Newsgroups: php.internals
Path: news.php.net
Xref: news.php.net php.internals:103702
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8;
 format=flowed
Content-Transfer-Encoding: 8bit
Date: Tue, 08 Jan 2019 20:26:59 +0200
To: "Legale.legale" <legale.legale@gmail.com>
Cc: internals@lists.php.net
In-Reply-To: <5c33a5db.1c69fb81.eed82.f2ab@mx.google.com>
References: <5c33a5db.1c69fb81.eed82.f2ab@mx.google.com>
Message-ID: <3fc7802760ad01d9f77ea6aa2f06b789@gmail.com>
User-Agent: Roundcube Webmail/1.3.8
Subject: Re: [PHP-DEV] Proposal: change precision for output functions
From: lauri.kentta@gmail.com (=?UTF-8?Q?Lauri_Kentt=C3=A4?=)

On 2019-01-07 21:17, Legale.legale wrote:
> I think your solution by changing precision is not good enough because
> float summation is still not working properly.
> 
> <?php
> var_dump(0.1 + 0.7);
> 
> returns:
> 
> 0.7999999999
> 
> expected: 0.8

As you may know, most computers use binary numbers. This applies to 
floating-point values as well. They have a limited number of binary 
digits. Many base-10 numbers cannot be expressed exactly in base-2 with 
limited number of digits. When you write 0.1 + 0.7, what you get (with 
double-precision floats) is closer to 0.1000000000000000055511 + 
0.6999999999999999555911 = 0.7999999999999999333866. This is expected 
and proper and well known.

You have to use formatting functions like number_format if you need 
neatly rounded base-10 output.

There are also a lot of libraries for precise base-10 calculations. They 
will be a lot slower than native binary floating-point calculations, 
though.

Regards,
Lauri Kenttä