How does PHP SimpleXML get evaluated as an expression? - php

Question

Consider this PHP code:
<?php
$xmlString =
'<'.'?xml version="1.0" encoding="UTF-8"?'.'>' .
'<rootEl>' .
'<a attrA="valA">xxx</a>' .
'<b attrB="valB"/>' .
'<c>oink</c>' .
'<d/>'.
'<e>' .
'<f>zzz</f>' .
'</e>' .
'</rootEl>';
$xml = simplexml_load_string( $xmlString );
foreach ( $xml->children() as $child ) {
echo "\$CHILD [" . $child->getName() . "]: " . $child->asXML();
echo "\n";
$insideIf1 = false;
if ($child) { $insideIf1 = true; }
$insideIf2 = false;
if (true == $child) { $insideIf2 = true; }
$insideIf3 = false;
if ((boolean)$child) { $insideIf3 = true; }
echo " - if(\$CHILD) return \"true\"; else return \"false\"; : " . (($insideIf1)?"true":"false")."\n";
echo " - if(true == \$CHILD) return \"true\"; else return \"false\"; : " . (($insideIf2)?"true":"false")."\n";
echo " - if((boolean)\$CHILD) return \"true\"; else return \"false\"; : " . (($insideIf3)?"true":"false")."\n";
echo " - ((\$CHILD)?\"true\":\"false\"): " . (($child)?"true":"false")."\n";
echo " - ((true == \$CHILD)?\"true\":\"false\"): " . ((true == $child)?"true":"false")."\n";
echo " - (((boolean)\$CHILD)?\"true\":\"false\"): " . (((boolean)$child)?"true":"false")."\n";
echo "\n";
}
As per PHP documentation (online, see http://php.net/manual/en/control-structures.if.php):
As described in the section about expressions, expression is evaluated
to its Boolean value. If expression evaluates to TRUE, PHP will
execute statement, and if it evaluates to FALSE - it'll ignore it.
More information about what values evaluate to FALSE can be found in
the 'Converting to boolean' section.
The "converting to boolean" section of this explanation states:
Converting to boolean
When converting to boolean, the following values are considered FALSE:
...omissis...
SimpleXML objects created from empty tags
If you execute the code above, these are the results on my PHP 5.5.9-1ubuntu4.5 (cli):
$CHILD [a]: <a attrA="valA">xxx</a>
- if($CHILD) return "true"; else return "false"; : true
- if(true == $CHILD) return "true"; else return "false"; : true
- if((boolean)$CHILD) return "true"; else return "false"; : true
- (($CHILD)?"true":"false"): true
- ((true == $CHILD)?"true":"false"): true
- (((boolean)$CHILD)?"true":"false"): true
$CHILD [b]: <b attrB="valB"/>
- if($CHILD) return "true"; else return "false"; : true
- if(true == $CHILD) return "true"; else return "false"; : false
- if((boolean)$CHILD) return "true"; else return "false"; : true
- (($CHILD)?"true":"false"): true
- ((true == $CHILD)?"true":"false"): false
- (((boolean)$CHILD)?"true":"false"): true
$CHILD [c]: <c>oink</c>
- if($CHILD) return "true"; else return "false"; : false
- if(true == $CHILD) return "true"; else return "false"; : true
- if((boolean)$CHILD) return "true"; else return "false"; : false
- (($CHILD)?"true":"false"): false
- ((true == $CHILD)?"true":"false"): true
- (((boolean)$CHILD)?"true":"false"): false
$CHILD [d]: <d/>
- if($CHILD) return "true"; else return "false"; : false
- if(true == $CHILD) return "true"; else return "false"; : false
- if((boolean)$CHILD) return "true"; else return "false"; : false
- (($CHILD)?"true":"false"): false
- ((true == $CHILD)?"true":"false"): false
- (((boolean)$CHILD)?"true":"false"): false
$CHILD [e]: <e><f>zzz</f></e>
- if($CHILD) return "true"; else return "false"; : true
- if(true == $CHILD) return "true"; else return "false"; : false
- if((boolean)$CHILD) return "true"; else return "false"; : true
- (($CHILD)?"true":"false"): true
- ((true == $CHILD)?"true":"false"): false
- (((boolean)$CHILD)?"true":"false"): true
Children elements "b" and "d" ARE empty, therefore I was expecting "false", and "a", "c", "e" ARE NOT empty, so I was expecting "true".
As is comes out, that statement in PHP documentation is absolutely NOT true, and the behaviour doesn't even seem consistent: implicit or explicit casts to boolean don't behave the same way (see case of element "c"), and if the bare SimpleXML object is used as "if" conditional expression (both "if" and ternary operator), somehow that expression evaluates differently from a cast to boolean.
Does anyone have an idea of what's going on here?

Answer 1

First of all, let's discard some of your code.
foreach ( $xml->children() as $child ) {
echo "\$CHILD [" . htmlspecialchars($child->getName()) . "]: " . htmlspecialchars($child->asXML());
echo "\n";
$insideIf1 = false;
if ($child) { $insideIf1 = true; }
$insideIf2 = false;
if (true == $child) { $insideIf2 = true; }
$insideIf3 = false;
if ((boolean)$child) { $insideIf3 = true; }
var_dump($child);
// echo intval($child)."\n";
echo 'VALUE EVALUATES TO: '. $child."\n";
echo " - if(\$CHILD) return \"true\"; else return \"false\"; : " . (($insideIf1)?"true":"false")."\n";
echo " - if(true == \$CHILD) return \"true\"; else return \"false\"; : " . (($insideIf2)?"true":"false")."\n";
echo " - if((boolean)\$CHILD) return \"true\"; else return \"false\"; : " . (($insideIf3)?"true":"false")."\n";
// This part evaluated exactly same as above. So no need to make confusion
echo "\n";
}
Next:
First and third test
Those are basically same. Object is considered true, if it has properties. So basically xml tag without attributes has no properties (just var_dump them).
Objects without properties (does not have attributes BUT NOTE THEY CAN have value - which is returned by iterator) considered false: c, d
Second test
This test on the other hand does not test object itself. But all values returned by it from casting to: int, bool and string.
One of those tests calls for __toString magic method. Which returns value. Then it evaluates to true (Elements a and c)
So its not random, but has special rules while considering it true or false.
Forgot to add
Note: SimpleXML has made a rule of adding iterative properties to most methods. They cannot be viewed using var_dump() or anything else which can examine objects.
This "strange" behaviour is result of it implementing inerators.
Edit:
As you stated in comment what made you analyze this behaviour. I think that you should never chose to check SimpleXML objects on if($obj) basis.
Instead, use SimpleXML::attributes() or strval to do all checks. Then you're sure what you're checking.
Edit2:
$xmlObject is considered true (is not empty) if it has Properties OR Chilren (subnodes) (that works fine)
To check if object has value you cast it to string.
In this case, empty does not mean if it has value, only if it has children or properties.