Newsgroups: php.internals
Path: news.php.net
Xref: news.php.net php.internals:107361
MIME-Version: 1.0
Date: Wed, 2 Oct 2019 15:14:21 +0200
Message-ID: <CAF+90c_1CU3VVExSpb0vphLxr7bfbRxWMhvyy2D4taUbYG1xtA@mail.gmail.com>
To: PHP internals <internals@lists.php.net>
Content-Type: multipart/alternative; boundary="0000000000003c5c640593ed40fd"
Subject: fdiv() function
From: nikita.ppv@gmail.com (Nikita Popov)

--0000000000003c5c640593ed40fd
Content-Type: text/plain; charset="UTF-8"

Hi internals,

I would like to add an fdiv() function in PHP 8, implemented in
https://github.com/php/php-src/pull/4769.

Per https://wiki.php.net/rfc/engine_warnings the main division operation
will switch to throwing DivisionByZeroError on division by zero. However, I
think it is important to expose a division with IEEE-754 semantics for
division by zero for the use-cases that do need this. It is fairly tricky
and not very efficient to implement this by hand in userland code,
primarily due to correct handling for negative zero.

The fdiv() function will perform a floating-point division while treating
division by zero as a perfectly legal operation, and will not emit any kind
of diagnostic in that case. Instead it will return the INF/-INF/NAN result
mandated by IEEE-754. It mirrors the existing fmod() function.

Does anyone see an issue with this addition?

Regards,
Nikita

--0000000000003c5c640593ed40fd--