RE: [ZEND-ENGINE-CVS] cvs: ZendEngine2(PHP_5_2) / zend_compile.c

OK cool. No time pressure I think. Whenever you have time. Andi > -----Original Message----- > From: Sara Golemon [mailto:pollita@php.net] > Sent: Monday, December 18, 2006 9:45 PM > To: andi@zend.com > Cc: zend-engine-cvs@lists.php.net > Subject: Re: [ZEND-ENGINE-CVS] cvs: ZendEngine2(PHP_5_2) / > zend_compile.c > > "(2+2) instanceof foo" winds up getting handled in the > runtime so it "fails right" (Is that a proper expression?)... > > I agree with the consistency argument though. Unless you or > Stas or someone does it first, I'll move the fix over to the > parser tomorrow. > > -Sara > > > It's probably more consistent. Very few places check outside the > > parser and it's usually in cases which are just hard to > implement in the rules. > > I haven't tested but I could guess this test doesn't do well with > > something like (2+2) instanceof foo (not IS_CONST). > > > > Andi > > > >> -----Original Message----- > >> From: Sara Golemon [mailto:pollita@php.net] > >> Sent: Monday, December 18, 2006 9:35 PM > >> To: andi@zend.com > >> Cc: zend-engine-cvs@lists.php.net > >> Subject: Re: [ZEND-ENGINE-CVS] cvs: ZendEngine2(PHP_5_2) / > >> zend_compile.c > >> > >> The primary reason I went this route rather than handling > it in the > >> parser (I did consider it), was so that a clearer error > message could > >> be given than "Unexpected T_STRING". If you'd rather, I'd > be happy > >> to move this fix over there... > >> > >> -Sara > >> > >>> Not sure this is the right fix. It probably makes more > sense to fix > >>> the parser itself and use r_variable instead of expr for > >> the instanceof rule. > >>> Did you try and have problems doing so? I just took a quick > >> look so it > >>> might be a bit more involved than that. > >>> >