// 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. The parser is invoked by calling // Parse. // package parser import ( "bytes"; "container/vector"; "fmt"; "go/ast"; "go/scanner"; "go/token"; "io"; "os"; "strings"; ) // A parser error is represented by an Error node. The position Pos, if // valid, points to the beginning of the offending token, and the error // condition is described by Msg. // type Error struct { Pos token.Position; Msg string; } func (e *Error) String() string { pos := ""; if e.Pos.IsValid() { pos = fmt.Sprintf("%d:%d: ", e.Pos.Line, e.Pos.Column); } return pos + e.Msg; } // Parser errors are returned as an ErrorList. type ErrorList []*Error // ErrorList implements the SortInterface. func (p ErrorList) Len() int { return len(p); } func (p ErrorList) Swap(i, j int) { p[i], p[j] = p[j], p[i]; } func (p ErrorList) Less(i, j int) bool { return p[i].Pos.Offset < p[j].Pos.Offset; } func (p ErrorList) String() string { switch len(p) { case 0: return "unspecified error"; case 1: return p[0].String(); } return fmt.Sprintf("%s (and %d more errors)", p[0].String(), len(p) - 1); } type interval struct { beg, end int; } // The parser structure holds the parser's internal state. type parser struct { errors vector.Vector; scanner scanner.Scanner; // Tracing/debugging mode uint; // parsing mode trace bool; // == (mode & Trace != 0) indent uint; // indentation used for tracing output // Comments // (comment indices and intervals index the comments vector) comments vector.Vector; // list of collected, unassociated comments lastComment int; // index of last trailing comment lastDoc interval; // last interval of consequtive free-standing comments // Next token pos token.Position; // token position tok token.Token; // one token look-ahead lit []byte; // token literal // Non-syntactic parser control optSemi bool; // true if semicolon separator is optional in statement list exprLev int; // < 0: in control clause, >= 0: in expression }; // noPos is used when there is no corresponding source position for a token var noPos token.Position; // ---------------------------------------------------------------------------- // Parsing support func (p *parser) printTrace(a ...) { const dots = ". . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . " ". . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . "; const n = uint(len(dots)); fmt.Printf("%5d:%3d: ", p.pos.Line, p.pos.Column); i := 2*p.indent; for ; i > n; i -= n { fmt.Print(dots); } fmt.Print(dots[0 : i]); fmt.Println(a); } func trace(p *parser, msg string) *parser { p.printTrace(msg, "("); p.indent++; return p; } func un/*trace*/(p *parser) { p.indent--; p.printTrace(")"); } // Advance to the next token. func (p *parser) next0() { // Because of one-token look-ahead, print the previous token // when tracing as it provides a more readable output. The // very first token (p.pos.Line == 0) is not initialized (it // is token.ILLEGAL), so don't print it . if p.trace && p.pos.Line > 0 { s := p.tok.String(); switch { case p.tok.IsLiteral(): p.printTrace(s, string(p.lit)); case p.tok.IsOperator(), p.tok.IsKeyword(): p.printTrace("\"" + s + "\""); default: p.printTrace(s); } } p.pos, p.tok, p.lit = p.scanner.Scan(); p.optSemi = false; } // Consume a comment, add it to the parser's comment list, // and return the line on which the comment ends. // func (p *parser) consumeComment() int { // For /*-style comments, the comment may end on a different line. // Scan the comment for '\n' chars and adjust the end line accordingly. // (Note that the position of the next token may be even further down // as there may be more whitespace lines after the comment.) endline := p.pos.Line; if p.lit[1] == '*' { for _, b := range p.lit { if b == '\n' { endline++; } } } p.comments.Push(&ast.Comment{p.pos, p.lit, endline}); p.next0(); return endline; } // Consume a group of adjacent comments and return the interval of // indices into the parser's comment list. An empty line or non-comment // token terminates a comment group. // func (p *parser) consumeCommentGroup() interval { beg := p.comments.Len(); endline := p.pos.Line; for p.tok == token.COMMENT && endline+1 >= p.pos.Line { endline = p.consumeComment(); } end := p.comments.Len(); return interval{beg, end}; } var empty interval; // Advance to the next non-comment token. func (p *parser) next() { p.lastComment = -1; p.lastDoc = empty; line := p.pos.Line; p.next0(); if p.tok == token.COMMENT { // the first comment may be a trailing comment if p.pos.Line == line { // comment is on same line as previous token; // it is not considered part of a free-standing comment group p.consumeComment(); if p.pos.Line != line { // the next token is on a different line, thus // the last comment is a trailing comment p.lastComment = p.comments.Len() - 1; } } // consume any successor comments group := empty; for p.tok == token.COMMENT { group = p.consumeCommentGroup(); } if group.end > 0 && p.comments.At(group.end - 1).(*ast.Comment).EndLine + 1 == p.pos.Line { // there is a comment group and the next token is following on the // line immediately after the group, thus the group may be used as // documentation p.lastDoc = group; } } } // Get current trailing comment, if any. func (p *parser) getComment() *ast.Comment { i := p.lastComment; if i < 0 { // no last comment return nil; } // get comment and remove it from the general list c := p.comments.At(i).(*ast.Comment); p.comments.Set(i, nil); // clear entry p.lastComment = -1; return c; } // Get current documentation comment group, if any. func (p *parser) getDoc() ast.Comments { doc := p.lastDoc; n := doc.end - doc.beg; if n == 0 { // no last comment group return nil; } // get comment group and remove if from the general list c := make(ast.Comments, n); for i := 0; i < n; i++ { c[i] = p.comments.At(doc.beg + i).(*ast.Comment); p.comments.Set(doc.beg + i, nil); // clear entry } p.lastDoc = empty; return c; } // The parser implements scanner.Error. func (p *parser) Error(pos token.Position, msg string) { // Don't collect errors that are on the same line as the previous error // in the hope to reduce the number of spurious errors due to incorrect // parser synchronization. if p.errors.Len() == 0 || p.errors.Last().(*Error).Pos.Line != pos.Line { p.errors.Push(&Error{pos, msg}); } } func (p *parser) errorExpected(pos token.Position, msg string) { msg = "expected " + msg; if pos.Offset == p.pos.Offset { // the error happened at the current position; // make the error message more specific msg += ", found '" + p.tok.String() + "'"; if p.tok.IsLiteral() { msg += " " + string(p.lit); } } p.Error(pos, msg); } func (p *parser) expect(tok token.Token) token.Position { pos := p.pos; if p.tok != tok { p.errorExpected(pos, "'" + tok.String() + "'"); } p.next(); // make progress in any case return pos; } // ---------------------------------------------------------------------------- // Common productions func (p *parser) tryType() ast.Expr func (p *parser) parseStringList(x *ast.StringLit) []*ast.StringLit func (p *parser) parseExpression() ast.Expr func (p *parser) parseStatement() ast.Stmt func (p *parser) parseDeclaration(getSemi bool) (decl ast.Decl, gotSemi bool) func (p *parser) parseIdent() *ast.Ident { if p.tok == token.IDENT { x := &ast.Ident{p.pos, string(p.lit)}; p.next(); return x; } p.expect(token.IDENT); // use expect() error handling return &ast.Ident{p.pos, ""}; } 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()); for p.tok == token.COMMA { p.next(); list.Push(p.parseExpression()); } // convert list 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")); } typ := p.tryType(); if typ == nil { p.errorExpected(p.pos, "type"); p.next(); // make progress return &ast.BadExpr{p.pos}; } return typ; } func (p *parser) parseQualifiedIdent() ast.Expr { if p.trace { defer un(trace(p, "QualifiedIdent")); } var x ast.Expr = p.parseIdent(); if p.tok == token.PERIOD { // first identifier is a package identifier p.next(); sel := p.parseIdent(); x = &ast.SelectorExpr{x, sel}; } return x; } func (p *parser) parseTypeName() ast.Expr { if p.trace { defer un(trace(p, "TypeName")); } return p.parseQualifiedIdent(); } func (p *parser) parseArrayType(ellipsisOk bool) ast.Expr { if p.trace { defer un(trace(p, "ArrayType")); } lbrack := p.expect(token.LBRACK); var len ast.Expr; if ellipsisOk && p.tok == token.ELLIPSIS { len = &ast.Ellipsis{p.pos}; p.next(); } else if p.tok != token.RBRACK { len = p.parseExpression(); } p.expect(token.RBRACK); elt := p.parseType(); return &ast.ArrayType{lbrack, len, elt}; } func (p *parser) makeIdentList(list *vector.Vector) []*ast.Ident { idents := make([]*ast.Ident, list.Len()); for i := 0; i < list.Len(); i++ { ident, isIdent := list.At(i).(*ast.Ident); if !isIdent { pos := list.At(i).(ast.Expr).Pos(); p.errorExpected(pos, "identifier"); idents[i] = &ast.Ident{pos, ""}; } idents[i] = ident; } return idents; } func (p *parser) parseFieldDecl() *ast.Field { if p.trace { defer un(trace(p, "FieldDecl")); } doc := p.getDoc(); // 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.StringLit; if p.tok == token.STRING { tag = p.parseStringList(nil); } // analyze case var idents []*ast.Ident; if typ != nil { // IdentifierList Type idents = p.makeIdentList(list); } else { // Type (anonymous field) if list.Len() == 1 { // TODO check that this looks like a type typ = list.At(0).(ast.Expr); } else { p.errorExpected(p.pos, "anonymous field"); typ = &ast.BadExpr{p.pos}; } } return &ast.Field{doc, idents, typ, tag, nil}; } func (p *parser) parseStructType() *ast.StructType { if p.trace { defer un(trace(p, "StructType")); } pos := p.expect(token.STRUCT); var lbrace, rbrace token.Position; var fields []*ast.Field; if p.tok == token.LBRACE { lbrace = p.pos; p.next(); list := vector.New(0); for p.tok != token.RBRACE && p.tok != token.EOF { f := p.parseFieldDecl(); list.Push(f); if p.tok == token.SEMICOLON { p.next(); f.Comment = p.getComment(); } else { f.Comment = p.getComment(); break; } } rbrace = p.expect(token.RBRACE); p.optSemi = 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{pos, lbrace, fields, rbrace}; } func (p *parser) parsePointerType() *ast.StarExpr { if p.trace { defer un(trace(p, "PointerType")); } star := p.expect(token.MUL); base := p.parseType(); return &ast.StarExpr{star, base}; } func (p *parser) tryParameterType(ellipsisOk bool) ast.Expr { if ellipsisOk && p.tok == token.ELLIPSIS { pos := p.pos; p.next(); if p.tok != token.RPAREN { // "..." always must be at the very end of a parameter list p.Error(pos, "expected type, found '...'"); } return &ast.Ellipsis{pos}; } return p.tryType(); } func (p *parser) parseParameterType(ellipsisOk bool) ast.Expr { typ := p.tryParameterType(ellipsisOk); if typ == nil { p.errorExpected(p.pos, "type"); p.next(); // make progress typ = &ast.BadExpr{p.pos}; } return typ; } func (p *parser) parseParameterDecl(ellipsisOk 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(ellipsisOk)); 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(ellipsisOk); return list, typ; } func (p *parser) parseParameterList(ellipsisOk bool) []*ast.Field { if p.trace { defer un(trace(p, "ParameterList")); } list, typ := p.parseParameterDecl(ellipsisOk); if typ != nil { // IdentifierList Type idents := p.makeIdentList(list); list.Init(0); list.Push(&ast.Field{nil, idents, typ, nil, nil}); for p.tok == token.COMMA { p.next(); idents := p.parseIdentList(nil); typ := p.parseParameterType(ellipsisOk); list.Push(&ast.Field{nil, idents, typ, nil, nil}); } } else { // Type { "," Type } (anonymous parameters) // convert list of types into list of *Param for i := 0; i < list.Len(); i++ { list.Set(i, &ast.Field{Type: list.At(i).(ast.Expr)}); } } // convert list params := make([]*ast.Field, list.Len()); for i := 0; i < list.Len(); i++ { params[i] = list.At(i).(*ast.Field); } return params; } func (p *parser) parseParameters(ellipsisOk 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(ellipsisOk); } p.expect(token.RPAREN); return params; } func (p *parser) parseResult() []*ast.Field { if p.trace { defer un(trace(p, "Result")); } var results []*ast.Field; if p.tok == token.LPAREN { results = p.parseParameters(false); } else if p.tok != token.FUNC { typ := p.tryType(); if typ != nil { results = make([]*ast.Field, 1); results[0] = &ast.Field{Type: typ}; } } return results; } func (p *parser) parseSignature() (params []*ast.Field, results []*ast.Field) { if p.trace { defer un(trace(p, "Signature")); } params = p.parseParameters(true); results = p.parseResult(); return params, results; } func (p *parser) parseFuncType() *ast.FuncType { if p.trace { defer un(trace(p, "FuncType")); } pos := p.expect(token.FUNC); params, results := p.parseSignature(); return &ast.FuncType{pos, params, results}; } func (p *parser) parseMethodSpec() *ast.Field { if p.trace { defer un(trace(p, "MethodSpec")); } doc := p.getDoc(); var idents []*ast.Ident; var typ ast.Expr; x := p.parseQualifiedIdent(); if tmp, isIdent := x.(*ast.Ident); isIdent && (p.tok == token.COMMA || p.tok == token.LPAREN) { // methods idents = p.parseIdentList(x); params, results := p.parseSignature(); typ = &ast.FuncType{noPos, params, results}; } else { // embedded interface typ = x; } return &ast.Field{doc, idents, typ, nil, nil}; } func (p *parser) parseInterfaceType() *ast.InterfaceType { if p.trace { defer un(trace(p, "InterfaceType")); } pos := p.expect(token.INTERFACE); var lbrace, rbrace token.Position; var methods []*ast.Field; if p.tok == token.LBRACE { lbrace = p.pos; p.next(); list := vector.New(0); for p.tok == token.IDENT { list.Push(p.parseMethodSpec()); if p.tok != token.RBRACE { p.expect(token.SEMICOLON); } } rbrace = p.expect(token.RBRACE); p.optSemi = 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{pos, lbrace, methods, rbrace}; } func (p *parser) parseMapType() *ast.MapType { if p.trace { defer un(trace(p, "MapType")); } pos := p.expect(token.MAP); p.expect(token.LBRACK); key := p.parseType(); p.expect(token.RBRACK); value := p.parseType(); return &ast.MapType{pos, key, value}; } func (p *parser) parseChanType() *ast.ChanType { if p.trace { defer un(trace(p, "ChanType")); } pos := p.pos; dir := ast.SEND | ast.RECV; if p.tok == token.CHAN { p.next(); if p.tok == token.ARROW { p.next(); dir = ast.SEND; } } else { p.expect(token.ARROW); p.expect(token.CHAN); dir = ast.RECV; } value := p.parseType(); return &ast.ChanType{pos, dir, value}; } func (p *parser) tryRawType(ellipsisOk bool) ast.Expr { switch p.tok { case token.IDENT: return p.parseTypeName(); case token.LBRACK: return p.parseArrayType(ellipsisOk); case token.STRUCT: return p.parseStructType(); case token.MUL: return p.parsePointerType(); case token.FUNC: return p.parseFuncType(); case token.INTERFACE: return p.parseInterfaceType(); case token.MAP: return p.parseMapType(); case token.CHAN, token.ARROW: return p.parseChanType(); case token.LPAREN: lparen := p.pos; p.next(); typ := p.parseType(); rparen := p.expect(token.RPAREN); return &ast.ParenExpr{lparen, typ, rparen}; } // no type found return nil; } func (p *parser) tryType() ast.Expr { return p.tryRawType(false); } // ---------------------------------------------------------------------------- // Blocks func makeStmtList(list *vector.Vector) []ast.Stmt { stats := make([]ast.Stmt, list.Len()); for i := 0; i < list.Len(); i++ { stats[i] = list.At(i).(ast.Stmt); } return stats; } func (p *parser) parseStatementList() []ast.Stmt { if p.trace { defer un(trace(p, "StatementList")); } list := vector.New(0); expectSemi := false; for p.tok != token.CASE && p.tok != token.DEFAULT && p.tok != token.RBRACE && p.tok != token.EOF { if expectSemi { p.expect(token.SEMICOLON); expectSemi = false; } list.Push(p.parseStatement()); if p.tok == token.SEMICOLON { p.next(); } else if p.optSemi { p.optSemi = false; // "consume" optional semicolon } else { expectSemi = true; } } return makeStmtList(list); } func (p *parser) parseBlockStmt() *ast.BlockStmt { if p.trace { defer un(trace(p, "BlockStmt")); } lbrace := p.expect(token.LBRACE); list := p.parseStatementList(); rbrace := p.expect(token.RBRACE); p.optSemi = true; return &ast.BlockStmt{lbrace, list, rbrace}; } // ---------------------------------------------------------------------------- // Expressions func (p *parser) parseStringList(x *ast.StringLit) []*ast.StringLit { if p.trace { defer un(trace(p, "StringList")); } list := vector.New(0); if x != nil { list.Push(x); } for p.tok == token.STRING { list.Push(&ast.StringLit{p.pos, p.lit}); p.next(); } // convert list strings := make([]*ast.StringLit, list.Len()); for i := 0; i < list.Len(); i++ { strings[i] = list.At(i).(*ast.StringLit); } return strings; } func (p *parser) parseFuncLit() ast.Expr { if p.trace { defer un(trace(p, "FuncLit")); } typ := p.parseFuncType(); p.exprLev++; body := p.parseBlockStmt(); p.optSemi = false; // function body requires separating ";" p.exprLev--; return &ast.FuncLit{typ, body}; } // parseOperand may return an expression or a raw type (incl. array // types of the form [...]T. Callers must verify the result. // 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: x := &ast.IntLit{p.pos, p.lit}; p.next(); return x; case token.FLOAT: x := &ast.FloatLit{p.pos, p.lit}; p.next(); return x; case token.CHAR: x := &ast.CharLit{p.pos, p.lit}; p.next(); return x; case token.STRING: x := &ast.StringLit{p.pos, p.lit}; p.next(); if p.tok == token.STRING { return &ast.StringList{p.parseStringList(x)}; } return x; case token.LPAREN: lparen := p.pos; p.next(); p.exprLev++; x := p.parseExpression(); p.exprLev--; rparen := p.expect(token.RPAREN); return &ast.ParenExpr{lparen, x, rparen}; case token.FUNC: return p.parseFuncLit(); default: t := p.tryRawType(true); // could be type for composite literal if t != nil { return t; } } p.errorExpected(p.pos, "operand"); p.next(); // make progress return &ast.BadExpr{p.pos}; } func (p *parser) parseSelectorOrTypeAssertion(x ast.Expr) ast.Expr { if p.trace { defer un(trace(p, "SelectorOrTypeAssertion")); } p.expect(token.PERIOD); if p.tok == token.IDENT { // selector sel := p.parseIdent(); return &ast.SelectorExpr{x, sel}; } // type assertion p.expect(token.LPAREN); var typ ast.Expr; if p.tok == token.TYPE { // special case for type switch typ = &ast.Ident{p.pos, "type"}; p.next(); } else { typ = p.parseType(); } p.expect(token.RPAREN); return &ast.TypeAssertExpr{x, typ}; } func (p *parser) parseIndex(x ast.Expr) ast.Expr { if p.trace { defer un(trace(p, "Index")); } p.expect(token.LBRACK); p.exprLev++; begin := p.parseExpression(); var end ast.Expr; if p.tok == token.COLON { p.next(); end = p.parseExpression(); } p.exprLev--; p.expect(token.RBRACK); return &ast.IndexExpr{x, begin, end}; } func (p *parser) parseCallOrConversion(fun ast.Expr) *ast.CallExpr { if p.trace { defer un(trace(p, "CallOrConversion")); } lparen := p.expect(token.LPAREN); var args []ast.Expr; if p.tok != token.RPAREN { args = p.parseExpressionList(); } rparen := p.expect(token.RPAREN); return &ast.CallExpr{fun, lparen, args, rparen}; } func (p *parser) parseElement() ast.Expr { if p.trace { defer un(trace(p, "Element")); } x := p.parseExpression(); if p.tok == token.COLON { colon := p.pos; p.next(); x = &ast.KeyValueExpr{x, colon, p.parseExpression()}; } return x; } func (p *parser) parseElementList() []ast.Expr { if p.trace { defer un(trace(p, "ElementList")); } list := vector.New(0); for p.tok != token.RBRACE && p.tok != token.EOF { list.Push(p.parseElement()); if p.tok == token.COMMA { p.next(); } else { break; } } // convert list elts := make([]ast.Expr, list.Len()); for i := 0; i < list.Len(); i++ { elts[i] = list.At(i).(ast.Expr); } return elts; } func (p *parser) parseCompositeLit(typ ast.Expr) ast.Expr { if p.trace { defer un(trace(p, "CompositeLit")); } lbrace := p.expect(token.LBRACE); var elts []ast.Expr; if p.tok != token.RBRACE { elts = p.parseElementList(); } rbrace := p.expect(token.RBRACE); return &ast.CompositeLit{typ, lbrace, elts, rbrace}; } // TODO Consider different approach to checking syntax after parsing: // Provide a arguments (set of flags) to parsing functions // restricting what they are syupposed to accept depending // on context. // checkExpr checks that x is an expression (and not a type). func (p *parser) checkExpr(x ast.Expr) ast.Expr { // TODO should provide predicate in AST nodes switch t := x.(type) { case *ast.BadExpr: case *ast.Ident: case *ast.IntLit: case *ast.FloatLit: case *ast.CharLit: case *ast.StringLit: case *ast.StringList: case *ast.FuncLit: case *ast.CompositeLit: case *ast.ParenExpr: case *ast.SelectorExpr: case *ast.IndexExpr: case *ast.TypeAssertExpr: case *ast.CallExpr: case *ast.StarExpr: case *ast.UnaryExpr: if t.Op == token.RANGE { // the range operator is only allowed at the top of a for statement p.errorExpected(x.Pos(), "expression"); x = &ast.BadExpr{x.Pos()}; } case *ast.BinaryExpr: default: // all other nodes are not proper expressions p.errorExpected(x.Pos(), "expression"); x = &ast.BadExpr{x.Pos()}; } return x; } // isTypeName returns true iff x is type name. func isTypeName(x ast.Expr) bool { // TODO should provide predicate in AST nodes switch t := x.(type) { case *ast.BadExpr: case *ast.Ident: case *ast.ParenExpr: return isTypeName(t.X); // TODO should (TypeName) be illegal? case *ast.SelectorExpr: return isTypeName(t.X); default: return false; // all other nodes are not type names } return true; } // isCompositeLitType returns true iff x is a legal composite literal type. func isCompositeLitType(x ast.Expr) bool { // TODO should provide predicate in AST nodes switch t := x.(type) { case *ast.BadExpr: case *ast.Ident: case *ast.ParenExpr: return isCompositeLitType(t.X); case *ast.SelectorExpr: return isTypeName(t.X); case *ast.ArrayType: case *ast.StructType: case *ast.MapType: default: return false; // all other nodes are not legal composite literal types } return true; } // checkExprOrType checks that x is an expression or a type // (and not a raw type such as [...]T). // func (p *parser) checkExprOrType(x ast.Expr) ast.Expr { // TODO should provide predicate in AST nodes switch t := x.(type) { case *ast.UnaryExpr: if t.Op == token.RANGE { // the range operator is only allowed at the top of a for statement p.errorExpected(x.Pos(), "expression"); x = &ast.BadExpr{x.Pos()}; } case *ast.ArrayType: if len, isEllipsis := t.Len.(*ast.Ellipsis); isEllipsis { p.Error(len.Pos(), "expected array length, found '...'"); x = &ast.BadExpr{x.Pos()}; } } // all other nodes are expressions or types return x; } func (p *parser) parsePrimaryExpr() ast.Expr { if p.trace { defer un(trace(p, "PrimaryExpr")); } x := p.parseOperand(); L: for { switch p.tok { case token.PERIOD: x = p.parseSelectorOrTypeAssertion(p.checkExpr(x)); case token.LBRACK: x = p.parseIndex(p.checkExpr(x)); case token.LPAREN: x = p.parseCallOrConversion(p.checkExprOrType(x)); case token.LBRACE: if isCompositeLitType(x) && (p.exprLev >= 0 || !isTypeName(x)) { x = p.parseCompositeLit(x); } else { break L; } default: break L; } } return p.checkExprOrType(x); } func (p *parser) parseUnaryExpr() ast.Expr { if p.trace { defer un(trace(p, "UnaryExpr")); } switch p.tok { case token.ADD, token.SUB, token.NOT, token.XOR, token.ARROW, token.AND, token.RANGE: pos, op := p.pos, p.tok; p.next(); x := p.parseUnaryExpr(); return &ast.UnaryExpr{pos, op, p.checkExpr(x)}; case token.MUL: // unary "*" expression or pointer type pos := p.pos; p.next(); x := p.parseUnaryExpr(); return &ast.StarExpr{pos, p.checkExprOrType(x)}; } return p.parsePrimaryExpr(); } func (p *parser) parseBinaryExpr(prec1 int) ast.Expr { if p.trace { defer un(trace(p, "BinaryExpr")); } x := p.parseUnaryExpr(); for prec := p.tok.Precedence(); prec >= prec1; prec-- { for p.tok.Precedence() == prec { pos, op := p.pos, p.tok; p.next(); y := p.parseBinaryExpr(prec + 1); x = &ast.BinaryExpr{p.checkExpr(x), pos, op, p.checkExpr(y)}; } } return x; } func (p *parser) parseExpression() ast.Expr { if p.trace { defer un(trace(p, "Expression")); } return p.parseBinaryExpr(token.LowestPrec + 1); } // ---------------------------------------------------------------------------- // Statements func (p *parser) parseSimpleStmt(labelOk bool) ast.Stmt { if p.trace { defer un(trace(p, "SimpleStmt")); } x := p.parseExpressionList(); switch p.tok { case token.COLON: // labeled statement p.next(); if labelOk && len(x) == 1 { if label, isIdent := x[0].(*ast.Ident); isIdent { return &ast.LabeledStmt{label, p.parseStatement()}; } } p.Error(x[0].Pos(), "illegal label declaration"); return &ast.BadStmt{x[0].Pos()}; 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, token.AND_NOT_ASSIGN: // assignment statement pos, tok := p.pos, p.tok; p.next(); y := p.parseExpressionList(); if len(x) > 1 && len(y) > 1 && len(x) != len(y) { p.Error(x[0].Pos(), "arity of lhs doesn't match rhs"); } return &ast.AssignStmt{x, pos, tok, y}; } if len(x) > 1 { p.Error(x[0].Pos(), "only one expression allowed"); // continue with first expression } if p.tok == token.INC || p.tok == token.DEC { // increment or decrement s := &ast.IncDecStmt{x[0], p.tok}; p.next(); // consume "++" or "--" return s; } // expression return &ast.ExprStmt{x[0]}; } func (p *parser) parseCallExpr() *ast.CallExpr { x := p.parseExpression(); if call, isCall := x.(*ast.CallExpr); isCall { return call; } p.errorExpected(x.Pos(), "function/method call"); return nil; } func (p *parser) parseGoStmt() ast.Stmt { if p.trace { defer un(trace(p, "GoStmt")); } pos := p.expect(token.GO); call := p.parseCallExpr(); if call != nil { return &ast.GoStmt{pos, call}; } return &ast.BadStmt{pos}; } func (p *parser) parseDeferStmt() ast.Stmt { if p.trace { defer un(trace(p, "DeferStmt")); } pos := p.expect(token.DEFER); call := p.parseCallExpr(); if call != nil { return &ast.DeferStmt{pos, call}; } return &ast.BadStmt{pos}; } func (p *parser) parseReturnStmt() *ast.ReturnStmt { if p.trace { defer un(trace(p, "ReturnStmt")); } pos := p.pos; p.expect(token.RETURN); var x []ast.Expr; if p.tok != token.SEMICOLON && p.tok != token.CASE && p.tok != token.DEFAULT && p.tok != token.RBRACE { x = p.parseExpressionList(); } return &ast.ReturnStmt{pos, x}; } func (p *parser) parseBranchStmt(tok token.Token) *ast.BranchStmt { if p.trace { defer un(trace(p, "BranchStmt")); } s := &ast.BranchStmt{p.pos, tok, nil}; p.expect(tok); if tok != token.FALLTHROUGH && p.tok == token.IDENT { s.Label = p.parseIdent(); } return s; } func (p *parser) isExpr(s ast.Stmt) bool { if s == nil { return true; } dummy, isExpr := s.(*ast.ExprStmt); return isExpr; } func (p *parser) makeExpr(s ast.Stmt) ast.Expr { if s == nil { return nil; } if es, isExpr := s.(*ast.ExprStmt); isExpr { return p.checkExpr(es.X); } p.Error(s.Pos(), "expected condition, found simple statement"); return &ast.BadExpr{s.Pos()}; } func (p *parser) parseControlClause(isForStmt bool) (s1, s2, s3 ast.Stmt) { if p.tok != token.LBRACE { prevLev := p.exprLev; p.exprLev = -1; if p.tok != token.SEMICOLON { s1 = p.parseSimpleStmt(false); } if p.tok == token.SEMICOLON { p.next(); if p.tok != token.LBRACE && p.tok != token.SEMICOLON { s2 = p.parseSimpleStmt(false); } if isForStmt { // for statements have a 3rd section p.expect(token.SEMICOLON); if p.tok != token.LBRACE { s3 = p.parseSimpleStmt(false); } } } else { s1, s2 = nil, s1; } p.exprLev = prevLev; } return s1, s2, s3; } func (p *parser) parseIfStmt() *ast.IfStmt { if p.trace { defer un(trace(p, "IfStmt")); } pos := p.expect(token.IF); s1, s2, dummy := p.parseControlClause(false); body := p.parseBlockStmt(); var else_ ast.Stmt; if p.tok == token.ELSE { p.next(); else_ = p.parseStatement(); } return &ast.IfStmt{pos, s1, p.makeExpr(s2), body, else_}; } func (p *parser) parseCaseClause() *ast.CaseClause { if p.trace { defer un(trace(p, "CaseClause")); } // SwitchCase pos := p.pos; var x []ast.Expr; if p.tok == token.CASE { p.next(); x = p.parseExpressionList(); } else { p.expect(token.DEFAULT); } colon := p.expect(token.COLON); body := p.parseStatementList(); return &ast.CaseClause{pos, x, colon, body}; } func (p *parser) parseTypeCaseClause() *ast.TypeCaseClause { if p.trace { defer un(trace(p, "TypeCaseClause")); } // TypeSwitchCase pos := p.pos; var typ ast.Expr; if p.tok == token.CASE { p.next(); typ = p.parseType(); } else { p.expect(token.DEFAULT); } colon := p.expect(token.COLON); body := p.parseStatementList(); return &ast.TypeCaseClause{pos, typ, colon, body}; } func (p *parser) parseSwitchStmt() ast.Stmt { if p.trace { defer un(trace(p, "SwitchStmt")); } pos := p.expect(token.SWITCH); s1, s2, dummy := p.parseControlClause(false); if p.isExpr(s2) { // expression switch lbrace := p.expect(token.LBRACE); cases := vector.New(0); for p.tok == token.CASE || p.tok == token.DEFAULT { cases.Push(p.parseCaseClause()); } rbrace := p.expect(token.RBRACE); p.optSemi = true; body := &ast.BlockStmt{lbrace, makeStmtList(cases), rbrace}; return &ast.SwitchStmt{pos, s1, p.makeExpr(s2), body}; } // type switch // TODO do all the checks! lbrace := p.expect(token.LBRACE); cases := vector.New(0); for p.tok == token.CASE || p.tok == token.DEFAULT { cases.Push(p.parseTypeCaseClause()); } rbrace := p.expect(token.RBRACE); p.optSemi = true; body := &ast.BlockStmt{lbrace, makeStmtList(cases), rbrace}; return &ast.TypeSwitchStmt{pos, s1, s2, body}; } func (p *parser) parseCommClause() *ast.CommClause { if p.trace { defer un(trace(p, "CommClause")); } // CommCase pos := p.pos; var tok token.Token; var lhs, rhs ast.Expr; if p.tok == token.CASE { p.next(); if p.tok == token.ARROW { // RecvExpr without assignment rhs = p.parseExpression(); } else { // SendExpr or RecvExpr rhs = p.parseExpression(); 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(); } else { p.expect(token.ARROW); // use expect() error handling } } // else SendExpr } } else { p.expect(token.DEFAULT); } colon := p.expect(token.COLON); body := p.parseStatementList(); return &ast.CommClause{pos, tok, lhs, rhs, colon, body}; } func (p *parser) parseSelectStmt() *ast.SelectStmt { if p.trace { defer un(trace(p, "SelectStmt")); } pos := p.expect(token.SELECT); lbrace := p.expect(token.LBRACE); cases := vector.New(0); for p.tok == token.CASE || p.tok == token.DEFAULT { cases.Push(p.parseCommClause()); } rbrace := p.expect(token.RBRACE); p.optSemi = true; body := &ast.BlockStmt{lbrace, makeStmtList(cases), rbrace}; return &ast.SelectStmt{pos, body}; } func (p *parser) parseForStmt() ast.Stmt { if p.trace { defer un(trace(p, "ForStmt")); } pos := p.expect(token.FOR); s1, s2, s3 := p.parseControlClause(true); body := p.parseBlockStmt(); if as, isAssign := s2.(*ast.AssignStmt); isAssign { // possibly a for statement with a range clause; check assignment operator if as.Tok != token.ASSIGN && as.Tok != token.DEFINE { p.errorExpected(as.TokPos, "'=' or ':='"); return &ast.BadStmt{pos}; } // check lhs var key, value ast.Expr; switch len(as.Lhs) { case 2: value = as.Lhs[1]; fallthrough; case 1: key = as.Lhs[0]; default: p.errorExpected(as.Lhs[0].Pos(), "1 or 2 expressions"); return &ast.BadStmt{pos}; } // check rhs if len(as.Rhs) != 1 { p.errorExpected(as.Rhs[0].Pos(), "1 expressions"); return &ast.BadStmt{pos}; } if rhs, isUnary := as.Rhs[0].(*ast.UnaryExpr); isUnary && rhs.Op == token.RANGE { // rhs is range expression; check lhs return &ast.RangeStmt{pos, key, value, as.TokPos, as.Tok, rhs.X, body} } else { p.errorExpected(s2.Pos(), "range clause"); return &ast.BadStmt{pos}; } } else { // regular for statement return &ast.ForStmt{pos, s1, p.makeExpr(s2), s3, body}; } panic(); // unreachable return nil; } func (p *parser) parseStatement() ast.Stmt { if p.trace { defer un(trace(p, "Statement")); } switch p.tok { case token.CONST, token.TYPE, token.VAR: decl, _ := p.parseDeclaration(false); // do not consume trailing semicolon return &ast.DeclStmt{decl}; 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.parseSimpleStmt(true); case token.GO: return p.parseGoStmt(); case token.DEFER: return p.parseDeferStmt(); case token.RETURN: return p.parseReturnStmt(); case token.BREAK, token.CONTINUE, token.GOTO, token.FALLTHROUGH: return p.parseBranchStmt(p.tok); case token.LBRACE: return p.parseBlockStmt(); case token.IF: return p.parseIfStmt(); case token.SWITCH: return p.parseSwitchStmt(); case token.SELECT: return p.parseSelectStmt(); case token.FOR: return p.parseForStmt(); case token.SEMICOLON, token.RBRACE: // don't consume the ";", it is the separator following the empty statement return &ast.EmptyStmt{p.pos}; } // no statement found p.errorExpected(p.pos, "statement"); p.next(); // make progress return &ast.BadStmt{p.pos}; } // ---------------------------------------------------------------------------- // Declarations type parseSpecFunction func(p *parser, doc ast.Comments, getSemi bool) (spec ast.Spec, gotSemi bool) // Consume semicolon if there is one and getSemi is set, and get any trailing comment. // Return the comment if any and indicate if a semicolon was consumed. // func (p *parser) parseComment(getSemi bool) (comment *ast.Comment, gotSemi bool) { if getSemi && p.tok == token.SEMICOLON { p.next(); gotSemi = true; } return p.getComment(), gotSemi; } func parseImportSpec(p *parser, doc ast.Comments, getSemi bool) (spec ast.Spec, gotSemi bool) { if p.trace { defer un(trace(p, "ImportSpec")); } var ident *ast.Ident; if p.tok == token.PERIOD { ident = &ast.Ident{p.pos, "."}; p.next(); } else if p.tok == token.IDENT { ident = p.parseIdent(); } var path []*ast.StringLit; if p.tok == token.STRING { path = p.parseStringList(nil); } else { p.expect(token.STRING); // use expect() error handling } comment, gotSemi := p.parseComment(getSemi); return &ast.ImportSpec{doc, ident, path, comment}, gotSemi; } func parseConstSpec(p *parser, doc ast.Comments, getSemi bool) (spec ast.Spec, gotSemi bool) { if p.trace { defer un(trace(p, "ConstSpec")); } idents := p.parseIdentList(nil); typ := p.tryType(); var values []ast.Expr; if typ != nil || p.tok == token.ASSIGN { p.expect(token.ASSIGN); values = p.parseExpressionList(); } comment, gotSemi := p.parseComment(getSemi); return &ast.ValueSpec{doc, idents, typ, values, comment}, gotSemi; } func parseTypeSpec(p *parser, doc ast.Comments, getSemi bool) (spec ast.Spec, gotSemi bool) { if p.trace { defer un(trace(p, "TypeSpec")); } ident := p.parseIdent(); typ := p.parseType(); comment, gotSemi := p.parseComment(getSemi); return &ast.TypeSpec{doc, ident, typ, comment}, gotSemi; } func parseVarSpec(p *parser, doc ast.Comments, getSemi bool) (spec ast.Spec, gotSemi bool) { if p.trace { defer un(trace(p, "VarSpec")); } idents := p.parseIdentList(nil); typ := p.tryType(); var values []ast.Expr; if typ == nil || p.tok == token.ASSIGN { p.expect(token.ASSIGN); values = p.parseExpressionList(); } comment, gotSemi := p.parseComment(getSemi); return &ast.ValueSpec{doc, idents, typ, values, comment}, gotSemi; } func (p *parser) parseGenDecl(keyword token.Token, f parseSpecFunction, getSemi bool) (decl *ast.GenDecl, gotSemi bool) { if p.trace { defer un(trace(p, keyword.String() + "Decl")); } doc := p.getDoc(); pos := p.expect(keyword); var lparen, rparen token.Position; list := vector.New(0); if p.tok == token.LPAREN { lparen = p.pos; p.next(); for p.tok != token.RPAREN && p.tok != token.EOF { doc := p.getDoc(); spec, semi := f(p, doc, true); // consume semicolon if any list.Push(spec); if !semi { break; } } rparen = p.expect(token.RPAREN); if getSemi && p.tok == token.SEMICOLON { p.next(); gotSemi = true; } else { p.optSemi = true; } } else { spec, semi := f(p, doc, getSemi); list.Push(spec); gotSemi = semi; } // convert vector specs := make([]ast.Spec, list.Len()); for i := 0; i < list.Len(); i++ { specs[i] = list.At(i); } return &ast.GenDecl{doc, pos, keyword, lparen, specs, rparen}, gotSemi; } func (p *parser) parseReceiver() *ast.Field { if p.trace { defer un(trace(p, "Receiver")); } pos := p.pos; par := p.parseParameters(false); // must have exactly one receiver if len(par) != 1 || len(par) == 1 && len(par[0].Names) > 1 { p.errorExpected(pos, "exactly one receiver"); return &ast.Field{Type: &ast.BadExpr{noPos}}; } recv := par[0]; // recv type must be TypeName or *TypeName base := recv.Type; if ptr, isPtr := base.(*ast.StarExpr); isPtr { base = ptr.X; } if !isTypeName(base) { p.errorExpected(base.Pos(), "type name"); } return recv; } func (p *parser) parseFunctionDecl() *ast.FuncDecl { if p.trace { defer un(trace(p, "FunctionDecl")); } doc := p.getDoc(); pos := p.expect(token.FUNC); var recv *ast.Field; if p.tok == token.LPAREN { recv = p.parseReceiver(); } ident := p.parseIdent(); params, results := p.parseSignature(); var body *ast.BlockStmt; if p.tok == token.LBRACE { body = p.parseBlockStmt(); } return &ast.FuncDecl{doc, recv, ident, &ast.FuncType{pos, params, results}, body}; } func (p *parser) parseDeclaration(getSemi bool) (decl ast.Decl, gotSemi bool) { if p.trace { defer un(trace(p, "Declaration")); } var f parseSpecFunction; switch p.tok { case token.CONST: f = parseConstSpec; case token.TYPE: f = parseTypeSpec; case token.VAR: f = parseVarSpec; case token.FUNC: decl = p.parseFunctionDecl(); // Do not use parseComment here to consume a semicolon // because we don't want to remove a trailing comment // from the list of unassociated comments. if getSemi && p.tok == token.SEMICOLON { p.next(); gotSemi = true; } return decl, gotSemi; default: pos := p.pos; p.errorExpected(pos, "declaration"); decl = &ast.BadDecl{pos}; gotSemi = getSemi && p.tok == token.SEMICOLON; p.next(); // make progress in any case return decl, gotSemi; } decl, gotSemi = p.parseGenDecl(p.tok, f, getSemi); // TODO 6g/spec issue return; } // ---------------------------------------------------------------------------- // Packages // The mode parameter to the Parse function is a set of flags (or 0). // They control the amount of source code parsed and other optional // parser functionality. // const ( PackageClauseOnly uint = 1 << iota; // parsing stops after package clause ImportsOnly; // parsing stops after import declarations ParseComments; // parse comments and add them to AST Trace; // print a trace of parsed productions ) func (p *parser) parsePackage() *ast.Program { if p.trace { defer un(trace(p, "Program")); } // package clause comment := p.getDoc(); pos := p.expect(token.PACKAGE); ident := p.parseIdent(); var decls []ast.Decl; // Don't bother parsing the rest if we had errors already. // Likely not a Go source file at all. if p.errors.Len() == 0 && p.mode & PackageClauseOnly == 0 { // import decls list := vector.New(0); for p.tok == token.IMPORT { decl, _ := p.parseGenDecl(token.IMPORT, parseImportSpec, true); // consume optional semicolon list.Push(decl); } if p.mode & ImportsOnly == 0 { // rest of package body for p.tok != token.EOF { decl, _ := p.parseDeclaration(true); // consume optional semicolon list.Push(decl); } } // convert declaration list decls = make([]ast.Decl, list.Len()); for i := 0; i < list.Len(); i++ { decls[i] = list.At(i).(ast.Decl); } } // convert comments list // 1) determine number of remaining comments n := 0; for i := 0; i < p.comments.Len(); i++ { if p.comments.At(i) != nil { n++; } } // 2) convert the remaining comments comments := make([]*ast.Comment, n); for i, j := 0, 0; i < p.comments.Len(); i++ { if p.comments.At(i) != nil { comments[j] = p.comments.At(i).(*ast.Comment); j++; } } return &ast.Program{comment, pos, ident, decls, comments}; } // ---------------------------------------------------------------------------- // Parsing of entire programs. func readSource(src interface{}) ([]byte, os.Error) { if src != nil { switch s := src.(type) { case string: return strings.Bytes(s), nil; case []byte: return s, nil; case *bytes.Buffer: // is io.Reader, but src is already available in []byte form if s != nil { return s.Data(), nil; } case io.Reader: var buf bytes.Buffer; n, err := io.Copy(s, &buf); if err != nil { return nil, err; } return buf.Data(), nil; } } return nil, os.ErrorString("invalid source"); } // scannerMode returns the scanner mode bits given the parser's mode bits. func scannerMode(mode uint) uint { if mode & ParseComments != 0 { return scanner.ScanComments; } return 0; } // Parse parses a Go program. // // The program source src may be provided in a variety of formats. At the // moment the following types are supported: string, []byte, and io.Reader. // The mode parameter controls the amount of source text parsed and other // optional parser functionality. // // Parse returns a complete AST if no error occured. Otherwise, if the // source couldn't be read, the returned program is nil and the error // indicates the specific failure. If the source was read but syntax // errors were found, the result is a partial AST (with ast.BadX nodes // representing the fragments of erroneous source code) and an ErrorList // describing the syntax errors. // func Parse(src interface{}, mode uint) (*ast.Program, os.Error) { data, err := readSource(src); if err != nil { return nil, err; } // initialize parser state var p parser; p.errors.Init(0); p.scanner.Init(data, &p, scannerMode(mode)); p.mode = mode; p.trace = mode & Trace != 0; // for convenience (p.trace is used frequently) p.comments.Init(0); p.next(); // parse program prog := p.parsePackage(); // convert errors list, if any if p.errors.Len() > 0 { errors := make(ErrorList, p.errors.Len()); for i := 0; i < p.errors.Len(); i++ { errors[i] = p.errors.At(i).(*Error); } return prog, errors; } return prog, nil; }