Re: 5.4 regression: non-existent sub-sub keys now have values

On 23 November 2011 15:31, Daniel Convissor <danielc@analysisandsolutions.com> wrote: > Hi Again Folks: > > On Wed, Nov 23, 2011 at 09:14:09AM -0500, Daniel Convissor wrote: >> $a = 'foo'; >> echo $a['blah'] . "\n"; >> >> But that second one echos out "f".  This is a huge WTF. > > Two things for the record on this front.  First, i've been actively > using PHP for, what, ten years or so, and have never run into this > behavior before.  Second, this behavior turns the following one liner: > >        if (isset($arr['package']['attribs']['version'])) { > > Into this: > >        if (is_array($arr) >                && array_key_exists('package', $arr) >                && is_array($arr['package']) >                && array_key_exists('attribs', > $arr['package']) >                && is_array($arr['package']['attribs']) >                && array_key_exists('version', > $arr['package']['attribs']) >                && > !empty($arr['package']['attribs']['version'])) >        { I agree with Daniel on this. Just looking for any test relating to isset() to see what tests will now fail. Neither of the isset() tests ext/standard/tests/general_functions look at arrays with associative indices. The behaviour with regard to unavailable associative indices has no tests. But using loose typing as a fallback for non existent keys seems really really wrong. Especially if the key is a constant. On what planet should ... isset($arr['exists']['test_existance']) should become isset($arr['exists'][0]) That just seems really wrong and the work around is awful. -- Richard Quadling Twitter : EE : Zend : PHPDoc : Fantasy Shopper @RQuadling : e-e.com/M_248814.html : bit.ly/9O8vFY : bit.ly/lFnVea : fan.sh/6/370