350377 - function () { return 3..();}

Status: RESOLVED INVALID

(Core :: JavaScript Engine, defect)

Description

[Biju]

by running Jesse's fuzzer (see bug 349611 ) found following....

for:-
------

function () {
    return 3..(<x/>);
}

2 Issues
--------

1)
* Compile with out syntax error
* Execution gives syntax error

3)
* uneval convert 3..(<x/>); to 3.(<x/>); 
* again uneval convert 3.(<x/>); to 3(<x/>); 


this also true for following

function () {
    return 3.(<x/>);
}

function () {
    return 3(<x/>);
}
function () {
    return 3(a);
}
function () {
    return 3('');
}
function () {
    return 3();
}

Comment 1

[Brendan Eich [:brendan]]

3. is a floating point literal.  3..(<x/>) is a filtering predicate operator whose left operand is 3. and whose filter expession is <x/>.  The lack of an early error is allowed by ECMA-262, see Chapter 16 -- in fact it's that chapter that allows early SyntaxErrors, where the default in the spec is all at runtime.

Any reason this is not INVALID?

/be

Comment 2

[Jesse Ruderman]

Regarding "issue 1", I see a TypeError, not a SyntaxError, so it's not surprising that it happens during runtime.

> 3..(<x/>)
TypeError on line 1: XML filtering predicate operator called on incompatible Number


Regarding "issue 3", I can't reproduce; maybe this was fixed in another bug.

> function () { return 3..(<x/>); }
function () { return (3).(<x/>); }

> function () { return (3).(<x/>); }
function () { return (3).(<x/>); }

Comment 3

[Biju]

(In reply to comment #1)
> Any reason this is not INVALID?

if you want to


(In reply to comment #2)
> Regarding "issue 1", I see a TypeError, not a SyntaxError, so it's not
> surprising that it happens during runtime.

right

Comment 4

[Brendan Eich [:brendan]]

Don't be shy about marking a bug INVALID.  If you don't think it's INVALID, please say why.  If you agree, no shame in so marking.  Lots of bugs, valid and not ;-).

/be