Newsgroups: php.internals
Path: news.php.net
Xref: news.php.net php.internals:61662
Mailing-List: contact internals-help@lists.php.net; run by ezmlm
Received-SPF: error (pb1.pair.com: domain kingsquare.nl from 212.54.42.167 cause and error)
Message-ID: <500E61F1.9040904@kingsquare.nl>
Date: Tue, 24 Jul 2012 10:50:57 +0200
Organization: Kingsquare Information Services
User-Agent: Mozilla/5.0 (Windows NT 6.1; rv:14.0) Gecko/20120713 Thunderbird/14.0
MIME-Version: 1.0
To: internals@lists.php.net
References: <500E566C.7040206@kingsquare.nl> <CA+Ky_uOTZN8HK-Q=EoEbew+g3ua-htxnSJgZow+O28S9O4+QZw@mail.gmail.com>
In-Reply-To: <CA+Ky_uOTZN8HK-Q=EoEbew+g3ua-htxnSJgZow+O28S9O4+QZw@mail.gmail.com>
Content-Type: text/plain; charset=UTF-8; format=flowed
Content-Transfer-Encoding: 7bit
Subject: Re: [PHP-DEV] Working with floats
From: robin@kingsquare.nl ("Kingsquare.nl - Robin Speekenbrink")

Thanks for the reply.

I know of the fact that converting the key to a string would be the 
'correct' way of using it, i was more curious as to the difference in 
the var_dump-s... But the precision setting is new to me, i'll have a 
look into that... That would indeed make the differences in the 
resulting $i visible and thus explainable to others.

Regards,

Robin Speekenbrink

Op dinsdag 24 juli 2012 10:42:45, Galen Wright-Watson schreef:
>
>
> On Tue, Jul 24, 2012 at 1:01 AM, Kingsquare.nl - Robin Speekenbrink
> <robin@kingsquare.nl <mailto:robin@kingsquare.nl>> wrote:
>
>     Hi all,
>     [...]
>
>     Why does the last of the following examples lower my key to 18 BUT
>     does the var dump of the float clearly state 'float(19)' as the
>     actual value:
>
>         <?php
>         [...]
>         $i =(float) 1.19;
>         $i -= 1;
>         $i *= 100;
>         var_dump($i);
>         $arr = array($i=>1);
>         var_dump($arr);
>         ?>
>
>
> When displaying floating point numbers, PHP uses the precision config
> option (http://php.net/precision) so as to shorten output (and to hide
> rounding errors in some circumstances). If you increase precision to
> 17 digits or so (64-bit IEEE floating point numbers have a decimal
> precision of around 15 to 17 digits), the var_dump will reveal that $i
> is actually 18.999999999999993.... When you use it as an array index,
> it ends up converted to an integer, which is done by truncation. If
> instead you were to convert $i to a string when using it as an index:
>
>     array("$i" => 1);
>     array((string)$i => 1);
>
> you'd get a result more along the lines that you expect.
>
>     I do know this is not really an internals thing, but after
>     fiddling with this for some time, i gave up (bug 32671 might
>     relate to this)
>
>
> It's central to what you see, but note that it was decided that the
> behavior was correct; the bug was in the documentation.
>
>     Again, if i'm to be regarded as a traditional n00b, i understand
>     as i've seen float / casting discussions before.
>
>
> I think those were more of the "how should PHP handle this" rather
> than the "why does it do this" variety. There is probably a more
> suitable venue for your question than the internals list; perhaps the
> general usage list. I can be a bit grumpy about these things, but from
> what I've seen, the PHP community likes to be inclusive, so I doubt
> it's a big deal.