Return preamble content when parsing files

This allows callers to provide patches with commit or email headers and
then retrieve that leading content for additional parsing without
having to identify where the first file in the patch starts.

Author: bluekeyes

gitdiff/parser.go

@@ -8,25 +8,24 @@ import (
 	"strings"
 )
 
-// Parse parses a patch with changes for one or more files. Any content
-// preceding the first file header is ignored. If an error occurs while
-// parsing, files will contain all files parsed before the error.
-func Parse(r io.Reader) ([]*File, error) {
+// Parse parses a patch with changes to one or more files. Any content before
+// the first file is returned as the second value. If an error occurs while
+// parsing, it returns all files parsed before the error.
+func Parse(r io.Reader) ([]*File, string, error) {
 	p := &parser{r: bufio.NewReader(r)}
 	if err := p.Next(); err != nil {
 		if err == io.EOF {
-			return nil, nil
+			return nil, "", nil
 		}
-		return nil, err
+		return nil, "", err
 	}
 
-	// TODO(bkeyes): capture non-file lines in between files
-
+	var preamble string
 	var files []*File
 	for {
-		file, err := p.ParseNextFileHeader()
+		file, pre, err := p.ParseNextFileHeader()
 		if err != nil {
-			return files, err
+			return files, preamble, err
 		}
 		if file == nil {
 			break
@@ -38,17 +37,20 @@ func Parse(r io.Reader) ([]*File, error) {
 		} {
 			n, err := fn(file)
 			if err != nil {
-				return files, err
+				return files, preamble, err
 			}
 			if n > 0 {
 				break
 			}
 		}
-		// file has fragment(s) from above or the patch is empty or invalid
+
+		if len(files) == 0 {
+			preamble = pre
+		}
 		files = append(files, file)
 	}
 
-	return files, nil
+	return files, preamble, nil
 }
 
 // TODO(bkeyes): consider exporting the parser type with configuration
@@ -71,9 +73,11 @@ type parser struct {
 	lines  [3]string
 }
 
-// ParseNextFileHeader finds and parses the next file header in the stream. It
-// returns nil if no headers are found before the end of the stream.
-func (p *parser) ParseNextFileHeader() (*File, error) {
+// ParseNextFileHeader finds and parses the next file header in the stream. If
+// a header is found, it returns a file and all input before the header. It
+// returns nil if no headers are found before the end of the input.
+func (p *parser) ParseNextFileHeader() (*File, string, error) {
+	var preamble strings.Builder
 	var file *File
 	for {
 		// check for disconnected fragment headers (corrupt patch)
@@ -83,36 +87,37 @@ func (p *parser) ParseNextFileHeader() (*File, error) {
 			goto NextLine
 		}
 		if frag != nil {
-			return nil, p.Errorf(-1, "patch fragment without file header: %s", frag.Header())
+			return nil, "", p.Errorf(-1, "patch fragment without file header: %s", frag.Header())
 		}
 
 		// check for a git-generated patch
 		file, err = p.ParseGitFileHeader()
 		if err != nil {
-			return nil, err
+			return nil, "", err
 		}
 		if file != nil {
-			return file, nil
+			return file, preamble.String(), nil
 		}
 
 		// check for a "traditional" patch
 		file, err = p.ParseTraditionalFileHeader()
 		if err != nil {
-			return nil, err
+			return nil, "", err
 		}
 		if file != nil {
-			return file, nil
+			return file, preamble.String(), nil
 		}
 
 	NextLine:
+		preamble.WriteString(p.Line(0))
 		if err := p.Next(); err != nil {
 			if err == io.EOF {
 				break
 			}
-			return nil, err
+			return nil, "", err
 		}
 	}
-	return nil, nil
+	return nil, "", nil
 }
 
 // ParseTextFragments parses text fragments until the next file header or the

gitdiff/parser_test.go

@@ -133,17 +133,18 @@ context line
 			}
 
 			if test.EndLine != p.Line(0) {
-				t.Errorf("incorrect position after parsing\nexpected: %q\nactual: %q", test.EndLine, p.Line(0))
+				t.Errorf("incorrect position after parsing\nexpected: %q\n  actual: %q", test.EndLine, p.Line(0))
 			}
 		})
 	}
 }
 
 func TestParseNextFileHeader(t *testing.T) {
 	tests := map[string]struct {
-		Input  string
-		Output *File
-		Err    bool
+		Input    string
+		Output   *File
+		Preamble string
+		Err      bool
 	}{
 		"gitHeader": {
 			Input: `commit 1acbae563cd6ef5750a82ee64e116c6eb065cb94
@@ -165,6 +166,13 @@ index cc34da1..1acbae5 100644
 				OldOIDPrefix: "cc34da1",
 				NewOIDPrefix: "1acbae5",
 			},
+			Preamble: `commit 1acbae563cd6ef5750a82ee64e116c6eb065cb94
+Author:	Morton Haypenny <mhaypenny@example.com>
+Date:	Tue Apr 2 22:30:00 2019 -0700
+
+    This is a sample commit message.
+
+`,
 		},
 		"traditionalHeader": {
 			Input: `
@@ -176,6 +184,7 @@ index cc34da1..1acbae5 100644
 				OldName: "file.txt",
 				NewName: "file.txt",
 			},
+			Preamble: "\n",
 		},
 		"noHeaders": {
 			Input: `
@@ -184,7 +193,8 @@ this is another line
 --- could this be a header?
 nope, it's just some dashes
 `,
-			Output: nil,
+			Output:   nil,
+			Preamble: "",
 		},
 		"detatchedFragmentLike": {
 			Input: `
@@ -206,7 +216,7 @@ a wild fragment appears?
 		t.Run(name, func(t *testing.T) {
 			p := newTestParser(test.Input, true)
 
-			f, err := p.ParseNextFileHeader()
+			f, pre, err := p.ParseNextFileHeader()
 			if test.Err {
 				if err == nil || err == io.EOF {
 					t.Fatalf("expected error parsing next file header, but got %v", err)
@@ -217,8 +227,11 @@ a wild fragment appears?
 				t.Fatalf("unexpected error parsing next file header: %v", err)
 			}
 
+			if test.Preamble != pre {
+				t.Errorf("incorrect preamble\nexpected: %q\n  actual: %q", test.Preamble, pre)
+			}
 			if !reflect.DeepEqual(test.Output, f) {
-				t.Errorf("incorrect file\nexpected: %+v\nactual: %+v", test.Output, f)
+				t.Errorf("incorrect file\nexpected: %+v\n  actual: %+v", test.Output, f)
 			}
 		})
 	}
@@ -266,9 +279,20 @@ func TestParse(t *testing.T) {
 		},
 	}
 
+	expectedPreamble := `commit 5d9790fec7d95aa223f3d20936340bf55ff3dcbe
+Author: Morton Haypenny <mhaypenny@example.com>
+Date:   Tue Apr 2 22:55:40 2019 -0700
+
+    A file with multiple fragments.
+
+    The content is arbitrary.
+
+`
+
 	tests := map[string]struct {
 		InputFile string
 		Output    []*File
+		Preamble  string
 		Err       bool
 	}{
 		"oneFile": {
@@ -283,6 +307,7 @@ func TestParse(t *testing.T) {
 					Fragments:    expectedFragments,
 				},
 			},
+			Preamble: expectedPreamble,
 		},
 		"twoFiles": {
 			InputFile: "testdata/two_files.patch",
@@ -304,6 +329,7 @@ func TestParse(t *testing.T) {
 					Fragments:    expectedFragments,
 				},
 			},
+			Preamble: expectedPreamble,
 		},
 	}
 
@@ -314,7 +340,7 @@ func TestParse(t *testing.T) {
 				t.Fatalf("unexpected error opening input file: %v", err)
 			}
 
-			files, err := Parse(f)
+			files, pre, err := Parse(f)
 			if test.Err {
 				if err == nil || err == io.EOF {
 					t.Fatalf("expected error parsing patch, but got %v", err)
@@ -328,11 +354,14 @@ func TestParse(t *testing.T) {
 			if len(test.Output) != len(files) {
 				t.Fatalf("incorrect number of parsed files: expected %d, actual %d", len(test.Output), len(files))
 			}
+			if test.Preamble != pre {
+				t.Errorf("incorrect preamble\nexpected: %q\n  actual: %q", test.Preamble, pre)
+			}
 			for i := range test.Output {
 				if !reflect.DeepEqual(test.Output[i], files[i]) {
 					exp, _ := json.MarshalIndent(test.Output[i], "", "  ")
 					act, _ := json.MarshalIndent(files[i], "", "  ")
-					t.Errorf("incorrect file at position %d\nexpected: %s\nactual: %s", i, exp, act)
+					t.Errorf("incorrect file at position %d\nexpected: %s\n  actual: %s", i, exp, act)
 				}
 			}
 		})

gitdiff/testdata/one_file.patch

@@ -2,7 +2,7 @@ commit 5d9790fec7d95aa223f3d20936340bf55ff3dcbe
 Author: Morton Haypenny <mhaypenny@example.com>
 Date:   Tue Apr 2 22:55:40 2019 -0700
 
-    A single file with multiple fragments.
+    A file with multiple fragments.
 
     The content is arbitrary.