note 47841 added to language.operators.comparison

Concerning floats: It is simply pointless to compare a float with the value "0.3". The number 0.3 is not exactly representable in binary. Period. So is the number 0.1, and an infinite number of others numbers. Just like 1/3 is not exactly representable in decimal. How would you code the test for your float to be one third? Maybe $myFloat == 0.33333333333 Hmm: you see: Everyone would agree that this test is not accurate. The test $myFloat == 0.3 is making exactly the same mistake. So the float which you think should be 0.3 is really something very close to it; if you print it in decimal, the conversion will end up with the closest decimal representation, which may well be "0.3". But "0.3" is also the "right display decimal" for hundreds of float values. The correct way to "compare" floats is: ( $myFloat - 0.3 ) < $EPSILON where $EPSILON is something like 1e-10, depending on your application. Approximate algorithms are the only right way to treat floats, which are only approximate by themselves. ---- Manual Page -- http://www.php.net/manual/en/language.operators.comparison.php Edit -- http://master.php.net/manage/user-notes.php?action=edit+47841 Delete -- http://master.php.net/manage/user-notes.php?action=delete+47841&report=yes Reason: bad code -- http://master.php.net/manage/user-notes.php?action=delete+47841&report=yes&reason=bad+code Reason: spam -- http://master.php.net/manage/user-notes.php?action=delete+47841&report=yes&reason=spam Reason: useless example -- http://master.php.net/manage/user-notes.php?action=delete+47841&report=yes&reason=useless+example Reason: contains commercial links -- http://master.php.net/manage/user-notes.php?action=delete+47841&report=yes&reason=contains+commercial+links Reason: useless note -- http://master.php.net/manage/user-notes.php?action=delete+47841&report=yes&reason=useless+note Reject -- http://master.php.net/manage/user-notes.php?action=reject+47841&report=yes Search -- http://master.php.net/manage/user-notes.php