fmt.Fscan: use UnreadRune to preserve data across calls.

Fixes #1055.
(About as well as it can be fixed without changing the API.)

R=rsc
CC=golang-dev
https://golang.org/cl/2156047

src/pkg/fmt/doc.go

@@ -129,9 +129,13 @@
 	All arguments to be scanned must be either pointers to basic
 	types or implementations of the Scanner interface.
 
-	Note: Fscan etc.  can read one character past the input
-	they return, which means that a loop calling a scan routine
-	may skip some of the input.  This is usually a problem only
-	when there is no space between input values.
+	Note: Fscan etc. can read one character (rune) past the
+	input they return, which means that a loop calling a scan
+	routine may skip some of the input.  This is usually a
+	problem only when there is no space between input values.
+	However, if the reader provided to Fscan implements UnreadRune,
+	that method will be used to save the character and successive
+	calls will not lose data.  To attach an UnreadRune method
+	to a reader without that capability, use bufio.NewReader.
 */
 package fmt

src/pkg/fmt/scan.go

@@ -22,6 +22,14 @@ type readRuner interface {
 	ReadRune() (rune int, size int, err os.Error)
 }
 
+// unreadRuner is the interface to something that can unread runes.
+// If the object provided to Scan does not satisfy this interface,
+// a local buffer will be used to back up the input, but its contents
+// will be lost when Scan returns.
+type unreadRuner interface {
+	UnreadRune() os.Error
+}
+
 // ScanState represents the scanner state passed to custom scanners.
 // Scanners may do rune-at-a-time scanning or ask the ScanState
 // to discover the next space-delimited token.
@@ -29,7 +37,7 @@ type ScanState interface {
 	// GetRune reads the next rune (Unicode code point) from the input.
 	GetRune() (rune int, err os.Error)
 	// UngetRune causes the next call to GetRune to return the rune.
-	UngetRune(rune int)
+	UngetRune()
 	// Width returns the value of the width option and whether it has been set.
 	// The unit is Unicode code points.
 	Width() (wid int, ok bool)
@@ -133,6 +141,7 @@ type ss struct {
 	buf        bytes.Buffer // token accumulator
 	nlIsSpace  bool         // whether newline counts as white space
 	peekRune   int          // one-rune lookahead
+	prevRune   int          // last rune returned by GetRune
 	atEOF      bool         // already read EOF
 	maxWid     int          // max width of field, in runes
 	widPresent bool         // width was specified
@@ -142,10 +151,14 @@ type ss struct {
 func (s *ss) GetRune() (rune int, err os.Error) {
 	if s.peekRune >= 0 {
 		rune = s.peekRune
+		s.prevRune = rune
 		s.peekRune = -1
 		return
 	}
 	rune, _, err = s.rr.ReadRune()
+	if err == nil {
+		s.prevRune = rune
+	}
 	return
 }
 
@@ -161,11 +174,14 @@ func (s *ss) getRune() (rune int) {
 	}
 	if s.peekRune >= 0 {
 		rune = s.peekRune
+		s.prevRune = rune
 		s.peekRune = -1
 		return
 	}
 	rune, _, err := s.rr.ReadRune()
-	if err != nil {
+	if err == nil {
+		s.prevRune = rune
+	} else if err != nil {
 		if err == os.EOF {
 			s.atEOF = true
 			return EOF
@@ -198,8 +214,12 @@ func (s *ss) mustGetRune() (rune int) {
 }
 
 
-func (s *ss) UngetRune(rune int) {
-	s.peekRune = rune
+func (s *ss) UngetRune() {
+	if u, ok := s.rr.(unreadRuner); ok {
+		u.UnreadRune()
+	} else {
+		s.peekRune = s.prevRune
+	}
 }
 
 func (s *ss) error(err os.Error) {
@@ -334,7 +354,7 @@ func (s *ss) skipSpace(stopAtNewline bool) {
 			return
 		}
 		if !unicode.IsSpace(rune) {
-			s.UngetRune(rune)
+			s.UngetRune()
 			break
 		}
 	}
@@ -352,7 +372,7 @@ func (s *ss) token() string {
 			break
 		}
 		if unicode.IsSpace(rune) {
-			s.UngetRune(rune)
+			s.UngetRune()
 			break
 		}
 		s.buf.WriteRune(rune)
@@ -386,7 +406,7 @@ func (s *ss) accept(ok string) bool {
 		}
 	}
 	if rune != EOF {
-		s.UngetRune(rune)
+		s.UngetRune()
 	}
 	return false
 }
@@ -681,7 +701,7 @@ func (s *ss) hexByte() (b byte, ok bool) {
 		return
 	}
 	if unicode.IsSpace(rune1) {
-		s.UngetRune(rune1)
+		s.UngetRune()
 		return
 	}
 	rune2 := s.mustGetRune()
@@ -892,7 +912,7 @@ func (s *ss) advance(format string) (i int) {
 		}
 		inputc := s.mustGetRune()
 		if fmtc != inputc {
-			s.UngetRune(inputc)
+			s.UngetRune()
 			return -1
 		}
 		i += w

src/pkg/fmt/scan_test.go

@@ -5,6 +5,7 @@
 package fmt_test
 
 import (
+	"bufio"
 	. "fmt"
 	"io"
 	"os"
@@ -582,3 +583,24 @@ func TestEOF(t *testing.T) {
 		t.Error("expected one EOF, got", ec.eofCount)
 	}
 }
+
+// Verify that, at least when using bufio, successive calls to Fscan do not lose runes.
+func TestUnreadRuneWithBufio(t *testing.T) {
+	r := bufio.NewReader(strings.NewReader("123αb"))
+	var i int
+	var a string
+	n, err := Fscanf(r, "%d", &i)
+	if n != 1 || err != nil {
+		t.Errorf("reading int expected one item, no errors; got %d %q", n, err)
+	}
+	if i != 123 {
+		t.Errorf("expected 123; got %d", i)
+	}
+	n, err = Fscanf(r, "%s", &a)
+	if n != 1 || err != nil {
+		t.Errorf("reading string expected one item, no errors; got %d %q", n, err)
+	}
+	if a != "αb" {
+		t.Errorf("expected αb; got %q", a)
+	}
+}