Arbitrary checkpoint

Apparently a gofmt happened in the meantime.

assembler.go

@@ -1,39 +1,37 @@
 package main
 
 import (
-	"errors"
-	"io"
-	"fmt"
 	"bufio"
-	"strings"
+	"errors"
+	"fmt"
+	"io"
 	"strconv"
+	"strings"
 )
 
 const (
-	WORD = iota          // [A-Za-z_-]+
+	WORD    = iota // [A-Za-z_-]+
-	NUMBER               // [0-9]+
+	NUMBER         // [0-9]+
-	NEWLINE              // \n
+	NEWLINE        // \n
-	ERROR                // Error
+	ERROR          // Error
 )
 
 type location struct {
-	line int
+	line   int
 	column int
 }
 
 type token struct {
-	location location    // Position of the token
+	location location // Position of the token
-	value string         // Text content of the token
+	value    string   // Text content of the token
-	kind int             // Kind of the token
+	kind     int      // Kind of the token
 }
 
 type tokenizer struct {
-	line int             // Current line
+	location location      // Current position
-	column int           // Current column
+	value    []byte        // Current token string
-
+	reader   *bufio.Reader // Reader
-	value []byte         // Current token string
+	tokens   chan<- token  // Output token channel
-	reader *bufio.Reader // Reader
-	tokens chan<- token  // Token channel
 }
 
 // -----------------------------------------------------------------------------
@@ -47,7 +45,9 @@ func isNumber(c byte) bool {
 }
 
 func isWordHead(c byte) bool {
-	if c >= 'a' && c <= 'z' { c -= 32 }
+	if c >= 'a' && c <= 'z' {
+		c -= 32
+	}
 	return c >= 'a' && c <= 'z' || c == '_'
 }
 
@@ -57,12 +57,13 @@ func isWordTail(c byte) bool {
 
 // -----------------------------------------------------------------------------
 
-func (t *tokenizer) send(kind int) {
+func (t *tokenizer) send(start location, kind int) {
-	// FIXME: track the beginning of the token
+	t.tokens <- token{start, string(t.value), kind}
-	t.tokens <- token{location{t.line, t.column}, string(t.value), kind}
 	t.value = []byte{}
 }
 
+// XXX: the handling could probably be simplified by extending the "byte"
+//   to also include a special value for io.EOF and other errors
 func (t *tokenizer) peek() (byte, error) {
 	buf, err := t.reader.Peek(1)
 	return buf[0], err
@@ -70,13 +71,15 @@ func (t *tokenizer) peek() (byte, error) {
 
 func (t *tokenizer) eat() (byte, error) {
 	c, err := t.reader.ReadByte()
-	if err != nil { return 0, err }
+	if err != nil {
+		return 0, err
+	}
 
 	if c == '\n' {
-		t.line++
+		t.location.line++
-		t.column = 0
+		t.location.column = 0
 	} else {
-		t.column++
+		t.location.column++
 	}
 	return c, nil
 }
@@ -87,63 +90,88 @@ func (t *tokenizer) step() error {
 	t.value = []byte{}
 	c, err := t.peek()
 
-	if err == io.EOF { return nil }
+	if err == io.EOF {
-	if err != nil { return err }
+		return nil
+	}
+	if err != nil {
+		return err
+	}
 
+	start := t.location
 	switch {
 	case isSpace(c):
-		t.eat()
+		c, err = t.eat()
 	case c == '\n':
 		c, err = t.eat()
 		t.value = append(t.value, c)
 
-		t.send(NEWLINE)
+		t.send(start, NEWLINE)
 	case isNumber(c):
 		c, err = t.eat()
 		t.value = append(t.value, c)
 
 		for {
 			c, err = t.peek()
-			if err == io.EOF { break }
+			if err == io.EOF {
-			if err != nil { return err }
+				break
+			}
+			if err != nil {
+				return err
+			}
 
-			if !isNumber(c) { break }
+			if !isNumber(c) {
+				break
+			}
 
-			c, _ = t.eat()
+			c, err = t.eat()
 			t.value = append(t.value, c)
 		}
-		t.send(NUMBER)
+		t.send(start, NUMBER)
 	case isWordHead(c):
 		c, err = t.eat()
 		t.value = append(t.value, c)
 
 		for {
 			c, err = t.peek()
-			if err == io.EOF { break }
+			if err == io.EOF {
-			if err != nil { return err }
+				break
+			}
+			if err != nil {
+				return err
+			}
 
-			if !isWordTail(c) { break }
+			if !isWordTail(c) {
+				break
+			}
 
-			c, _ = t.eat()
+			c, err = t.eat()
 			t.value = append(t.value, c)
 		}
-		t.send(WORD)
+		t.send(start, WORD)
 	case c == '/':
 		c, err = t.eat()
 		t.value = append(t.value, c)
 
 		c, err = t.peek()
-		if err != nil { return err }
+		if err != nil {
+			return err
+		}
 
 		if c != '/' {
 			return errors.New("unrecognized input")
 		}
 		for {
 			c, err = t.peek()
-			if err == io.EOF { break }
+			if err == io.EOF {
-			if err != nil { return err }
+				break
+			}
+			if err != nil {
+				return err
+			}
 
-			if c == '\n' { break }
+			if c == '\n' {
+				break
+			}
 			t.eat()
 		}
 	default:
@@ -154,10 +182,9 @@ func (t *tokenizer) step() error {
 
 func tokenize(r io.Reader, tokens chan<- token) {
 	t := tokenizer{
-		line: 1,
+		location: location{line: 1, column: 0},
-		column: 0,
+		tokens:   tokens,
-		tokens: tokens,
+		reader:   bufio.NewReader(r),
-		reader: bufio.NewReader(r),
 	}
 	for {
 		err := t.step()
@@ -165,9 +192,8 @@ func tokenize(r io.Reader, tokens chan<- token) {
 			break
 		}
 		if err != nil {
-			t.tokens <- token{location{t.line, t.column},
+			t.tokens <- token{t.location, fmt.Sprintf("line %d, column %d: %s",
-				fmt.Sprintf("line %d, column %d: %s",
+				t.location.line, t.location.column, err.Error()), ERROR}
-					t.line, t.column, err.Error()), ERROR}
 			break
 		}
 	}
@@ -190,7 +216,7 @@ const (
 	IDATA
 )
 
-var instructions = map[string]int {
+var instructions = map[string]int{
 	"HLT": IHALT,
 	"COB": IHALT,
 	"ADD": IADD,
@@ -206,7 +232,7 @@ var instructions = map[string]int {
 }
 
 type instruction struct {
-	id int
+	id     int
 	target string
 	number int
 }
@@ -221,7 +247,9 @@ type assembler struct {
 
 func (a *assembler) step() (bool, error) {
 	token, ok := <-a.tokens
-	if !ok { return false, nil }
+	if !ok {
+		return false, nil
+	}
 
 	// TODO: add token location information to returned errors
 
@@ -297,7 +325,7 @@ func Assemble(r io.Reader) (code []int16, err error) {
 		}
 	}
 
-	for _, x := range(a.output) {
+	for _, x := range a.output {
 		n := x.id * 100
 		if len(x.target) != 0 {
 			if resolved, ok := a.labels[x.target]; !ok {

main.go

@@ -1,8 +1,8 @@
 package main
 
 import (
-	"os"
 	"fmt"
+	"os"
 )
 
 func main() {