*: parsing a spec to a Hierarchy struct

Signed-off-by: Vincent Batts <vbatts@hashbangbash.com>

hierarchy.go

@@ -0,0 +1,65 @@
+package mtree
+
+import (
+	"fmt"
+	"io"
+	"sort"
+	"strings"
+)
+
+// DirectoryHierarchy is the mapped structure for an mtree directory hierarchy
+// spec
+type DirectoryHierarchy struct {
+	Entries []Entry
+}
+
+// WriteTo simplifies the output of the resulting hierarchy spec
+func (dh DirectoryHierarchy) WriteTo(w io.Writer) (n int64, err error) {
+	sort.Sort(byPos(dh.Entries))
+	var sum int64
+	for _, e := range dh.Entries {
+		i, err := io.WriteString(w, e.String()+"\n")
+		if err != nil {
+			return sum, err
+		}
+		sum += int64(i)
+	}
+	return sum, nil
+}
+
+type byPos []Entry
+
+func (bp byPos) Len() int           { return len(bp) }
+func (bp byPos) Less(i, j int) bool { return bp[i].Pos < bp[j].Pos }
+func (bp byPos) Swap(i, j int)      { bp[i], bp[j] = bp[j], bp[i] }
+
+// Entry is each component of content in the mtree spec file
+type Entry struct {
+	Pos      int      // order in the spec
+	Raw      string   // file or directory name
+	Name     string   // file or directory name
+	Keywords []string // TODO(vbatts) maybe a keyword typed set of values?
+	Type     EntryType
+}
+
+func (e Entry) String() string {
+	if e.Raw != "" {
+		return e.Raw
+	}
+	// TODO(vbatts) if type is RelativeType and a keyword of not type=dir
+	return fmt.Sprintf("%s %s", e.Name, strings.Join(e.Keywords, " "))
+}
+
+// EntryType are the formats of lines in an mtree spec file
+type EntryType int
+
+// The types of lines to be found in an mtree spec file
+const (
+	SignatureType EntryType = iota // first line of the file, like `#mtree v2.0`
+	BlankType                      // blank lines are ignored
+	CommentType                    // Lines beginning with `#` are ignored
+	SpecialType                    // line that has `/` prefix issue a "special" command (currently only /set and /unset)
+	RelativeType                   // if the first white-space delimited word does not have a '/' in it. Options/keywords are applied.
+	DotDotType                     // .. - A relative path step. keywords/options are ignored
+	FullType                       // if the first word on the line has a `/` after the first character, it interpretted as a file pathname with options
+)

mtree.go

@@ -1,58 +0,0 @@
-package mtree
-
-import "io"
-
-// Positioner responds with the newline delimited position
-type Positioner interface {
-	Pos() int
-}
-
-// DirectoryHierarchy is the mapped structure for an mtree directory hierarchy
-// spec
-type DirectoryHierarchy struct {
-	Comments []Comment
-	Entries  []Entry
-}
-
-// WriteTo simplifies the output of the resulting hierarchy spec
-func (dh DirectoryHierarchy) WriteTo(w io.Writer) (n int64, err error) {
-	return 0, nil
-}
-
-// Comment stores arbitrary metadata for the spec. Sometimes "user", "machine",
-// "tree", and "date". But most of the time, it includes the relative path of
-// the directory being stepped into. Or a "signature" like `#mtree v2.0`,
-type Comment struct {
-	Position int
-	Str      string
-	// TODO(vbatts) include a comment line parser
-}
-
-// Pos returns the line of this comment
-func (c Comment) Pos() int {
-	return c.Position
-}
-
-// Entry is each component of content in the mtree spec file
-type Entry struct {
-	Position int      // order in the spec
-	Name     string   // file or directory name
-	Keywords []string // TODO(vbatts) maybe a keyword typed set of values?
-	str      string   // raw string. needed?
-	Type     EntryType
-}
-
-// Pos returns the line of this comment
-func (e Entry) Pos() int {
-	return e.Position
-}
-
-type EntryType int
-
-const (
-	SpecialType   int = iota // line that has `/` prefix issue a "special" command (currently only /set and /unset)
-	FileType                 // indented line
-	DirectoryType            // ^ matched line, that is not /set
-	PathStepType             // .. - keywords/options are ignored
-	FullType                 // if the first word on the line has a `/` after the first character, it interpretted as a file pathname with options
-)

mtree_test.go

@@ -1,11 +1,8 @@
 package mtree
 
 import (
-	"bufio"
 	"fmt"
-	"log"
 	"os"
-	"strings"
 	"testing"
 )
 
@@ -18,24 +15,22 @@ func TestParser(t *testing.T) {
 		func() {
 			fh, err := os.Open(file)
 			if err != nil {
-				log.Println(err)
+				t.Error(err)
 				return
 			}
 			defer fh.Close()
 
-			s := bufio.NewScanner(fh)
+			dh, err := ParseSpec(fh)
-			for s.Scan() {
+			if err != nil {
-				str := s.Text()
+				t.Error(err)
-				switch {
-				case strings.HasPrefix(str, "#"):
-					continue
-				default:
 			}
-				fmt.Printf("%q\n", str)
+			fmt.Printf("%q", dh)
-			}
+
-			if err := s.Err(); err != nil {
+			_, err = dh.WriteTo(os.Stdout)
-				log.Println("ERROR:", err)
+			if err != nil {
+				t.Error(err)
 			}
+
 		}()
 	}
 }

parse.go

@@ -0,0 +1,77 @@
+package mtree
+
+import (
+	"bufio"
+	"io"
+	"strings"
+)
+
+func ParseSpec(r io.Reader) (*DirectoryHierarchy, error) {
+	s := bufio.NewScanner(r)
+	i := int(0)
+	dh := DirectoryHierarchy{}
+	for s.Scan() {
+		str := s.Text()
+		e := Entry{Pos: i}
+		switch {
+		case strings.HasPrefix(str, "#"):
+			e.Raw = str
+			if strings.HasPrefix(str, "#mtree") {
+				e.Type = SignatureType
+			} else {
+				e.Type = CommentType
+				// from here, the comment could be "# key: value" metadata
+				// or a relative path hint
+			}
+		case str == "":
+			e.Type = BlankType
+			// nothing else to do here
+		case strings.HasPrefix(str, "/"):
+			e.Type = SpecialType
+			// collapse any escaped newlines
+			for {
+				if strings.HasSuffix(str, `\`) {
+					str = str[:len(str)-1]
+					s.Scan()
+					str += s.Text()
+				} else {
+					break
+				}
+			}
+			// parse the options
+			f := strings.Fields(str)
+			e.Name = f[0]
+			e.Keywords = f[1:]
+		case len(strings.Fields(str)) > 0 && strings.Fields(str)[0] == "..":
+			e.Type = DotDotType
+			e.Raw = str
+			// nothing else to do here
+		case len(strings.Fields(str)) > 0:
+			// collapse any escaped newlines
+			for {
+				if strings.HasSuffix(str, `\`) {
+					str = str[:len(str)-1]
+					s.Scan()
+					str += s.Text()
+				} else {
+					break
+				}
+			}
+			// parse the options
+			f := strings.Fields(str)
+			if strings.Contains(str, "/") {
+				e.Type = FullType
+			} else {
+				e.Type = RelativeType
+			}
+			e.Name = f[0]
+			e.Keywords = f[1:]
+		default:
+			// TODO(vbatts) log a warning?
+			continue
+		}
+		dh.Entries = append(dh.Entries, e)
+		i++
+	}
+	return &dh, s.Err()
+}