0000560: === for arrays is implemented like the java == but it should be implemented differently!

Notes
(0000573) bago 12-24-05 10:38	I think this test is also interesting: <?php $foo = array(2=>'a',1=>'b'); $bar = $foo; ksort($foo); if ($bar===$foo) print("true"); else print("false"); flush(); $foo = array('a',1=>'b'); $bar = $foo; ksort($foo); if ($bar===$foo) print("true"); else print("false"); flush(); return; ?> this script in the original php return "falsetrue". after fixing the above problem you should run this test also!

(0000574) bago 12-24-05 11:24	I changed Value.java:582 from: return this == rValue.toValue(); to: return this.getType()==rValue.toValue().getType() && this.eq(rValue.toValue()); This seems to work better but the previous "ksort" example does not mimic the original php behaviour either way.

(0000575) bago 12-24-05 11:45	Using the patch described at 0000574: <?php $a = array ('bar' => 1, 'foo' => 2,); $b = array ('foo' => 2, 'bar' => 1,); print("B ==: ".($b == $a)." "); print("B ===: ".($b === $a)." "); ksort($b); print("A ==: ".($b == $a)." "); print("A ===: ".($b === $a)." "); ?> Original php result: B ==: 1 B ===: A ==: 1 A ===: 1 Patched quercus result: B ==: 1 B ===: 1 A ==: 1 A ===: 1 It seems that the php behaviour does not match its own documentation. The only difference between == and === should be in the type checking, while this prove that php check the sorting, too! I'm not sure wether quercus should try to mimic this behaviour or follow the documentation.

(0000587) ferg 12-27-05 16:32	php/0360

Relationships