunk
+				addEqualLines(h, lines, last, last+edge)
+				u.Hunks = append(u.Hunks, h)
+			}
+			h = &Hunk{
+				FromLine: op.I1 + 1,
+				ToLine:   op.J1 + 1,
+			}
+			// add the edge to the new hunk
+			delta := addEqualLines(h, lines, op.I1-edge, op.I1)
+			h.FromLine -= delta
+			h.ToLine -= delta
+		}
+		last = op.I1
+		switch op.Kind {
+		case Delete:
+			for i := op.I1; i < op.I2; i++ {
+				h.Lines = append(h.Lines, Line{Kind: Delete, Content: lines[i]})
+				last++
+			}
+		case Insert:
+			for _, c := range op.Content {
+				h.Lines = append(h.Lines, Line{Kind: Insert, Content: c})
+			}
+		default:
+			// all other op types ignored
+		}
+	}
+	if h != nil {
+		// add the edge to the final hunk
+		addEqualLines(h, lines, last, last+edge)
+		u.Hunks = append(u.Hunks, h)
+	}
+	return u
+}
+
+func addEqualLines(h *Hunk, lines []string, start, end int) int {
+	delta := 0
+	for i := start; i < end; i++ {
+		if i < 0 {
+			continue
+		}
+		if i >= len(lines) {
+			return delta
+		}
+		h.Lines = append(h.Lines, Line{Kind: Equal, Content: lines[i]})
+		delta++
+	}
+	return delta
+}
+
+func (u Unified) Format(f fmt.State, r rune) {
+	fmt.Fprintf(f, "--- %s\n", u.From)
+	fmt.Fprintf(f, "+++ %s\n", u.To)
+	for _, hunk := range u.Hunks {
+		fromCount, toCount := 0, 0
+		for _, l := range hunk.Lines {
+			switch l.Kind {
+			case Delete:
+				fromCount++
+			case Insert:
+				toCount++
+			default:
+				fromCount++
+				toCount++
+			}
+		}
+		fmt.Fprint(f, "@@")
+		if fromCount > 1 {
+			fmt.Fprintf(f, " -%d,%d", hunk.FromLine, fromCount)
+		} else {
+			fmt.Fprintf(f, " -%d", hunk.FromLine)
+		}
+		if toCount > 1 {
+			fmt.Fprintf(f, " +%d,%d", hunk.ToLine, toCount)
+		} else {
+			fmt.Fprintf(f, " +%d", hunk.ToLine)
+		}
+		fmt.Fprint(f, " @@\n")
+		for _, l := range hunk.Lines {
+			switch l.Kind {
+			case Delete:
+				fmt.Fprintf(f, "-%s", l.Content)
+			case Insert:
+				fmt.Fprintf(f, "+%s", l.Content)
+			default:
+				fmt.Fprintf(f, " %s", l.Content)
+			}
+			if !strings.HasSuffix(l.Content, "\n") {
+				fmt.Fprintf(f, "\n\\ No newline at end of file\n")
+			}
+		}
+	}
+}