parse ... differently

lex.go

@@ -1,6 +1,7 @@
 package main
 
 import (
+	"bufio"
 	"errors"
 	"fmt"
 	"io"
@@ -14,6 +15,8 @@ func (t tokenType) String() string {
 	switch t {
 	case t_error:
 		return "t_error"
+	case t_eof:
+		return "t_eof"
 	case t_string:
 		return "t_string"
 	case t_name:
@@ -22,6 +25,8 @@ func (t tokenType) String() string {
 		return "t_type"
 	case t_equals:
 		return "t_equals"
+	case t_comment:
+		return "t_comment"
 	default:
 		panic(fmt.Sprintf("unknown token type: %v", t))
 	}
@@ -29,6 +34,7 @@ func (t tokenType) String() string {
 
 const (
 	t_error   tokenType = iota // a stored lex error
+	t_eof                      // end of file token
 	t_string                   // a string literal
 	t_name                     // a name
 	t_type                     // a type
@@ -59,6 +65,7 @@ func (l *lexer) lex() {
 		switch err {
 		case nil:
 		case io.EOF:
+			l.out <- token{t_eof, ""}
 			return
 		default:
 			l.out <- token{t_error, err.Error()}
@@ -98,9 +105,9 @@ func lexString(in string) chan token {
 	return lex(r)
 }
 
-func lex(r io.RuneReader) chan token {
+func lex(r io.Reader) chan token {
 	l := lexer{
-		in:     r,
+		in:     bufio.NewReader(r),
 		out:    make(chan token),
 		backup: make([]rune, 0, 4),
 	}
@@ -131,6 +138,8 @@ func lexRoot(l *lexer) (stateFn, error) {
 		return lexRoot, nil
 	case r == '"', r == '`':
 		return lexStringLiteral(r), nil
+	case r == '#':
+		return lexComment, nil
 	case unicode.IsSpace(r):
 		return lexRoot, nil
 	case unicode.IsLower(r):
@@ -144,6 +153,26 @@ func lexRoot(l *lexer) (stateFn, error) {
 	}
 }
 
+func lexComment(l *lexer) (stateFn, error) {
+	r, err := l.next()
+	switch err {
+	case io.EOF:
+		l.emit(t_comment)
+		return nil, io.EOF
+	case nil:
+	default:
+		return nil, err
+	}
+	switch {
+	case r == '\n':
+		l.emit(t_comment)
+		return lexRoot, nil
+	default:
+		l.keep(r)
+		return lexComment, nil
+	}
+}
+
 func lexStringLiteral(delim rune) stateFn {
 	return func(l *lexer) (stateFn, error) {
 		r, err := l.next()

lex_test.go

@@ -30,6 +30,11 @@ var primitivesTests = []struct {
 		{t_equals, "="},
 		{t_string, "sam"},
 	}},
+	{`# this is a comment`, []token{{t_comment, " this is a comment"}}},
+	{`
+    # comment line one
+    # comment line two
+    `, []token{{t_comment, " comment line one"}, {t_comment, " comment line two"}}},
 }
 
 func TestLexPrimities(t *testing.T) {
@@ -39,6 +44,7 @@ func TestLexPrimities(t *testing.T) {
 			t.Error(err)
 			continue
 		}
+		tokens = tokens[:len(tokens)-1]
 		if len(tokens) != len(test.out) {
 			t.Errorf("expected %d token, saw %d: %v", len(test.out), len(tokens), tokens)
 			continue

node.go

@@ -0,0 +1,128 @@
+package main
+
+import (
+	"bytes"
+	"fmt"
+)
+
+type nodeType int
+
+const (
+	n_error nodeType = iota
+	n_root
+	n_comment
+	n_assignment
+)
+
+type node interface {
+	Type() nodeType
+	parse(*parser) error
+}
+
+type rootNode struct {
+	children []node
+}
+
+func newRootNode() node {
+	return &rootNode{children: make([]node, 0, 8)}
+}
+
+func (n *rootNode) Type() nodeType {
+	return n_root
+}
+
+func (n *rootNode) parse(p *parser) error {
+	for {
+		t := p.next()
+		switch t.t {
+		case t_error:
+			return fmt.Errorf("parse error: saw lex error while parsing root node: %v", t.s)
+		case t_eof:
+			return nil
+		case t_comment:
+			shit := commentNode(t.s)
+			n.addChild(&shit)
+		case t_name:
+			nn := &assignmentNode{name: t.s}
+			if err := nn.parse(p); err != nil {
+				return err
+			}
+			n.addChild(nn)
+		default:
+			return fmt.Errorf("parse error: unexpected token type %v while parsing root node", t.t)
+		}
+	}
+}
+
+func (n *rootNode) addChild(child node) {
+	if n.children == nil {
+		n.children = make([]node, 0, 8)
+	}
+	n.children = append(n.children, child)
+}
+
+func (n *rootNode) String() string {
+	var buf bytes.Buffer
+	buf.WriteString("{")
+	for _, child := range n.children {
+		fmt.Fprintf(&buf, "%s, ", child)
+	}
+	if len(n.children) > 0 {
+		buf.Truncate(buf.Len() - 2)
+	}
+	buf.WriteString("}")
+	return buf.String()
+}
+
+type commentNode string
+
+func (n commentNode) Type() nodeType {
+	return n_comment
+}
+
+func (n commentNode) parse(p *parser) error {
+	return nil
+}
+
+func (n commentNode) String() string {
+	return fmt.Sprintf("{comment: %s}", string(n))
+}
+
+type assignmentNode struct {
+	name  string
+	value interface{}
+}
+
+func (n assignmentNode) Type() nodeType {
+	return n_assignment
+}
+
+func (n *assignmentNode) parse(p *parser) error {
+	t := p.next()
+	switch t.t {
+	case t_error:
+		return fmt.Errorf("parse error: saw lex error while parsing assignment node: %v", t.s)
+	case t_eof:
+		return fmt.Errorf("parse error: unexpected eof in assignment node")
+	case t_equals:
+	default:
+		return fmt.Errorf("parse error: unexpected %v token after name, expected =", t.t)
+	}
+
+	t = p.next()
+	switch t.t {
+	case t_error:
+		return fmt.Errorf("parse error: saw lex error while parsing assignment node: %v", t.s)
+	case t_eof:
+		return fmt.Errorf("parse error: unexpected eof in assignment node")
+	case t_string:
+		n.value = t.s
+		return nil
+	default:
+		return fmt.Errorf("parse error: unexpected %v token after =, expected some kind of value", t.t)
+	}
+}
+
+func (n *assignmentNode) String() string {
+	return fmt.Sprintf("{assign: name=%s, val=%s}", n.name, n.value)
+}

parse.go

@@ -1,134 +1,42 @@
 package main
 
 import (
-	"fmt"
 	"io"
 )
 
-const (
+const ()
-	e_no_error parseErrorType = iota
-	e_lex_error
-	e_unexpected_eof
-	e_unexpected_token
-)
-
-type parseErrorType int
-
-type parseError struct {
-	t parseErrorType
-	m string
-}
-
-func (p parseError) Error() string {
-	return fmt.Sprintf("parse error: %s", p.m)
-}
-
-func parseErrorf(t parseErrorType, tpl string, args ...interface{}) error {
-	return parseError{t: t, m: fmt.Sprintf(tpl, args...)}
-}
-
-type parseFn func(*parser) (parseFn, error)
 
-func parseRoot(p *parser) (parseFn, error) {
+func parse(r io.Reader) (node, error) {
-	if err := p.next(); err != nil {
+	p := &parser{
-		return nil, err
+		root:   newRootNode(),
-	}
+		input:  lex(r),
-	switch p.cur.t {
+		backup: make([]token, 0, 8),
-	case t_name:
-		return parseAfterName(p.cur.s), nil
-	default:
-		return nil, parseErrorf(e_unexpected_token, "unexpected %s token in parseRoot", p.cur.t)
-	}
-}
-
-func parseAfterName(name string) parseFn {
-	return func(p *parser) (parseFn, error) {
-		switch err := p.next(); err {
-		case io.EOF:
-			return nil, parseErrorf(e_unexpected_eof, "unexpected eof after name %s", name)
-		case nil:
-		default:
-			return nil, err
-		}
-
-		switch p.cur.t {
-		case t_equals:
-			return parseAssign(name), nil
-		default:
-			return nil, parseErrorf(e_unexpected_token, "unexpected %s token in parseAfterName", p.cur.t)
-		}
 	}
-}
+	if err := p.parse(); err != nil {
-
-func parseAssign(name string) parseFn {
-	return func(p *parser) (parseFn, error) {
-		switch err := p.next(); err {
-		case io.EOF:
-			return nil, parseErrorf(e_unexpected_eof, "unexpected eof when trying to parse value for name %s", name)
-		case nil:
-		default:
 		return nil, err
 	}
-
+	return p.root, nil
-		switch p.cur.t {
-		case t_string:
-			p.out.setUnique(name, p.cur.s)
-			return parseRoot, nil
-		default:
-			return nil, parseErrorf(e_unexpected_token, "unexpected %s token in parseAssign", p.cur.t)
-		}
-	}
 }
 
 type parser struct {
-	in  chan token
+	root   node
-	cur token
+	input  chan token
-	out *Config
+	backup []token
-}
-
-func (p *parser) next() error {
-	t, ok := <-p.in
-	if !ok {
-		return io.EOF
-	}
-	if t.t == t_error {
-		return parseError{e_lex_error, t.s}
-	}
-	p.cur = t
-	return nil
 }
 
-func (p *parser) run() error {
+func (p *parser) parse() error {
-	fn := parseRoot
+	if p.root == nil {
-	var err error
+		p.root = newRootNode()
-	for {
-		fn, err = fn(p)
-		switch err {
-		case io.EOF:
-			return nil
-		case nil:
-		default:
-			return err
-		}
 	}
+	return p.root.parse(p)
 }
 
-type assignment struct {
+// returns the next token and advances the input stream
-	name  string
+func (p *parser) next() token {
-	value interface{}
+	if len(p.backup) > 0 {
-}
+		t := p.backup[len(p.backup)-1]
-
+		p.backup = p.backup[:len(p.backup)-1]
-func parse(in chan token) (*Config, error) {
+		return t
-	p := &parser{
-		in:  in,
-		out: new(Config),
 	}
-	if err := p.run(); err != nil {
+	return <-p.input
-		return nil, err
-	}
-	return p.out, nil
-}
-
-func parseString(in string) (*Config, error) {
-	return parse(lexString(in))
 }

parse_test.go

@@ -1,127 +1,76 @@
 package main
 
 import (
+	"strings"
 	"testing"
 )
 
 var parseTests = []parseTest{
 	{
-		in:   ``,
+		source: ``,
-		desc: "an empty string is a valid config",
+		root: &rootNode{
-		configTests: []configTest{
+			children: []node{},
+		},
+	},
 	{
-				desc: "undefined name field should not exist",
+		source: `# just a comment`,
-				pass: inv(hasKey("name")),
+		root: &rootNode{
+			children: []node{
+				commentNode(" just a comment"),
 			},
 		},
 	},
 	{
-		in:        `name `,
+		source: `name = "jordan"`,
-		desc:      "eof after name",
+		root: &rootNode{
-		errorType: e_unexpected_eof,
+			children: []node{
+				&assignmentNode{
+					name:  "name",
+					value: "jordan",
+				},
 			},
-	{
-		in:        `firstname lastname`,
-		desc:      "two names in a row",
-		errorType: e_unexpected_token,
 		},
-	{
-		in:        `name = `,
-		desc:      "eof after equals",
-		errorType: e_unexpected_eof,
 	},
 	{
-		in:   `name = "jordan"`,
+		source: `
-		desc: "assign a value",
+        first_name = "jordan"
-		configTests: []configTest{
+        last_name = "orelli"
-			{
+        `,
-				desc: "should have name",
+		root: &rootNode{},
-				pass: hasValue("name", "jordan"),
 	},
+	{
+		source: `
+        # personal info
+        first_name = "jordan"
+        last_name = "orelli"
+        `,
+		root: &rootNode{},
 	},
+	{
+		source: `
+        first_name = "jordan" # yep, that's my name
+        last_name = "orelli"  # comments should be able to follow other shit
+        `,
+		root: &rootNode{},
 	},
 }
 
-// a boolean statement about a config struct
-type configPredicate func(*Config) bool
-
-// a suite of tests for parsing potm input
 type parseTest struct {
-	in          string
+	source string
-	desc        string
+	root   *rootNode
-	configTests []configTest
-	errorType   parseErrorType
 }
 
 func (p *parseTest) run(t *testing.T) {
-	c, err := parseString(p.in)
+	r := strings.NewReader(p.source)
+	n, err := parse(r)
 	if err != nil {
-		t.Logf("test %s has error %v", p.desc, err)
+		t.Errorf("parse error: %v", err)
-		e, ok := err.(parseError)
-		if !ok {
-			t.Errorf("unexpected error: %s", e)
-			return
-		}
-		if p.errorType == e.t {
-			t.Logf("OK: got expected error type %v for %s", e.t, p.desc)
-		} else {
-			t.Errorf("unexpected parse error: %s", e)
 		return
 	}
+	if n.Type() != n_root {
+		t.Errorf("we expected a root node object, but instead we got: %s", n.Type())
 	}
-	t.Logf("parsed config for %s", p.desc)
+	t.Logf("output: %v", n)
-	t.Log(c)
-	p.runConfigTests(t, c)
-}
-
-func (p *parseTest) runConfigTests(t *testing.T, c *Config) {
-	ok := true
-	for _, test := range p.configTests {
-		if test.pass(c) {
-			t.Logf("OK: %s", test.desc)
-		} else {
-			t.Errorf("config predicate failed: %s", test.desc)
-			ok = false
-		}
-	}
-	if ok {
-		t.Logf("OK: %s", p.desc)
-	}
-}
-
-// an individual test for confirming that a parsed config struct meets some
-// predicate
-type configTest struct {
-	desc string
-	pass configPredicate
-}
-
-// inverts a given config predicate
-func inv(fn configPredicate) configPredicate {
-	return func(c *Config) bool {
-		return !fn(c)
-	}
-}
-
-func hasKey(s string) configPredicate {
-	return func(c *Config) bool {
-		return c.hasKey(s)
-	}
-}
-
-func hasValue(key string, expected interface{}) configPredicate {
-	switch t := expected.(type) {
-	case string:
-		return hasStringValue(key, t)
-	default:
-		panic("no we can't do that yet")
-	}
-}
-
-func hasStringValue(key string, expected string) configPredicate {
-	return func(c *Config) bool {
-		return c.GetString(key) == expected
-	}
 }
 
 func TestParse(t *testing.T) {