1
0
mirror of https://github.com/golang/go synced 2024-11-20 10:54:49 -07:00
go/usr/gri/pretty/parser.go

1692 lines
34 KiB
Go
Raw Normal View History

// Copyright 2009 The Go Authors. All rights reserved.
// Use of this source code is governed by a BSD-style
// license that can be found in the LICENSE file.
// A parser for Go source text. The input is a stream of lexical tokens
// provided via the Scanner interface. The output is an abstract syntax
// tree (AST) representing the Go source.
//
// A client may parse the entire program (ParseProgram), only the package
// clause (ParsePackageClause), or the package clause and the import
// declarations (ParseImportDecls).
//
package Parser
import (
"fmt";
"vector";
"token";
"scanner";
"ast";
)
// A Parser holds the parser's internal state while processing
// a given text. It can be allocated as part of another data
// structure but must be initialized via Init before use.
//
2009-01-20 15:40:40 -07:00
type Parser struct {
scanner *scanner.Scanner;
err scanner.ErrorHandler;
// Tracing/debugging
trace bool;
indent uint;
comments vector.Vector;
last_comment ast.CommentGroup;
// The next token
loc scanner.Location; // token location
tok int; // one token look-ahead
val []byte; // token value
// Non-syntactic parser control
opt_semi bool; // true if semicolon separator is optional in statement list
expr_lev int; // < 0: in control clause, >= 0: in expression
};
// When we don't have a location use noloc.
// TODO make sure we always have a location.
var noloc scanner.Location;
// ----------------------------------------------------------------------------
// Helper functions
func unreachable() {
panic("unreachable");
}
// ----------------------------------------------------------------------------
// Parsing support
func (P *Parser) printIndent() {
i := P.indent;
// reduce printing time by a factor of 2 or more
for ; i > 10; i -= 10 {
fmt.Printf(". . . . . . . . . . ");
}
for ; i > 0; i-- {
fmt.Printf(". ");
}
}
func trace(P *Parser, msg string) *Parser {
P.printIndent();
fmt.Printf("%s (\n", msg);
P.indent++;
return P;
}
func un/*trace*/(P *Parser) {
P.indent--;
P.printIndent();
fmt.Printf(")\n");
}
func (P *Parser) next0() {
P.loc, P.tok, P.val = P.scanner.Scan();
P.opt_semi = false;
if P.trace {
P.printIndent();
switch P.tok {
case token.IDENT, token.INT, token.FLOAT, token.CHAR, token.STRING:
fmt.Printf("%d:%d: %s = %s\n", P.loc.Line, P.loc.Col, token.TokenString(P.tok), P.val);
case token.LPAREN:
// don't print '(' - screws up selection in terminal window
fmt.Printf("%d:%d: LPAREN\n", P.loc.Line, P.loc.Col);
case token.RPAREN:
// don't print ')' - screws up selection in terminal window
fmt.Printf("%d:%d: RPAREN\n", P.loc.Line, P.loc.Col);
default:
fmt.Printf("%d:%d: %s\n", P.loc.Line, P.loc.Col, token.TokenString(P.tok));
}
}
}
func (P *Parser) getComment() *ast.Comment {
defer P.next0();
// for /*-style comments, the comment may end on a different line
endline := P.loc.Line;
if P.val[1] == '*' {
for i, b := range P.val {
if b == '\n' {
endline++;
}
}
}
return &ast.Comment{P.loc, endline, P.val};
}
func (P *Parser) getCommentGroup() ast.CommentGroup {
list := vector.New(0);
// group adjacent comments
// (an empty line terminates a group)
endline := P.loc.Line;
for P.tok == token.COMMENT && endline+1 >= P.loc.Line {
c := P.getComment();
list.Push(c);
endline = c.EndLine;
}
// convert list
group := make(ast.CommentGroup, list.Len());
for i := 0; i < list.Len(); i++ {
group[i] = list.At(i).(*ast.Comment);
}
return group;
}
func (P *Parser) getLastComment() ast.CommentGroup {
c := P.last_comment;
if c != nil && c[len(c) - 1].EndLine + 1 < P.loc.Line {
// empty line between last comment and current token,
// at least one line of space between last comment
// and current token; ignore this comment
return nil;
}
return c;
}
func (P *Parser) next() {
P.next0();
P.last_comment = nil;
for P.tok == token.COMMENT {
P.last_comment = P.getCommentGroup();
P.comments.Push(P.last_comment);
}
}
func (P *Parser) Init(scanner *scanner.Scanner, err scanner.ErrorHandler, trace bool) {
P.scanner = scanner;
P.err = err;
P.trace = trace;
P.comments.Init(0);
P.next();
}
func (P *Parser) error(loc scanner.Location, msg string) {
P.err.Error(loc, msg);
}
func (P *Parser) expect(tok int) {
if P.tok != tok {
msg := "expected '" + token.TokenString(tok) + "', found '" + token.TokenString(P.tok) + "'";
if token.IsLiteral(P.tok) {
msg += " " + string(P.val);
}
P.error(P.loc, msg);
}
P.next(); // make progress in any case
}
// ----------------------------------------------------------------------------
// Common productions
func (P *Parser) tryType() ast.Expr;
func (P *Parser) parseExpression(prec int) ast.Expr;
func (P *Parser) parseStatement() ast.Stat;
func (P *Parser) parseDeclaration() ast.Decl;
func (P *Parser) parseIdent() *ast.Ident {
if P.trace {
defer un(trace(P, "Ident"));
}
2009-01-20 15:40:40 -07:00
if P.tok == token.IDENT {
x := &ast.Ident{P.loc, string(P.val)};
P.next();
return x;
}
P.expect(token.IDENT); // use expect() error handling
return &ast.Ident{P.loc, ""};
}
func (P *Parser) parseIdentList(x ast.Expr) []*ast.Ident {
if P.trace {
defer un(trace(P, "IdentList"));
}
list := vector.New(0);
if x == nil {
x = P.parseIdent();
}
list.Push(x);
for P.tok == token.COMMA {
P.next();
list.Push(P.parseIdent());
}
// convert vector
idents := make([]*ast.Ident, list.Len());
for i := 0; i < list.Len(); i++ {
idents[i] = list.At(i).(*ast.Ident);
}
return idents;
}
func (P *Parser) parseExpressionList() []ast.Expr {
if P.trace {
defer un(trace(P, "ExpressionList"));
}
list := vector.New(0);
list.Push(P.parseExpression(1)); // TODO should use a const instead of 1
for P.tok == token.COMMA {
P.next();
list.Push(P.parseExpression(1)); // TODO should use a const instead of 1
}
// convert vector
exprs := make([]ast.Expr, list.Len());
for i := 0; i < list.Len(); i++ {
exprs[i] = list.At(i).(ast.Expr);
}
return exprs;
}
// ----------------------------------------------------------------------------
// Types
func (P *Parser) parseType() ast.Expr {
if P.trace {
defer un(trace(P, "Type"));
}
t := P.tryType();
if t == nil {
P.error(P.loc, "type expected");
t = &ast.BadExpr{P.loc};
}
return t;
}
func (P *Parser) parseVarType() ast.Expr {
if P.trace {
defer un(trace(P, "VarType"));
}
return P.parseType();
}
func (P *Parser) parseQualifiedIdent() ast.Expr {
if P.trace {
defer un(trace(P, "QualifiedIdent"));
}
var x ast.Expr = P.parseIdent();
for P.tok == token.PERIOD {
loc := P.loc;
P.next();
y := P.parseIdent();
x = &ast.Selector{loc, x, y};
}
return x;
}
func (P *Parser) parseTypeName() ast.Expr {
if P.trace {
defer un(trace(P, "TypeName"));
}
return P.parseQualifiedIdent();
}
func (P *Parser) parseArrayType() *ast.ArrayType {
if P.trace {
defer un(trace(P, "ArrayType"));
}
loc := P.loc;
P.expect(token.LBRACK);
var len ast.Expr;
if P.tok == token.ELLIPSIS {
len = &ast.Ellipsis{P.loc};
P.next();
} else if P.tok != token.RBRACK {
len = P.parseExpression(1);
}
P.expect(token.RBRACK);
elt := P.parseType();
return &ast.ArrayType{loc, len, elt};
}
func (P *Parser) parseChannelType() *ast.ChannelType {
if P.trace {
defer un(trace(P, "ChannelType"));
}
loc := P.loc;
mode := ast.FULL;
if P.tok == token.CHAN {
P.next();
if P.tok == token.ARROW {
P.next();
mode = ast.SEND;
}
} else {
P.expect(token.ARROW);
P.expect(token.CHAN);
mode = ast.RECV;
}
val := P.parseVarType();
return &ast.ChannelType{loc, mode, val};
}
func (P *Parser) tryParameterType() ast.Expr {
if P.tok == token.ELLIPSIS {
loc := P.loc;
P.next();
return &ast.Ellipsis{loc};
}
return P.tryType();
}
func (P *Parser) parseParameterType() ast.Expr {
typ := P.tryParameterType();
if typ == nil {
P.error(P.loc, "type expected");
typ = &ast.BadExpr{P.loc};
}
return typ;
}
func (P *Parser) parseParameterDecl(ellipsis_ok bool) (*vector.Vector, ast.Expr) {
if P.trace {
defer un(trace(P, "ParameterDecl"));
}
// a list of identifiers looks like a list of type names
list := vector.New(0);
for {
// TODO do not allow ()'s here
list.Push(P.parseParameterType());
if P.tok == token.COMMA {
P.next();
} else {
break;
}
}
// if we had a list of identifiers, it must be followed by a type
typ := P.tryParameterType();
return list, typ;
}
func (P *Parser) parseParameterList(ellipsis_ok bool) []*ast.Field {
if P.trace {
defer un(trace(P, "ParameterList"));
}
list, typ := P.parseParameterDecl(false);
if typ != nil {
// IdentifierList Type
// convert list of identifiers into []*Ident
idents := make([]*ast.Ident, list.Len());
for i := 0; i < list.Len(); i++ {
idents[i] = list.At(i).(*ast.Ident);
}
list.Init(0);
list.Push(&ast.Field{idents, typ, nil, nil});
for P.tok == token.COMMA {
P.next();
idents := P.parseIdentList(nil);
typ := P.parseParameterType();
list.Push(&ast.Field{idents, typ, nil, nil});
}
} else {
// Type { "," Type }
// convert list of types into list of *Param
for i := 0; i < list.Len(); i++ {
list.Set(i, &ast.Field{nil, list.At(i).(ast.Expr), nil, nil});
}
}
// convert list
params := make([]*ast.Field, list.Len());
for i := 0; i < list.Len(); i++ {
params[i] = list.At(i).(*ast.Field);
}
return params;
}
// TODO make sure Go spec is updated
func (P *Parser) parseParameters(ellipsis_ok bool) []*ast.Field {
if P.trace {
defer un(trace(P, "Parameters"));
}
var params []*ast.Field;
P.expect(token.LPAREN);
if P.tok != token.RPAREN {
params = P.parseParameterList(ellipsis_ok);
}
P.expect(token.RPAREN);
return params;
}
func (P *Parser) parseResult() []*ast.Field {
if P.trace {
defer un(trace(P, "Result"));
}
var result []*ast.Field;
if P.tok == token.LPAREN {
result = P.parseParameters(false);
} else if P.tok != token.FUNC {
typ := P.tryType();
if typ != nil {
result = make([]*ast.Field, 1);
result[0] = &ast.Field{nil, typ, nil, nil};
}
}
return result;
}
// Function types
//
// (params)
// (params) type
// (params) (results)
func (P *Parser) parseSignature() *ast.Signature {
if P.trace {
defer un(trace(P, "Signature"));
}
params := P.parseParameters(true); // TODO find better solution
//t.End = P.loc;
result := P.parseResult();
return &ast.Signature{params, result};
}
func (P *Parser) parseFunctionType() *ast.FunctionType {
if P.trace {
defer un(trace(P, "FunctionType"));
}
loc := P.loc;
P.expect(token.FUNC);
sig := P.parseSignature();
return &ast.FunctionType{loc, sig};
}
func (P *Parser) parseMethodSpec() *ast.Field {
if P.trace {
defer un(trace(P, "MethodSpec"));
}
var idents []*ast.Ident;
var typ ast.Expr;
x := P.parseQualifiedIdent();
if tmp, is_ident := x.(*ast.Ident); is_ident && (P.tok == token.COMMA || P.tok == token.LPAREN) {
// method(s)
idents = P.parseIdentList(x);
typ = &ast.FunctionType{noloc, P.parseSignature()};
} else {
// embedded interface
typ = x;
}
return &ast.Field{idents, typ, nil, nil};
}
func (P *Parser) parseInterfaceType() *ast.InterfaceType {
if P.trace {
defer un(trace(P, "InterfaceType"));
}
loc := P.loc;
var end scanner.Location;
var methods []*ast.Field;
P.expect(token.INTERFACE);
if P.tok == token.LBRACE {
P.next();
list := vector.New(0);
for P.tok == token.IDENT {
list.Push(P.parseMethodSpec());
if P.tok != token.RBRACE {
P.expect(token.SEMICOLON);
}
}
end = P.loc;
P.expect(token.RBRACE);
P.opt_semi = true;
// convert vector
methods = make([]*ast.Field, list.Len());
for i := list.Len() - 1; i >= 0; i-- {
methods[i] = list.At(i).(*ast.Field);
}
}
return &ast.InterfaceType{loc, methods, end};
}
func (P *Parser) parseMapType() *ast.MapType {
if P.trace {
defer un(trace(P, "MapType"));
}
loc := P.loc;
P.expect(token.MAP);
P.expect(token.LBRACK);
key := P.parseVarType();
P.expect(token.RBRACK);
val := P.parseVarType();
return &ast.MapType{loc, key, val};
}
func (P *Parser) parseStringLit() ast.Expr
func (P *Parser) parseFieldDecl() *ast.Field {
if P.trace {
defer un(trace(P, "FieldDecl"));
}
comment := P.getLastComment();
// a list of identifiers looks like a list of type names
list := vector.New(0);
for {
// TODO do not allow ()'s here
list.Push(P.parseType());
if P.tok == token.COMMA {
P.next();
} else {
break;
}
}
// if we had a list of identifiers, it must be followed by a type
typ := P.tryType();
// optional tag
var tag ast.Expr;
if P.tok == token.STRING {
tag = P.parseStringLit();
}
// analyze case
var idents []*ast.Ident;
if typ != nil {
// non-empty identifier list followed by a type
idents = make([]*ast.Ident, list.Len());
for i := 0; i < list.Len(); i++ {
if ident, is_ident := list.At(i).(*ast.Ident); is_ident {
idents[i] = ident;
} else {
P.error(list.At(i).(ast.Expr).Loc(), "identifier expected");
}
}
} else {
// anonymous field
if list.Len() == 1 {
// TODO should do more checks here
typ = list.At(0).(ast.Expr);
} else {
P.error(P.loc, "anonymous field expected");
}
}
return &ast.Field{idents, typ, tag, comment};
}
func (P *Parser) parseStructType() ast.Expr {
if P.trace {
defer un(trace(P, "StructType"));
}
loc := P.loc;
var end scanner.Location;
var fields []*ast.Field;
P.expect(token.STRUCT);
if P.tok == token.LBRACE {
P.next();
list := vector.New(0);
for P.tok != token.RBRACE && P.tok != token.EOF {
list.Push(P.parseFieldDecl());
if P.tok == token.SEMICOLON {
P.next();
} else {
break;
}
}
if P.tok == token.SEMICOLON {
P.next();
}
end = P.loc;
P.expect(token.RBRACE);
P.opt_semi = true;
// convert vector
fields = make([]*ast.Field, list.Len());
for i := list.Len() - 1; i >= 0; i-- {
fields[i] = list.At(i).(*ast.Field);
}
}
return &ast.StructType{loc, fields, end};
}
func (P *Parser) parsePointerType() ast.Expr {
if P.trace {
defer un(trace(P, "PointerType"));
}
loc := P.loc;
P.expect(token.MUL);
base := P.parseType();
return &ast.PointerType{loc, base};
}
func (P *Parser) tryType() ast.Expr {
if P.trace {
defer un(trace(P, "Type (try)"));
}
switch P.tok {
case token.IDENT: return P.parseTypeName();
case token.LBRACK: return P.parseArrayType();
case token.CHAN, token.ARROW: return P.parseChannelType();
case token.INTERFACE: return P.parseInterfaceType();
case token.FUNC: return P.parseFunctionType();
case token.MAP: return P.parseMapType();
case token.STRUCT: return P.parseStructType();
case token.MUL: return P.parsePointerType();
case token.LPAREN:
loc := P.loc;
P.next();
t := P.parseType();
P.expect(token.RPAREN);
return &ast.Group{loc, t};
}
// no type found
return nil;
}
// ----------------------------------------------------------------------------
// Blocks
func (P *Parser) parseStatementList(list *vector.Vector) {
if P.trace {
defer un(trace(P, "StatementList"));
}
expect_semi := false;
for P.tok != token.CASE && P.tok != token.DEFAULT && P.tok != token.RBRACE && P.tok != token.EOF {
if expect_semi {
P.expect(token.SEMICOLON);
expect_semi = false;
}
list.Push(P.parseStatement());
if P.tok == token.SEMICOLON {
P.next();
} else if P.opt_semi {
P.opt_semi = false; // "consume" optional semicolon
} else {
expect_semi = true;
}
}
}
func (P *Parser) parseBlock(tok int) *ast.Block {
if P.trace {
defer un(trace(P, "Block"));
}
b := ast.NewBlock(P.loc, tok);
P.expect(tok);
P.parseStatementList(b.List);
if tok == token.LBRACE {
b.End = P.loc;
P.expect(token.RBRACE);
P.opt_semi = true;
}
return b;
}
// ----------------------------------------------------------------------------
// Expressions
func (P *Parser) parseFunctionLit() ast.Expr {
if P.trace {
defer un(trace(P, "FunctionLit"));
}
loc := P.loc;
P.expect(token.FUNC);
typ := P.parseSignature();
P.expr_lev++;
body := P.parseBlock(token.LBRACE);
P.expr_lev--;
return &ast.FunctionLit{loc, typ, body};
}
func (P *Parser) parseStringLit() ast.Expr {
if P.trace {
defer un(trace(P, "StringLit"));
}
var x ast.Expr = &ast.BasicLit{P.loc, P.tok, P.val};
P.expect(token.STRING); // always satisfied
for P.tok == token.STRING {
y := &ast.BasicLit{P.loc, P.tok, P.val};
P.next();
x = &ast.ConcatExpr{x, y};
}
return x;
}
func (P *Parser) parseOperand() ast.Expr {
if P.trace {
defer un(trace(P, "Operand"));
}
switch P.tok {
case token.IDENT:
return P.parseIdent();
case token.INT, token.FLOAT, token.CHAR:
x := &ast.BasicLit{P.loc, P.tok, P.val};
P.next();
return x;
case token.STRING:
return P.parseStringLit();
case token.LPAREN:
loc := P.loc;
P.next();
P.expr_lev++;
x := P.parseExpression(1);
P.expr_lev--;
P.expect(token.RPAREN);
return &ast.Group{loc, x};
case token.FUNC:
return P.parseFunctionLit();
default:
t := P.tryType();
if t != nil {
return t;
} else {
P.error(P.loc, "operand expected");
P.next(); // make progress
}
}
return &ast.BadExpr{P.loc};
}
func (P *Parser) parseSelectorOrTypeGuard(x ast.Expr) ast.Expr {
if P.trace {
defer un(trace(P, "SelectorOrTypeGuard"));
}
loc := P.loc;
P.expect(token.PERIOD);
if P.tok == token.IDENT {
x = &ast.Selector{loc, x, P.parseIdent()};
} else {
P.expect(token.LPAREN);
var typ ast.Expr;
if P.tok == token.TYPE {
typ = &ast.TypeType{P.loc};
P.next();
} else {
typ = P.parseType();
}
x = &ast.TypeGuard{loc, x, typ};
P.expect(token.RPAREN);
}
return x;
}
func (P *Parser) parseIndex(x ast.Expr) ast.Expr {
if P.trace {
defer un(trace(P, "IndexOrSlice"));
}
loc := P.loc;
P.expect(token.LBRACK);
P.expr_lev++;
i := P.parseExpression(0);
P.expr_lev--;
P.expect(token.RBRACK);
return &ast.Index{loc, x, i};
}
func (P *Parser) parseBinaryExpr(prec1 int) ast.Expr
func (P *Parser) parseCompositeElements(close int) ast.Expr {
x := P.parseExpression(0);
if P.tok == token.COMMA {
loc := P.loc;
P.next();
// first element determines mode
singles := true;
if t, is_binary := x.(*ast.BinaryExpr); is_binary && t.Tok == token.COLON {
singles = false;
}
var last *ast.BinaryExpr;
for P.tok != close && P.tok != token.EOF {
y := P.parseExpression(0);
if singles {
if t, is_binary := y.(*ast.BinaryExpr); is_binary && t.Tok == token.COLON {
P.error(t.X.Loc(), "single value expected; found pair");
}
} else {
if t, is_binary := y.(*ast.BinaryExpr); !is_binary || t.Tok != token.COLON {
P.error(y.Loc(), "key:value pair expected; found single value");
}
}
if last == nil {
last = &ast.BinaryExpr{loc, token.COMMA, x, y};
x = last;
} else {
last.Y = &ast.BinaryExpr{loc, token.COMMA, last.Y, y};
last = last.Y.(*ast.BinaryExpr);
}
if P.tok == token.COMMA {
loc = P.loc;
P.next();
} else {
break;
}
}
}
return x;
}
func (P *Parser) parseCallOrCompositeLit(f ast.Expr, open, close int) ast.Expr {
if P.trace {
defer un(trace(P, "CallOrCompositeLit"));
}
loc := P.loc;
P.expect(open);
var args ast.Expr;
if P.tok != close {
args = P.parseCompositeElements(close);
}
P.expect(close);
return &ast.Call{loc, open, f, args};
}
func (P *Parser) parsePrimaryExpr() ast.Expr {
if P.trace {
defer un(trace(P, "PrimaryExpr"));
}
x := P.parseOperand();
for {
switch P.tok {
case token.PERIOD: x = P.parseSelectorOrTypeGuard(x);
case token.LBRACK: x = P.parseIndex(x);
// TODO fix once we have decided on literal/conversion syntax
case token.LPAREN: x = P.parseCallOrCompositeLit(x, token.LPAREN, token.RPAREN);
case token.LBRACE:
if P.expr_lev >= 0 {
x = P.parseCallOrCompositeLit(x, token.LBRACE, token.RBRACE);
} else {
return x;
}
default:
return x;
}
}
unreachable();
return nil;
}
func (P *Parser) parseUnaryExpr() ast.Expr {
if P.trace {
defer un(trace(P, "UnaryExpr"));
}
switch P.tok {
case token.ADD, token.SUB, token.MUL, token.NOT, token.XOR, token.ARROW, token.AND:
loc, tok := P.loc, P.tok;
P.next();
y := P.parseUnaryExpr();
return &ast.UnaryExpr{loc, tok, y};
/*
if lit, ok := y.(*ast.TypeLit); ok && tok == token.MUL {
// pointer type
t := ast.NewType(pos, ast.POINTER);
t.Elt = lit.Typ;
return &ast.TypeLit{t};
} else {
return &ast.UnaryExpr{loc, tok, y};
}
*/
}
return P.parsePrimaryExpr();
}
func (P *Parser) parseBinaryExpr(prec1 int) ast.Expr {
if P.trace {
defer un(trace(P, "BinaryExpr"));
}
x := P.parseUnaryExpr();
for prec := token.Precedence(P.tok); prec >= prec1; prec-- {
for token.Precedence(P.tok) == prec {
loc, tok := P.loc, P.tok;
P.next();
y := P.parseBinaryExpr(prec + 1);
x = &ast.BinaryExpr{loc, tok, x, y};
}
}
return x;
}
func (P *Parser) parseExpression(prec int) ast.Expr {
if P.trace {
defer un(trace(P, "Expression"));
}
if prec < 0 {
panic("precedence must be >= 0");
}
return P.parseBinaryExpr(prec);
}
// ----------------------------------------------------------------------------
// Statements
const /* mode */ (
label_ok = 1 << iota;
range_ok;
)
func (P *Parser) parseSimpleStat(mode int) ast.Stat {
if P.trace {
defer un(trace(P, "SimpleStat"));
}
loc := P.loc;
x := P.parseExpressionList();
switch P.tok {
case token.COLON:
// labeled statement
loc := P.loc;
P.expect(token.COLON);
P.opt_semi = true;
if mode & label_ok != 0 && len(x) == 1 {
if label, is_ident := x[0].(*ast.Ident); is_ident {
return &ast.LabeledStat{loc, label, P.parseStatement()};
}
}
P.error(loc, "illegal label declaration");
return nil;
case
token.DEFINE, token.ASSIGN, token.ADD_ASSIGN,
token.SUB_ASSIGN, token.MUL_ASSIGN, token.QUO_ASSIGN,
token.REM_ASSIGN, token.AND_ASSIGN, token.OR_ASSIGN,
token.XOR_ASSIGN, token.SHL_ASSIGN, token.SHR_ASSIGN:
// assignment statement or range clause
loc, tok := P.loc, P.tok;
P.next();
if mode & range_ok != 0 && P.tok == token.RANGE {
// range clause
P.next();
if len(x) != 1 && len(x) != 2 {
P.error(loc, "expected 1 or 2 expressions on lhs of range clause");
}
if tok != token.DEFINE && tok != token.ASSIGN {
P.error(loc, "expected '=' or ':=', found '" + token.TokenString(tok) + "'");
}
y := P.parseExpression(1);
return &ast.RangeClause{loc, tok, x, y};
} else {
// assignment statement
y := P.parseExpressionList();
xl, yl := len(x), len(y);
if xl > 1 && yl > 1 && xl != yl {
P.error(loc, "arity of lhs doesn't match rhs"); // TODO use better loc for error
}
if xl == 1 && yl == 1 {
// common case - use smaller node
return &ast.AssignmentStat{loc, tok, x[0], y[0]};
} else {
// general case
return &ast.TupleAssignStat{loc, tok, x, y};
}
}
default:
if len(x) != 1 {
P.error(loc, "only one expression allowed");
}
if P.tok == token.INC || P.tok == token.DEC {
s := &ast.IncDecStat{P.loc, P.tok, x[0]};
P.next(); // consume "++" or "--"
return s;
}
// TODO change ILLEGAL -> NONE
return &ast.ExpressionStat{loc, token.ILLEGAL, x[0]};
}
unreachable();
return nil;
}
func (P *Parser) parseInvocationStat(keyword int) *ast.ExpressionStat {
if P.trace {
defer un(trace(P, "InvocationStat"));
}
loc := P.loc;
P.expect(keyword);
return &ast.ExpressionStat{loc, keyword, P.parseExpression(1)};
}
func (P *Parser) parseReturnStat() *ast.ReturnStat {
if P.trace {
defer un(trace(P, "ReturnStat"));
}
loc := P.loc;
P.expect(token.RETURN);
var x []ast.Expr;
if P.tok != token.SEMICOLON && P.tok != token.RBRACE {
x = P.parseExpressionList();
}
return &ast.ReturnStat{loc, x};
}
func (P *Parser) parseControlFlowStat(tok int) *ast.ControlFlowStat {
if P.trace {
defer un(trace(P, "ControlFlowStat"));
}
s := &ast.ControlFlowStat{P.loc, tok, nil};
P.expect(tok);
if tok != token.FALLTHROUGH && P.tok == token.IDENT {
s.Label = P.parseIdent();
}
return s;
}
func (P *Parser) parseControlClause(isForStat bool) (init ast.Stat, expr ast.Expr, post ast.Stat) {
if P.trace {
defer un(trace(P, "ControlClause"));
}
if P.tok != token.LBRACE {
prev_lev := P.expr_lev;
P.expr_lev = -1;
if P.tok != token.SEMICOLON {
mode := 0;
if isForStat {
mode = range_ok;
}
init = P.parseSimpleStat(mode);
}
if dummy, is_range := init.(*ast.RangeClause); !is_range {
if P.tok == token.SEMICOLON {
P.next();
if P.tok != token.SEMICOLON && P.tok != token.LBRACE {
expr = P.parseExpression(1);
}
if isForStat {
P.expect(token.SEMICOLON);
if P.tok != token.LBRACE {
post = P.parseSimpleStat(0);
}
}
} else {
if init != nil { // guard in case of errors
if s, is_expr_stat := init.(*ast.ExpressionStat); is_expr_stat {
expr, init = s.Expr, nil;
} else {
P.error(noloc, "illegal control clause");
}
}
}
}
P.expr_lev = prev_lev;
}
return init, expr, post;
}
func (P *Parser) parseIfStat() *ast.IfStat {
if P.trace {
defer un(trace(P, "IfStat"));
}
loc := P.loc;
P.expect(token.IF);
init, cond, dummy := P.parseControlClause(false);
body := P.parseBlock(token.LBRACE);
var else_ ast.Stat;
if P.tok == token.ELSE {
P.next();
else_ = P.parseStatement();
}
return &ast.IfStat{loc, init, cond, body, else_};
}
func (P *Parser) parseForStat() *ast.ForStat {
if P.trace {
defer un(trace(P, "ForStat"));
}
loc := P.loc;
P.expect(token.FOR);
init, cond, post := P.parseControlClause(true);
body := P.parseBlock(token.LBRACE);
return &ast.ForStat{loc, init, cond, post, body};
}
func (P *Parser) asIdent(x ast.Expr) *ast.Ident {
if name, ok := x.(*ast.Ident); ok {
return name;
}
P.error(x.Loc(), "identifier expected");
return &ast.Ident{noloc, "BAD"};
}
func (P *Parser) isTypeSwitch(init ast.Stat) (lhs *ast.Ident, rhs ast.Expr) {
if assign, ok := init.(*ast.AssignmentStat); ok {
if guard, ok := assign.Rhs.(*ast.TypeGuard); ok {
if tmp, ok := guard.Typ.(*ast.TypeType); ok {
// we appear to have a type switch
// TODO various error checks
return P.asIdent(assign.Lhs), guard.X;
}
}
}
return nil, nil;
}
func (P *Parser) parseCaseClause() *ast.CaseClause {
if P.trace {
defer un(trace(P, "CaseClause"));
}
// SwitchCase
loc := P.loc;
var x []ast.Expr;
if P.tok == token.CASE {
P.next();
x = P.parseExpressionList();
} else {
P.expect(token.DEFAULT);
}
return &ast.CaseClause{loc, x, P.parseBlock(token.COLON)};
}
func (P *Parser) parseSwitchStat() *ast.SwitchStat {
if P.trace {
defer un(trace(P, "SwitchStat"));
}
loc := P.loc;
P.expect(token.SWITCH);
init, tag, post := P.parseControlClause(false);
body := ast.NewBlock(P.loc, token.LBRACE);
P.expect(token.LBRACE);
for P.tok != token.RBRACE && P.tok != token.EOF {
body.List.Push(P.parseCaseClause());
}
body.End = P.loc;
P.expect(token.RBRACE);
P.opt_semi = true;
if lhs, rhs := P.isTypeSwitch(init); lhs != nil {
if tag != nil {
P.error(loc, "illegal type switch clause");
}
// TODO fix location
init = &ast.TypeSwitchClause{loc, lhs, rhs};
}
return &ast.SwitchStat{loc, init, tag, body};
}
func (P *Parser) parseCommClause() *ast.CommClause {
if P.trace {
defer un(trace(P, "CommClause"));
}
// CommCase
loc := P.loc;
var tok int;
var lhs, rhs ast.Expr;
if P.tok == token.CASE {
P.next();
if P.tok == token.ARROW {
// RecvExpr without assignment
rhs = P.parseExpression(1);
} else {
// SendExpr or RecvExpr
rhs = P.parseExpression(1);
if P.tok == token.ASSIGN || P.tok == token.DEFINE {
// RecvExpr with assignment
tok = P.tok;
P.next();
lhs = rhs;
if P.tok == token.ARROW {
rhs = P.parseExpression(1);
} else {
P.expect(token.ARROW); // use expect() error handling
}
}
// else SendExpr
}
} else {
P.expect(token.DEFAULT);
}
return &ast.CommClause{loc, tok, lhs, rhs, P.parseBlock(token.COLON)};
}
func (P *Parser) parseSelectStat() *ast.SelectStat {
if P.trace {
defer un(trace(P, "SelectStat"));
}
loc := P.loc;
P.expect(token.SELECT);
body := ast.NewBlock(P.loc, token.LBRACE);
P.expect(token.LBRACE);
for P.tok != token.RBRACE && P.tok != token.EOF {
body.List.Push(P.parseCommClause());
}
body.End = P.loc;
P.expect(token.RBRACE);
P.opt_semi = true;
return &ast.SelectStat{loc, body};
}
func (P *Parser) parseStatement() ast.Stat {
if P.trace {
defer un(trace(P, "Statement"));
}
switch P.tok {
case token.CONST, token.TYPE, token.VAR:
return &ast.DeclarationStat{P.parseDeclaration()};
case
// tokens that may start a top-level expression
token.IDENT, token.INT, token.FLOAT, token.CHAR, token.STRING, token.FUNC, token.LPAREN, // operand
token.LBRACK, token.STRUCT, // composite type
token.MUL, token.AND, token.ARROW: // unary operators
return P.parseSimpleStat(label_ok);
case token.GO, token.DEFER:
return P.parseInvocationStat(P.tok);
case token.RETURN:
return P.parseReturnStat();
case token.BREAK, token.CONTINUE, token.GOTO, token.FALLTHROUGH:
return P.parseControlFlowStat(P.tok);
case token.LBRACE:
return &ast.CompositeStat{P.parseBlock(token.LBRACE)};
case token.IF:
return P.parseIfStat();
case token.FOR:
return P.parseForStat();
case token.SWITCH:
return P.parseSwitchStat();
case token.SELECT:
return P.parseSelectStat();
case token.SEMICOLON, token.RBRACE:
// don't consume the ";", it is the separator following the empty statement
return &ast.EmptyStat{P.loc};
}
// no statement found
P.error(P.loc, "statement expected");
return &ast.BadStat{P.loc};
}
// ----------------------------------------------------------------------------
// Declarations
func (P *Parser) parseImportSpec(loc scanner.Location) *ast.ImportDecl {
if P.trace {
defer un(trace(P, "ImportSpec"));
}
var ident *ast.Ident;
if P.tok == token.PERIOD {
P.error(P.loc, `"import ." not yet handled properly`);
P.next();
} else if P.tok == token.IDENT {
ident = P.parseIdent();
}
var path ast.Expr;
if P.tok == token.STRING {
path = P.parseStringLit();
} else {
P.expect(token.STRING); // use expect() error handling
}
return &ast.ImportDecl{loc, ident, path};
}
func (P *Parser) parseConstSpec(loc scanner.Location, comment ast.CommentGroup) *ast.ConstDecl {
if P.trace {
defer un(trace(P, "ConstSpec"));
}
names := P.parseIdentList(nil);
typ := P.tryType();
var values []ast.Expr;
if typ != nil || P.tok == token.ASSIGN {
P.expect(token.ASSIGN);
values = P.parseExpressionList();
}
return &ast.ConstDecl{loc, names, typ, values, comment};
}
func (P *Parser) parseTypeSpec(loc scanner.Location, comment ast.CommentGroup) *ast.TypeDecl {
if P.trace {
defer un(trace(P, "TypeSpec"));
}
ident := P.parseIdent();
typ := P.parseType();
return &ast.TypeDecl{loc, ident, typ, comment};
}
func (P *Parser) parseVarSpec(loc scanner.Location, comment ast.CommentGroup) *ast.VarDecl {
if P.trace {
defer un(trace(P, "VarSpec"));
}
names := P.parseIdentList(nil);
typ := P.tryType();
var values []ast.Expr;
if typ == nil || P.tok == token.ASSIGN {
P.expect(token.ASSIGN);
values = P.parseExpressionList();
}
return &ast.VarDecl{loc, names, typ, values, comment};
}
func (P *Parser) parseSpec(loc scanner.Location, comment ast.CommentGroup, keyword int) ast.Decl {
switch keyword {
case token.IMPORT: return P.parseImportSpec(loc);
case token.CONST: return P.parseConstSpec(loc, comment);
case token.TYPE: return P.parseTypeSpec(loc, comment);
case token.VAR: return P.parseVarSpec(loc, comment);
}
unreachable();
return nil;
}
func (P *Parser) parseDecl(keyword int) ast.Decl {
if P.trace {
defer un(trace(P, "Decl"));
}
comment := P.getLastComment();
loc := P.loc;
P.expect(keyword);
if P.tok == token.LPAREN {
P.next();
list := vector.New(0);
for P.tok != token.RPAREN && P.tok != token.EOF {
list.Push(P.parseSpec(noloc, nil, keyword));
if P.tok == token.SEMICOLON {
P.next();
} else {
break;
}
}
end := P.loc;
P.expect(token.RPAREN);
P.opt_semi = true;
// convert vector
decls := make([]ast.Decl, list.Len());
for i := 0; i < list.Len(); i++ {
decls[i] = list.At(i).(ast.Decl);
}
return &ast.DeclList{loc, keyword, decls, end};
}
return P.parseSpec(loc, comment, keyword);
}
// Function and method declarations
//
// func ident (params)
// func ident (params) type
// func ident (params) (results)
// func (recv) ident (params)
// func (recv) ident (params) type
// func (recv) ident (params) (results)
func (P *Parser) parseFunctionDecl() *ast.FuncDecl {
if P.trace {
defer un(trace(P, "FunctionDecl"));
}
comment := P.getLastComment();
loc := P.loc;
P.expect(token.FUNC);
var recv *ast.Field;
if P.tok == token.LPAREN {
loc := P.loc;
tmp := P.parseParameters(true);
if len(tmp) == 1 {
recv = tmp[0];
} else {
P.error(loc, "must have exactly one receiver");
}
}
ident := P.parseIdent();
sig := P.parseSignature();
var body *ast.Block;
if P.tok == token.LBRACE {
body = P.parseBlock(token.LBRACE);
}
return &ast.FuncDecl{loc, recv, ident, sig, body, comment};
}
func (P *Parser) parseDeclaration() ast.Decl {
if P.trace {
defer un(trace(P, "Declaration"));
}
switch P.tok {
case token.CONST, token.TYPE, token.VAR:
return P.parseDecl(P.tok);
case token.FUNC:
return P.parseFunctionDecl();
}
loc := P.loc;
P.error(loc, "declaration expected");
P.next(); // make progress
return &ast.BadDecl{loc};
}
// ----------------------------------------------------------------------------
// Program
func (P *Parser) getComments() []ast.CommentGroup {
// convert comments vector
list := make([]ast.CommentGroup, P.comments.Len());
for i := 0; i < P.comments.Len(); i++ {
list[i] = P.comments.At(i).(ast.CommentGroup);
}
return list;
}
// The Parse function is parametrized with one of the following
// constants. They control how much of the source text is parsed.
//
const (
ParseEntirePackage = iota;
ParseImportDeclsOnly;
ParsePackageClauseOnly;
)
// Parse parses the source...
//
// foo bar
//
func (P *Parser) Parse(mode int) *ast.Program {
if P.trace {
defer un(trace(P, "Program"));
}
// package clause
comment := P.getLastComment();
loc := P.loc;
P.expect(token.PACKAGE);
name := P.parseIdent();
if P.tok == token.SEMICOLON {
// common error
P.error(P.loc, "extra semicolon");
P.next();
}
var decls []ast.Decl;
if mode <= ParseImportDeclsOnly {
// import decls
list := vector.New(0);
for P.tok == token.IMPORT {
list.Push(P.parseDecl(token.IMPORT));
if P.tok == token.SEMICOLON {
P.next();
}
}
if mode <= ParseEntirePackage {
// rest of package body
for P.tok != token.EOF {
list.Push(P.parseDeclaration());
if P.tok == token.SEMICOLON {
P.next();
}
}
}
// convert list
decls = make([]ast.Decl, list.Len());
for i := 0; i < list.Len(); i++ {
decls[i] = list.At(i).(ast.Decl);
}
}
return &ast.Program{loc, name, decls, comment, P.getComments()};
}