/* -*- Mode: java; tab-width: 8; indent-tabs-mode: nil; c-basic-offset: 4 -*-
*
* The contents of this file are subject to the Netscape Public
* License Version 1.1 (the "License"); you may not use this file
* except in compliance with the License. You may obtain a copy of
* the License at http://www.mozilla.org/NPL/
*
* Software distributed under the License is distributed on an "AS
* IS" basis, WITHOUT WARRANTY OF ANY KIND, either express oqr
* implied. See the License for the specific language governing
* rights and limitations under the License.
*
* The Original Code is Rhino code, released
* May 6, 1999.
*
* The Initial Developer of the Original Code is Netscape
* Communications Corporation. Portions created by Netscape are
* Copyright (C) 1997-1999 Netscape Communications Corporation. All
* Rights Reserved.
*
* Contributor(s):
* Mike Ang
* Mike McCabe
*
* Alternatively, the contents of this file may be used under the
* terms of the GNU Public License (the "GPL"), in which case the
* provisions of the GPL are applicable instead of those above.
* If you wish to allow use of your version of this file only
* under the terms of the GPL and not to allow others to use your
* version of this file under the NPL, indicate your decision by
* deleting the provisions above and replace them with the notice
* and other provisions required by the GPL. If you do not delete
* the provisions above, a recipient may use your version of this
* file under either the NPL or the GPL.
*/
// Modified by Google
package com.google.gwt.dev.js.rhino;
import java.io.IOException;
/**
* This class implements the JavaScript parser.
*
* It is based on the C source files jsparse.c and jsparse.h in the jsref
* package.
*
* @see TokenStream
*
* @author Mike McCabe
* @author Brendan Eich
*/
public class Parser {
public Parser(IRFactory nf) {
this.nf = nf;
}
private void mustMatchToken(TokenStream ts, int toMatch, String messageId)
throws IOException, JavaScriptException {
int tt;
if ((tt = ts.getToken()) != toMatch) {
reportError(ts, messageId);
ts.ungetToken(tt); // In case the parser decides to continue
}
}
private void reportError(TokenStream ts, String messageId)
throws JavaScriptException {
this.ok = false;
ts.reportSyntaxError(messageId, null);
/*
* Throw an exception to unwind the recursive descent parse. We use
* JavaScriptException here even though it is really a different use of the
* exception than it is usually used for.
*/
throw new JavaScriptException(messageId);
}
/*
* Build a parse tree from the given TokenStream.
*
* @param ts the TokenStream to parse
*
* @return an Object representing the parsed program. If the parse fails, null
* will be returned. (The parse failure will result in a call to the current
* Context's ErrorReporter.)
*/
public Object parse(TokenStream ts) throws IOException {
this.ok = true;
sourceTop = 0;
functionNumber = 0;
int tt; // last token from getToken();
int baseLineno = ts.getLineno(); // line number where source starts
/*
* so we have something to add nodes to until we've collected all the source
*/
Object tempBlock = nf.createLeaf(TokenStream.BLOCK);
((Node) tempBlock).setIsSyntheticBlock(true);
while (true) {
ts.flags |= ts.TSF_REGEXP;
tt = ts.getToken();
ts.flags &= ~ts.TSF_REGEXP;
if (tt <= ts.EOF) {
break;
}
if (tt == ts.FUNCTION) {
try {
nf.addChildToBack(tempBlock, function(ts, false));
} catch (JavaScriptException e) {
this.ok = false;
break;
}
} else {
ts.ungetToken(tt);
nf.addChildToBack(tempBlock, statement(ts));
}
}
if (!this.ok) {
// XXX ts.clearPushback() call here?
return null;
}
Object pn = nf.createScript(tempBlock, ts.getSourceName(), baseLineno, ts
.getLineno(), sourceToString(0));
((Node) pn).setIsSyntheticBlock(true);
return pn;
}
/*
* The C version of this function takes an argument list, which doesn't seem
* to be needed for tree generation... it'd only be useful for checking
* argument hiding, which I'm not doing anyway...
*/
private Object parseFunctionBody(TokenStream ts) throws IOException {
int oldflags = ts.flags;
ts.flags &= ~(TokenStream.TSF_RETURN_EXPR | TokenStream.TSF_RETURN_VOID);
ts.flags |= TokenStream.TSF_FUNCTION;
Object pn = nf.createBlock(ts.getLineno());
try {
int tt;
while ((tt = ts.peekToken()) > ts.EOF && tt != ts.RC) {
if (tt == TokenStream.FUNCTION) {
ts.getToken();
nf.addChildToBack(pn, function(ts, false));
} else {
nf.addChildToBack(pn, statement(ts));
}
}
} catch (JavaScriptException e) {
this.ok = false;
} finally {
// also in finally block:
// flushNewLines, clearPushback.
ts.flags = oldflags;
}
return pn;
}
private Object function(TokenStream ts, boolean isExpr) throws IOException,
JavaScriptException {
int baseLineno = ts.getLineno(); // line number where source starts
String name;
Object memberExprNode = null;
if (ts.matchToken(ts.NAME)) {
name = ts.getString();
if (!ts.matchToken(ts.LP)) {
if (Context.getContext().hasFeature(
Context.FEATURE_MEMBER_EXPR_AS_FUNCTION_NAME)) {
// Extension to ECMA: if 'function <name>' does not follow
// by '(', assume <name> starts memberExpr
sourceAddString(ts.NAME, name);
Object memberExprHead = nf.createName(name);
name = null;
memberExprNode = memberExprTail(ts, false, memberExprHead);
}
mustMatchToken(ts, ts.LP, "msg.no.paren.parms");
}
} else if (ts.matchToken(ts.LP)) {
// Anonymous function
name = null;
} else {
name = null;
if (Context.getContext().hasFeature(
Context.FEATURE_MEMBER_EXPR_AS_FUNCTION_NAME)) {
// Note that memberExpr can not start with '(' like
// in (1+2).toString, because 'function (' already
// processed as anonymous function
memberExprNode = memberExpr(ts, false);
}
mustMatchToken(ts, ts.LP, "msg.no.paren.parms");
}
if (memberExprNode != null) {
// transform 'function' <memberExpr> to <memberExpr> = function
// even in the decompilated source
sourceAdd((char) ts.ASSIGN);
sourceAdd((char) ts.NOP);
}
// save a reference to the function in the enclosing source.
sourceAdd((char) ts.FUNCTION);
sourceAdd((char) functionNumber);
++functionNumber;
// Save current source top to restore it on exit not to include
// function to parent source
int savedSourceTop = sourceTop;
int savedFunctionNumber = functionNumber;
Object args;
Object body;
String source;
try {
functionNumber = 0;
// FUNCTION as the first token in a Source means it's a function
// definition, and not a reference.
sourceAdd((char) ts.FUNCTION);
if (name != null) {
sourceAddString(ts.NAME, name);
}
sourceAdd((char) ts.LP);
args = nf.createLeaf(ts.LP);
if (!ts.matchToken(ts.GWT)) {
boolean first = true;
do {
if (!first)
sourceAdd((char) ts.COMMA);
first = false;
mustMatchToken(ts, ts.NAME, "msg.no.parm");
String s = ts.getString();
nf.addChildToBack(args, nf.createName(s));
sourceAddString(ts.NAME, s);
} while (ts.matchToken(ts.COMMA));
mustMatchToken(ts, ts.GWT, "msg.no.paren.after.parms");
}
sourceAdd((char) ts.GWT);
mustMatchToken(ts, ts.LC, "msg.no.brace.body");
sourceAdd((char) ts.LC);
sourceAdd((char) ts.EOL);
body = parseFunctionBody(ts);
mustMatchToken(ts, ts.RC, "msg.no.brace.after.body");
sourceAdd((char) ts.RC);
// skip the last EOL so nested functions work...
// name might be null;
source = sourceToString(savedSourceTop);
} finally {
sourceTop = savedSourceTop;
functionNumber = savedFunctionNumber;
}
Object pn = nf.createFunction(name, args, body, ts.getSourceName(),
baseLineno, ts.getLineno(), source, isExpr || memberExprNode != null);
if (memberExprNode != null) {
pn = nf.createBinary(ts.ASSIGN, ts.NOP, memberExprNode, pn);
}
// Add EOL but only if function is not part of expression, in which
// case it gets SEMI + EOL from Statement.
if (!isExpr) {
if (memberExprNode != null) {
// Add ';' to make 'function x.f(){}' and 'x.f = function(){}'
// to print the same strings when decompiling
sourceAdd((char) ts.SEMI);
}
sourceAdd((char) ts.EOL);
wellTerminated(ts, ts.FUNCTION);
}
return pn;
}
private Object statements(TokenStream ts) throws IOException {
Object pn = nf.createBlock(ts.getLineno());
int tt;
while ((tt = ts.peekToken()) > ts.EOF && tt != ts.RC) {
nf.addChildToBack(pn, statement(ts));
}
return pn;
}
private Object condition(TokenStream ts) throws IOException,
JavaScriptException {
Object pn;
mustMatchToken(ts, ts.LP, "msg.no.paren.cond");
sourceAdd((char) ts.LP);
pn = expr(ts, false);
mustMatchToken(ts, ts.GWT, "msg.no.paren.after.cond");
sourceAdd((char) ts.GWT);
// there's a check here in jsparse.c that corrects = to ==
return pn;
}
private boolean wellTerminated(TokenStream ts, int lastExprType)
throws IOException, JavaScriptException {
int tt = ts.peekTokenSameLine();
if (tt == ts.ERROR) {
return false;
}
if (tt != ts.EOF && tt != ts.EOL && tt != ts.SEMI && tt != ts.RC) {
int version = Context.getContext().getLanguageVersion();
if ((tt == ts.FUNCTION || lastExprType == ts.FUNCTION)
&& (version < Context.VERSION_1_2)) {
/*
* Checking against version < 1.2 and version >= 1.0 in the above line
* breaks old javascript, so we keep it this way for now... XXX warning
* needed?
*/
return true;
} else {
reportError(ts, "msg.no.semi.stmt");
}
}
return true;
}
// match a NAME; return null if no match.
private String matchLabel(TokenStream ts) throws IOException,
JavaScriptException {
int lineno = ts.getLineno();
String label = null;
int tt;
tt = ts.peekTokenSameLine();
if (tt == ts.NAME) {
ts.getToken();
label = ts.getString();
}
if (lineno == ts.getLineno())
wellTerminated(ts, ts.ERROR);
return label;
}
private Object statement(TokenStream ts) throws IOException {
try {
return statementHelper(ts);
} catch (JavaScriptException e) {
// skip to end of statement
int lineno = ts.getLineno();
int t;
do {
t = ts.getToken();
} while (t != TokenStream.SEMI && t != TokenStream.EOL
&& t != TokenStream.EOF && t != TokenStream.ERROR);
return nf.createExprStatement(nf.createName("error"), lineno);
}
}
/**
* Whether the "catch (e: e instanceof Exception) { ... }" syntax is
* implemented.
*/
private Object statementHelper(TokenStream ts) throws IOException,
JavaScriptException {
Object pn = null;
// If skipsemi == true, don't add SEMI + EOL to source at the
// end of this statment. For compound statements, IF/FOR etc.
boolean skipsemi = false;
int tt;
int lastExprType = 0; // For wellTerminated. 0 to avoid warning.
tt = ts.getToken();
switch (tt) {
case TokenStream.IF: {
skipsemi = true;
sourceAdd((char) ts.IF);
int lineno = ts.getLineno();
Object cond = condition(ts);
sourceAdd((char) ts.LC);
sourceAdd((char) ts.EOL);
Object ifTrue = statement(ts);
Object ifFalse = null;
if (ts.matchToken(ts.ELSE)) {
sourceAdd((char) ts.RC);
sourceAdd((char) ts.ELSE);
sourceAdd((char) ts.LC);
sourceAdd((char) ts.EOL);
ifFalse = statement(ts);
}
sourceAdd((char) ts.RC);
sourceAdd((char) ts.EOL);
pn = nf.createIf(cond, ifTrue, ifFalse, lineno);
break;
}
case TokenStream.SWITCH: {
skipsemi = true;
sourceAdd((char) ts.SWITCH);
pn = nf.createSwitch(ts.getLineno());
Object cur_case = null; // to kill warning
Object case_statements;
mustMatchToken(ts, ts.LP, "msg.no.paren.switch");
sourceAdd((char) ts.LP);
nf.addChildToBack(pn, expr(ts, false));
mustMatchToken(ts, ts.GWT, "msg.no.paren.after.switch");
sourceAdd((char) ts.GWT);
mustMatchToken(ts, ts.LC, "msg.no.brace.switch");
sourceAdd((char) ts.LC);
sourceAdd((char) ts.EOL);
while ((tt = ts.getToken()) != ts.RC && tt != ts.EOF) {
switch (tt) {
case TokenStream.CASE:
sourceAdd((char) ts.CASE);
cur_case = nf.createUnary(ts.CASE, expr(ts, false));
sourceAdd((char) ts.COLON);
sourceAdd((char) ts.EOL);
break;
case TokenStream.DEFAULT:
cur_case = nf.createLeaf(ts.DEFAULT);
sourceAdd((char) ts.DEFAULT);
sourceAdd((char) ts.COLON);
sourceAdd((char) ts.EOL);
// XXX check that there isn't more than one default
break;
default:
reportError(ts, "msg.bad.switch");
break;
}
mustMatchToken(ts, ts.COLON, "msg.no.colon.case");
case_statements = nf.createLeaf(TokenStream.BLOCK);
((Node) case_statements).setIsSyntheticBlock(true);
while ((tt = ts.peekToken()) != ts.RC && tt != ts.CASE
&& tt != ts.DEFAULT && tt != ts.EOF) {
nf.addChildToBack(case_statements, statement(ts));
}
// assert cur_case
nf.addChildToBack(cur_case, case_statements);
nf.addChildToBack(pn, cur_case);
}
sourceAdd((char) ts.RC);
sourceAdd((char) ts.EOL);
break;
}
case TokenStream.WHILE: {
skipsemi = true;
sourceAdd((char) ts.WHILE);
int lineno = ts.getLineno();
Object cond = condition(ts);
sourceAdd((char) ts.LC);
sourceAdd((char) ts.EOL);
Object body = statement(ts);
sourceAdd((char) ts.RC);
sourceAdd((char) ts.EOL);
pn = nf.createWhile(cond, body, lineno);
break;
}
case TokenStream.DO: {
sourceAdd((char) ts.DO);
sourceAdd((char) ts.LC);
sourceAdd((char) ts.EOL);
int lineno = ts.getLineno();
Object body = statement(ts);
sourceAdd((char) ts.RC);
mustMatchToken(ts, ts.WHILE, "msg.no.while.do");
sourceAdd((char) ts.WHILE);
Object cond = condition(ts);
pn = nf.createDoWhile(body, cond, lineno);
break;
}
case TokenStream.FOR: {
skipsemi = true;
sourceAdd((char) ts.FOR);
int lineno = ts.getLineno();
Object init; // Node init is also foo in 'foo in Object'
Object cond; // Node cond is also object in 'foo in Object'
Object incr = null; // to kill warning
Object body;
mustMatchToken(ts, ts.LP, "msg.no.paren.for");
sourceAdd((char) ts.LP);
tt = ts.peekToken();
if (tt == ts.SEMI) {
init = nf.createLeaf(ts.VOID);
} else {
if (tt == ts.VAR) {
// set init to a var list or initial
ts.getToken(); // throw away the 'var' token
init = variables(ts, true);
} else {
init = expr(ts, true);
}
}
tt = ts.peekToken();
if (tt == ts.RELOP && ts.getOp() == ts.IN) {
ts.matchToken(ts.RELOP);
sourceAdd((char) ts.IN);
// 'cond' is the object over which we're iterating
cond = expr(ts, false);
} else { // ordinary for loop
mustMatchToken(ts, ts.SEMI, "msg.no.semi.for");
sourceAdd((char) ts.SEMI);
if (ts.peekToken() == ts.SEMI) {
// no loop condition
cond = nf.createLeaf(ts.VOID);
} else {
cond = expr(ts, false);
}
mustMatchToken(ts, ts.SEMI, "msg.no.semi.for.cond");
sourceAdd((char) ts.SEMI);
if (ts.peekToken() == ts.GWT) {
incr = nf.createLeaf(ts.VOID);
} else {
incr = expr(ts, false);
}
}
mustMatchToken(ts, ts.GWT, "msg.no.paren.for.ctrl");
sourceAdd((char) ts.GWT);
sourceAdd((char) ts.LC);
sourceAdd((char) ts.EOL);
body = statement(ts);
sourceAdd((char) ts.RC);
sourceAdd((char) ts.EOL);
if (incr == null) {
// cond could be null if 'in obj' got eaten by the init node.
pn = nf.createForIn(init, cond, body, lineno);
} else {
pn = nf.createFor(init, cond, incr, body, lineno);
}
break;
}
case TokenStream.TRY: {
int lineno = ts.getLineno();
Object tryblock;
Object catchblocks = null;
Object finallyblock = null;
skipsemi = true;
sourceAdd((char) ts.TRY);
sourceAdd((char) ts.LC);
sourceAdd((char) ts.EOL);
tryblock = statement(ts);
sourceAdd((char) ts.RC);
sourceAdd((char) ts.EOL);
catchblocks = nf.createLeaf(TokenStream.BLOCK);
boolean sawDefaultCatch = false;
int peek = ts.peekToken();
if (peek == ts.CATCH) {
while (ts.matchToken(ts.CATCH)) {
if (sawDefaultCatch) {
reportError(ts, "msg.catch.unreachable");
}
sourceAdd((char) ts.CATCH);
mustMatchToken(ts, ts.LP, "msg.no.paren.catch");
sourceAdd((char) ts.LP);
mustMatchToken(ts, ts.NAME, "msg.bad.catchcond");
String varName = ts.getString();
sourceAddString(ts.NAME, varName);
Object catchCond = null;
if (ts.matchToken(ts.IF)) {
sourceAdd((char) ts.IF);
catchCond = expr(ts, false);
} else {
sawDefaultCatch = true;
}
mustMatchToken(ts, ts.GWT, "msg.bad.catchcond");
sourceAdd((char) ts.GWT);
mustMatchToken(ts, ts.LC, "msg.no.brace.catchblock");
sourceAdd((char) ts.LC);
sourceAdd((char) ts.EOL);
nf.addChildToBack(catchblocks, nf.createCatch(varName, catchCond,
statements(ts), ts.getLineno()));
mustMatchToken(ts, ts.RC, "msg.no.brace.after.body");
sourceAdd((char) ts.RC);
sourceAdd((char) ts.EOL);
}
} else if (peek != ts.FINALLY) {
mustMatchToken(ts, ts.FINALLY, "msg.try.no.catchfinally");
}
if (ts.matchToken(ts.FINALLY)) {
sourceAdd((char) ts.FINALLY);
sourceAdd((char) ts.LC);
sourceAdd((char) ts.EOL);
finallyblock = statement(ts);
sourceAdd((char) ts.RC);
sourceAdd((char) ts.EOL);
}
pn = nf.createTryCatchFinally(tryblock, catchblocks, finallyblock,
lineno);
break;
}
case TokenStream.THROW: {
int lineno = ts.getLineno();
sourceAdd((char) ts.THROW);
pn = nf.createThrow(expr(ts, false), lineno);
if (lineno == ts.getLineno())
wellTerminated(ts, ts.ERROR);
break;
}
case TokenStream.BREAK: {
int lineno = ts.getLineno();
sourceAdd((char) ts.BREAK);
// matchLabel only matches if there is one
String label = matchLabel(ts);
if (label != null) {
sourceAddString(ts.NAME, label);
}
pn = nf.createBreak(label, lineno);
break;
}
case TokenStream.CONTINUE: {
int lineno = ts.getLineno();
sourceAdd((char) ts.CONTINUE);
// matchLabel only matches if there is one
String label = matchLabel(ts);
if (label != null) {
sourceAddString(ts.NAME, label);
}
pn = nf.createContinue(label, lineno);
break;
}
case TokenStream.DEBUGGER: {
int lineno = ts.getLineno();
sourceAdd((char) ts.DEBUGGER);
pn = nf.createDebugger(lineno);
break;
}
case TokenStream.WITH: {
// bruce: we don't support this is JSNI code because it's impossible
// to identify bindings even passably well
//
reportError(ts, "msg.jsni.unsupported.with");
skipsemi = true;
sourceAdd((char) ts.WITH);
int lineno = ts.getLineno();
mustMatchToken(ts, ts.LP, "msg.no.paren.with");
sourceAdd((char) ts.LP);
Object obj = expr(ts, false);
mustMatchToken(ts, ts.GWT, "msg.no.paren.after.with");
sourceAdd((char) ts.GWT);
sourceAdd((char) ts.LC);
sourceAdd((char) ts.EOL);
Object body = statement(ts);
sourceAdd((char) ts.RC);
sourceAdd((char) ts.EOL);
pn = nf.createWith(obj, body, lineno);
break;
}
case TokenStream.VAR: {
int lineno = ts.getLineno();
pn = variables(ts, false);
if (ts.getLineno() == lineno)
wellTerminated(ts, ts.ERROR);
break;
}
case TokenStream.RETURN: {
Object retExpr = null;
int lineno = 0;
sourceAdd((char) ts.RETURN);
// bail if we're not in a (toplevel) function
if ((ts.flags & ts.TSF_FUNCTION) == 0)
reportError(ts, "msg.bad.return");
/* This is ugly, but we don't want to require a semicolon. */
ts.flags |= ts.TSF_REGEXP;
tt = ts.peekTokenSameLine();
ts.flags &= ~ts.TSF_REGEXP;
if (tt != ts.EOF && tt != ts.EOL && tt != ts.SEMI && tt != ts.RC) {
lineno = ts.getLineno();
retExpr = expr(ts, false);
if (ts.getLineno() == lineno)
wellTerminated(ts, ts.ERROR);
ts.flags |= ts.TSF_RETURN_EXPR;
} else {
ts.flags |= ts.TSF_RETURN_VOID;
}
// XXX ASSERT pn
pn = nf.createReturn(retExpr, lineno);
break;
}
case TokenStream.LC:
skipsemi = true;
pn = statements(ts);
mustMatchToken(ts, ts.RC, "msg.no.brace.block");
break;
case TokenStream.ERROR:
// Fall thru, to have a node for error recovery to work on
case TokenStream.EOL:
case TokenStream.SEMI:
pn = nf.createLeaf(ts.VOID);
skipsemi = true;
break;
default: {
lastExprType = tt;
int tokenno = ts.getTokenno();
ts.ungetToken(tt);
int lineno = ts.getLineno();
pn = expr(ts, false);
if (ts.peekToken() == ts.COLON) {
/*
* check that the last thing the tokenizer returned was a NAME and
* that only one token was consumed.
*/
if (lastExprType != ts.NAME || (ts.getTokenno() != tokenno))
reportError(ts, "msg.bad.label");
ts.getToken(); // eat the COLON
/*
* in the C source, the label is associated with the statement that
* follows: nf.addChildToBack(pn, statement(ts));
*/
String name = ts.getString();
pn = nf.createLabel(name, lineno);
// bruce: added to make it easier to bind labels to the
// statements they modify
//
nf.addChildToBack(pn, statement(ts));
// depend on decompiling lookahead to guess that that
// last name was a label.
sourceAdd((char) ts.COLON);
sourceAdd((char) ts.EOL);
return pn;
}
if (lastExprType == ts.FUNCTION) {
if (nf.getLeafType(pn) != ts.FUNCTION) {
reportError(ts, "msg.syntax");
}
}
pn = nf.createExprStatement(pn, lineno);
/*
* Check explicitly against (multi-line) function statement.
*
* lastExprEndLine is a hack to fix an automatic semicolon insertion
* problem with function expressions; the ts.getLineno() == lineno check
* was firing after a function definition even though the next statement
* was on a new line, because speculative getToken calls advanced the
* line number even when they didn't succeed.
*/
if (ts.getLineno() == lineno
|| (lastExprType == ts.FUNCTION && ts.getLineno() == lastExprEndLine)) {
wellTerminated(ts, lastExprType);
}
break;
}
}
ts.matchToken(ts.SEMI);
if (!skipsemi) {
sourceAdd((char) ts.SEMI);
sourceAdd((char) ts.EOL);
}
return pn;
}
private Object variables(TokenStream ts, boolean inForInit)
throws IOException, JavaScriptException {
Object pn = nf.createVariables(ts.getLineno());
boolean first = true;
sourceAdd((char) ts.VAR);
for (;;) {
Object name;
Object init;
mustMatchToken(ts, ts.NAME, "msg.bad.var");
String s = ts.getString();
if (!first)
sourceAdd((char) ts.COMMA);
first = false;
sourceAddString(ts.NAME, s);
name = nf.createName(s);
// omitted check for argument hiding
if (ts.matchToken(ts.ASSIGN)) {
if (ts.getOp() != ts.NOP)
reportError(ts, "msg.bad.var.init");
sourceAdd((char) ts.ASSIGN);
sourceAdd((char) ts.NOP);
init = assignExpr(ts, inForInit);
nf.addChildToBack(name, init);
}
nf.addChildToBack(pn, name);
if (!ts.matchToken(ts.COMMA))
break;
}
return pn;
}
private Object expr(TokenStream ts, boolean inForInit) throws IOException,
JavaScriptException {
Object pn = assignExpr(ts, inForInit);
while (ts.matchToken(ts.COMMA)) {
sourceAdd((char) ts.COMMA);
pn = nf.createBinary(ts.COMMA, pn, assignExpr(ts, inForInit));
}
return pn;
}
private Object assignExpr(TokenStream ts, boolean inForInit)
throws IOException, JavaScriptException {
Object pn = condExpr(ts, inForInit);
if (ts.matchToken(ts.ASSIGN)) {
// omitted: "invalid assignment left-hand side" check.
sourceAdd((char) ts.ASSIGN);
sourceAdd((char) ts.getOp());
pn = nf
.createBinary(ts.ASSIGN, ts.getOp(), pn, assignExpr(ts, inForInit));
}
return pn;
}
private Object condExpr(TokenStream ts, boolean inForInit)
throws IOException, JavaScriptException {
Object ifTrue;
Object ifFalse;
Object pn = orExpr(ts, inForInit);
if (ts.matchToken(ts.HOOK)) {
sourceAdd((char) ts.HOOK);
ifTrue = assignExpr(ts, false);
mustMatchToken(ts, ts.COLON, "msg.no.colon.cond");
sourceAdd((char) ts.COLON);
ifFalse = assignExpr(ts, inForInit);
return nf.createTernary(pn, ifTrue, ifFalse);
}
return pn;
}
private Object orExpr(TokenStream ts, boolean inForInit) throws IOException,
JavaScriptException {
Object pn = andExpr(ts, inForInit);
if (ts.matchToken(ts.OR)) {
sourceAdd((char) ts.OR);
pn = nf.createBinary(ts.OR, pn, orExpr(ts, inForInit));
}
return pn;
}
private Object andExpr(TokenStream ts, boolean inForInit) throws IOException,
JavaScriptException {
Object pn = bitOrExpr(ts, inForInit);
if (ts.matchToken(ts.AND)) {
sourceAdd((char) ts.AND);
pn = nf.createBinary(ts.AND, pn, andExpr(ts, inForInit));
}
return pn;
}
private Object bitOrExpr(TokenStream ts, boolean inForInit)
throws IOException, JavaScriptException {
Object pn = bitXorExpr(ts, inForInit);
while (ts.matchToken(ts.BITOR)) {
sourceAdd((char) ts.BITOR);
pn = nf.createBinary(ts.BITOR, pn, bitXorExpr(ts, inForInit));
}
return pn;
}
private Object bitXorExpr(TokenStream ts, boolean inForInit)
throws IOException, JavaScriptException {
Object pn = bitAndExpr(ts, inForInit);
while (ts.matchToken(ts.BITXOR)) {
sourceAdd((char) ts.BITXOR);
pn = nf.createBinary(ts.BITXOR, pn, bitAndExpr(ts, inForInit));
}
return pn;
}
private Object bitAndExpr(TokenStream ts, boolean inForInit)
throws IOException, JavaScriptException {
Object pn = eqExpr(ts, inForInit);
while (ts.matchToken(ts.BITAND)) {
sourceAdd((char) ts.BITAND);
pn = nf.createBinary(ts.BITAND, pn, eqExpr(ts, inForInit));
}
return pn;
}
private Object eqExpr(TokenStream ts, boolean inForInit) throws IOException,
JavaScriptException {
Object pn = relExpr(ts, inForInit);
while (ts.matchToken(ts.EQOP)) {
sourceAdd((char) ts.EQOP);
sourceAdd((char) ts.getOp());
pn = nf.createBinary(ts.EQOP, ts.getOp(), pn, relExpr(ts, inForInit));
}
return pn;
}
private Object relExpr(TokenStream ts, boolean inForInit) throws IOException,
JavaScriptException {
Object pn = shiftExpr(ts);
while (ts.matchToken(ts.RELOP)) {
int op = ts.getOp();
if (inForInit && op == ts.IN) {
ts.ungetToken(ts.RELOP);
break;
}
sourceAdd((char) ts.RELOP);
sourceAdd((char) op);
pn = nf.createBinary(ts.RELOP, op, pn, shiftExpr(ts));
}
return pn;
}
private Object shiftExpr(TokenStream ts) throws IOException,
JavaScriptException {
Object pn = addExpr(ts);
while (ts.matchToken(ts.SHOP)) {
sourceAdd((char) ts.SHOP);
sourceAdd((char) ts.getOp());
pn = nf.createBinary(ts.SHOP, ts.getOp(), pn, addExpr(ts));
}
return pn;
}
private Object addExpr(TokenStream ts) throws IOException,
JavaScriptException {
int tt;
Object pn = mulExpr(ts);
while ((tt = ts.getToken()) == ts.ADD || tt == ts.SUB) {
sourceAdd((char) tt);
// flushNewLines
pn = nf.createBinary(tt, pn, mulExpr(ts));
}
ts.ungetToken(tt);
return pn;
}
private Object mulExpr(TokenStream ts) throws IOException,
JavaScriptException {
int tt;
Object pn = unaryExpr(ts);
while ((tt = ts.peekToken()) == ts.MUL || tt == ts.DIV || tt == ts.MOD) {
tt = ts.getToken();
sourceAdd((char) tt);
pn = nf.createBinary(tt, pn, unaryExpr(ts));
}
return pn;
}
private Object unaryExpr(TokenStream ts) throws IOException,
JavaScriptException {
int tt;
ts.flags |= ts.TSF_REGEXP;
tt = ts.getToken();
ts.flags &= ~ts.TSF_REGEXP;
switch (tt) {
case TokenStream.UNARYOP:
sourceAdd((char) ts.UNARYOP);
sourceAdd((char) ts.getOp());
return nf.createUnary(ts.UNARYOP, ts.getOp(), unaryExpr(ts));
case TokenStream.ADD:
case TokenStream.SUB:
sourceAdd((char) ts.UNARYOP);
sourceAdd((char) tt);
return nf.createUnary(ts.UNARYOP, tt, unaryExpr(ts));
case TokenStream.INC:
case TokenStream.DEC:
sourceAdd((char) tt);
return nf.createUnary(tt, ts.PRE, memberExpr(ts, true));
case TokenStream.DELPROP:
sourceAdd((char) ts.DELPROP);
return nf.createUnary(ts.DELPROP, unaryExpr(ts));
case TokenStream.ERROR:
break;
default:
ts.ungetToken(tt);
int lineno = ts.getLineno();
Object pn = memberExpr(ts, true);
/*
* don't look across a newline boundary for a postfix incop.
*
* the rhino scanner seems to work differently than the js scanner here;
* in js, it works to have the line number check precede the peekToken
* calls. It'd be better if they had similar behavior...
*/
int peeked;
if (((peeked = ts.peekToken()) == ts.INC || peeked == ts.DEC)
&& ts.getLineno() == lineno) {
int pf = ts.getToken();
sourceAdd((char) pf);
return nf.createUnary(pf, ts.POST, pn);
}
return pn;
}
return nf.createName("err"); // Only reached on error. Try to continue.
}
private Object argumentList(TokenStream ts, Object listNode)
throws IOException, JavaScriptException {
boolean matched;
ts.flags |= ts.TSF_REGEXP;
matched = ts.matchToken(ts.GWT);
ts.flags &= ~ts.TSF_REGEXP;
if (!matched) {
boolean first = true;
do {
if (!first)
sourceAdd((char) ts.COMMA);
first = false;
nf.addChildToBack(listNode, assignExpr(ts, false));
} while (ts.matchToken(ts.COMMA));
mustMatchToken(ts, ts.GWT, "msg.no.paren.arg");
}
sourceAdd((char) ts.GWT);
return listNode;
}
private Object memberExpr(TokenStream ts, boolean allowCallSyntax)
throws IOException, JavaScriptException {
int tt;
Object pn;
/* Check for new expressions. */
ts.flags |= ts.TSF_REGEXP;
tt = ts.peekToken();
ts.flags &= ~ts.TSF_REGEXP;
if (tt == ts.NEW) {
/* Eat the NEW token. */
ts.getToken();
sourceAdd((char) ts.NEW);
/* Make a NEW node to append to. */
pn = nf.createLeaf(ts.NEW);
nf.addChildToBack(pn, memberExpr(ts, false));
if (ts.matchToken(ts.LP)) {
sourceAdd((char) ts.LP);
/* Add the arguments to pn, if any are supplied. */
pn = argumentList(ts, pn);
}
/*
* XXX there's a check in the C source against "too many constructor
* arguments" - how many do we claim to support?
*/
/*
* Experimental syntax: allow an object literal to follow a new
* expression, which will mean a kind of anonymous class built with the
* JavaAdapter. the object literal will be passed as an additional
* argument to the constructor.
*/
tt = ts.peekToken();
if (tt == ts.LC) {
nf.addChildToBack(pn, primaryExpr(ts));
}
} else {
pn = primaryExpr(ts);
}
return memberExprTail(ts, allowCallSyntax, pn);
}
private Object memberExprTail(TokenStream ts, boolean allowCallSyntax,
Object pn) throws IOException, JavaScriptException {
lastExprEndLine = ts.getLineno();
int tt;
while ((tt = ts.getToken()) > ts.EOF) {
if (tt == ts.DOT) {
sourceAdd((char) ts.DOT);
mustMatchToken(ts, ts.NAME, "msg.no.name.after.dot");
String s = ts.getString();
sourceAddString(ts.NAME, s);
pn = nf.createBinary(ts.DOT, pn, nf.createName(ts.getString()));
/*
* pn = nf.createBinary(ts.DOT, pn, memberExpr(ts)) is the version in
* Brendan's IR C version. Not in ECMA... does it reflect the 'new'
* operator syntax he mentioned?
*/
lastExprEndLine = ts.getLineno();
} else if (tt == ts.LB) {
sourceAdd((char) ts.LB);
pn = nf.createBinary(ts.LB, pn, expr(ts, false));
mustMatchToken(ts, ts.RB, "msg.no.bracket.index");
sourceAdd((char) ts.RB);
lastExprEndLine = ts.getLineno();
} else if (allowCallSyntax && tt == ts.LP) {
/* make a call node */
pn = nf.createUnary(ts.CALL, pn);
sourceAdd((char) ts.LP);
/* Add the arguments to pn, if any are supplied. */
pn = argumentList(ts, pn);
lastExprEndLine = ts.getLineno();
} else {
ts.ungetToken(tt);
break;
}
}
return pn;
}
private Object primaryExpr(TokenStream ts) throws IOException,
JavaScriptException {
int tt;
Object pn;
ts.flags |= ts.TSF_REGEXP;
tt = ts.getToken();
ts.flags &= ~ts.TSF_REGEXP;
switch (tt) {
case TokenStream.FUNCTION:
return function(ts, true);
case TokenStream.LB: {
sourceAdd((char) ts.LB);
pn = nf.createLeaf(ts.ARRAYLIT);
ts.flags |= ts.TSF_REGEXP;
boolean matched = ts.matchToken(ts.RB);
ts.flags &= ~ts.TSF_REGEXP;
if (!matched) {
boolean first = true;
do {
ts.flags |= ts.TSF_REGEXP;
tt = ts.peekToken();
ts.flags &= ~ts.TSF_REGEXP;
if (!first)
sourceAdd((char) ts.COMMA);
else
first = false;
if (tt == ts.RB) { // to fix [,,,].length behavior...
break;
}
if (tt == ts.COMMA) {
nf.addChildToBack(pn, nf.createLeaf(ts.PRIMARY, ts.UNDEFINED));
} else {
nf.addChildToBack(pn, assignExpr(ts, false));
}
} while (ts.matchToken(ts.COMMA));
mustMatchToken(ts, ts.RB, "msg.no.bracket.arg");
}
sourceAdd((char) ts.RB);
return nf.createArrayLiteral(pn);
}
case TokenStream.LC: {
pn = nf.createLeaf(ts.OBJLIT);
sourceAdd((char) ts.LC);
if (!ts.matchToken(ts.RC)) {
boolean first = true;
commaloop : do {
Object property;
if (!first)
sourceAdd((char) ts.COMMA);
else
first = false;
tt = ts.getToken();
switch (tt) {
// map NAMEs to STRINGs in object literal context.
case TokenStream.NAME:
case TokenStream.STRING:
String s = ts.getString();
sourceAddString(ts.NAME, s);
property = nf.createString(ts.getString());
break;
case TokenStream.NUMBER:
double n = ts.getNumber();
sourceAddNumber(n);
property = nf.createNumber(n);
break;
case TokenStream.RC:
// trailing comma is OK.
ts.ungetToken(tt);
break commaloop;
default:
reportError(ts, "msg.bad.prop");
break commaloop;
}
mustMatchToken(ts, ts.COLON, "msg.no.colon.prop");
// OBJLIT is used as ':' in object literal for
// decompilation to solve spacing ambiguity.
sourceAdd((char) ts.OBJLIT);
nf.addChildToBack(pn, property);
nf.addChildToBack(pn, assignExpr(ts, false));
} while (ts.matchToken(ts.COMMA));
mustMatchToken(ts, ts.RC, "msg.no.brace.prop");
}
sourceAdd((char) ts.RC);
return nf.createObjectLiteral(pn);
}
case TokenStream.LP:
/*
* Brendan's IR-jsparse.c makes a new node tagged with TOK_LP here...
* I'm not sure I understand why. Isn't the grouping already implicit in
* the structure of the parse tree? also TOK_LP is already overloaded (I
* think) in the C IR as 'function call.'
*/
sourceAdd((char) ts.LP);
pn = expr(ts, false);
sourceAdd((char) ts.GWT);
mustMatchToken(ts, ts.GWT, "msg.no.paren");
return pn;
case TokenStream.NAME:
String name = ts.getString();
sourceAddString(ts.NAME, name);
return nf.createName(name);
case TokenStream.NUMBER:
double n = ts.getNumber();
sourceAddNumber(n);
return nf.createNumber(n);
case TokenStream.STRING:
String s = ts.getString();
sourceAddString(ts.STRING, s);
return nf.createString(s);
case TokenStream.REGEXP: {
String flags = ts.regExpFlags;
ts.regExpFlags = null;
String re = ts.getString();
sourceAddString(ts.REGEXP, '/' + re + '/' + flags);
return nf.createRegExp(re, flags);
}
case TokenStream.PRIMARY:
sourceAdd((char) ts.PRIMARY);
sourceAdd((char) ts.getOp());
return nf.createLeaf(ts.PRIMARY, ts.getOp());
case TokenStream.RESERVED:
reportError(ts, "msg.reserved.id");
break;
case TokenStream.ERROR:
/* the scanner or one of its subroutines reported the error. */
break;
default:
reportError(ts, "msg.syntax");
break;
}
return null; // should never reach here
}
/**
* The following methods save decompilation information about the source.
* Source information is returned from the parser as a String associated with
* function nodes and with the toplevel script. When saved in the constant
* pool of a class, this string will be UTF-8 encoded, and token values will
* occupy a single byte.
*
* Source is saved (mostly) as token numbers. The tokens saved pretty much
* correspond to the token stream of a 'canonical' representation of the input
* program, as directed by the parser. (There were a few cases where tokens
* could have been left out where decompiler could easily reconstruct them,
* but I left them in for clarity). (I also looked adding source collection to
* TokenStream instead, where I could have limited the changes to a few lines
* in getToken... but this wouldn't have saved any space in the resulting
* source representation, and would have meant that I'd have to duplicate
* parser logic in the decompiler to disambiguate situations where newlines
* are important.) NativeFunction.decompile expands the tokens back into their
* string representations, using simple lookahead to correct spacing and
* indentation.
*
* Token types with associated ops (ASSIGN, SHOP, PRIMARY, etc.) are saved as
* two-token pairs. Number tokens are stored inline, as a NUMBER token, a
* character representing the type, and either 1 or 4 characters representing
* the bit-encoding of the number. String types NAME, STRING and OBJECT are
* currently stored as a token type, followed by a character giving the length
* of the string (assumed to be less than 2^16), followed by the characters of
* the string inlined into the source string. Changing this to some reference
* to to the string in the compiled class' constant pool would probably save a
* lot of space... but would require some method of deriving the final
* constant pool entry from information available at parse time.
*
* Nested functions need a similar mechanism... fortunately the nested
* functions for a given function are generated in source order. Nested
* functions are encoded as FUNCTION followed by a function number (encoded as
* a character), which is enough information to find the proper generated
* NativeFunction instance.
*
*/
private void sourceAdd(char c) {
if (sourceTop == sourceBuffer.length) {
increaseSourceCapacity(sourceTop + 1);
}
sourceBuffer[sourceTop] = c;
++sourceTop;
}
private void sourceAddString(int type, String str) {
int L = str.length();
// java string length < 2^16?
if (Context.check && L > Character.MAX_VALUE)
Context.codeBug();
if (sourceTop + L + 2 > sourceBuffer.length) {
increaseSourceCapacity(sourceTop + L + 2);
}
sourceAdd((char) type);
sourceAdd((char) L);
str.getChars(0, L, sourceBuffer, sourceTop);
sourceTop += L;
}
private void sourceAddNumber(double n) {
sourceAdd((char) TokenStream.NUMBER);
/*
* encode the number in the source stream. Save as NUMBER type (char | char
* char char char) where type is 'D' - double, 'S' - short, 'J' - long.
*
* We need to retain float vs. integer type info to keep the behavior of
* liveconnect type-guessing the same after decompilation. (Liveconnect
* tries to present 1.0 to Java as a float/double) OPT: This is no longer
* true. We could compress the format.
*
* This may not be the most space-efficient encoding; the chars created
* below may take up to 3 bytes in constant pool UTF-8 encoding, so a Double
* could take up to 12 bytes.
*/
long lbits = (long) n;
if (lbits != n) {
// if it's floating point, save as a Double bit pattern.
// (12/15/97 our scanner only returns Double for f.p.)
lbits = Double.doubleToLongBits(n);
sourceAdd('D');
sourceAdd((char) (lbits >> 48));
sourceAdd((char) (lbits >> 32));
sourceAdd((char) (lbits >> 16));
sourceAdd((char) lbits);
} else {
// we can ignore negative values, bc they're already prefixed
// by UNARYOP SUB
if (Context.check && lbits < 0)
Context.codeBug();
// will it fit in a char?
// this gives a short encoding for integer values up to 2^16.
if (lbits <= Character.MAX_VALUE) {
sourceAdd('S');
sourceAdd((char) lbits);
} else { // Integral, but won't fit in a char. Store as a long.
sourceAdd('J');
sourceAdd((char) (lbits >> 48));
sourceAdd((char) (lbits >> 32));
sourceAdd((char) (lbits >> 16));
sourceAdd((char) lbits);
}
}
}
private void increaseSourceCapacity(int minimalCapacity) {
// Call this only when capacity increase is must
if (Context.check && minimalCapacity <= sourceBuffer.length)
Context.codeBug();
int newCapacity = sourceBuffer.length * 2;
if (newCapacity < minimalCapacity) {
newCapacity = minimalCapacity;
}
char[] tmp = new char[newCapacity];
System.arraycopy(sourceBuffer, 0, tmp, 0, sourceTop);
sourceBuffer = tmp;
}
private String sourceToString(int offset) {
if (Context.check && (offset < 0 || sourceTop < offset))
Context.codeBug();
return new String(sourceBuffer, offset, sourceTop - offset);
}
private int lastExprEndLine; // Hack to handle function expr termination.
private IRFactory nf;
private ErrorReporter er;
private boolean ok; // Did the parse encounter an error?
private char[] sourceBuffer = new char[128];
private int sourceTop;
private int functionNumber;
}