Newsgroups: php.internals
Path: news.php.net
Xref: news.php.net php.internals:111036
MIME-Version: 1.0
References: <CAF+90c821boRO6YsO0SH0JbDb8E8eTHfEK=e4Y5_z9P7+OOnUg@mail.gmail.com>
 <CAJHtzny=SX_ojsohApAY8Xe7c_EMAKGbNe110QQVh7i5yGFm_A@mail.gmail.com>
 <CAF+90c-6iEBPjWuzA6=bAr=85N=JNV0+GqL3_maak3yqVP-e=Q@mail.gmail.com>
 <CAF+90c9DxVz2Ro3yJ0-jDuPXg5Ffe3F9ERE8wHt=7ERjOm01Sw@mail.gmail.com>
 <CAF+90c_vyLakwP=1QrWUJh2rwJtcsMtPSYzyqxqOeXecGYcQTg@mail.gmail.com>
 <529e7a72-8bd7-b4f4-a987-0e88d37e47b5@telia.com> <CAF+90c_sxBHXv6Sk1eaY-2bJG24aefDE2oLEeDS8fuuFH8PoZA@mail.gmail.com>
 <44637D3E-5F0B-4AEF-A154-21C8FF20ED91@stitcher.io> <CAF+90c_Hv+xR3TTsHJU+KPosBQ3SjuTSaRf6vdaYYvTN7R9RZQ@mail.gmail.com>
 <35A1BC6F-D147-4692-8A12-A13B5ABF78D5@stitcher.io> <CAF+90c-=rM3S-zBcjdoY9GzaSp_m_UA-kkHrRQ-cgUUOwX_Kpw@mail.gmail.com>
 <CANZOY=DT7o7Pgfms3arZ2Z72Co0Y2gXqYJhHa9+PQP8_WMKuZA@mail.gmail.com>
In-Reply-To: <CANZOY=DT7o7Pgfms3arZ2Z72Co0Y2gXqYJhHa9+PQP8_WMKuZA@mail.gmail.com>
Date: Thu, 16 Jul 2020 09:51:23 +0200
Message-ID: <CAF+90c_q8jCKFemxPthCs-iPnHDuEvgCKa03HaSPOe62e+vwgg@mail.gmail.com>
To: Arnold Daniels <arnold.adaniels.nl@gmail.com>
Cc: PHP Internals <internals@lists.php.net>
Content-Type: multipart/alternative; boundary="0000000000007fa4fb05aa8a4f32"
Subject: Re: [PHP-DEV] [RFC] Saner string to number comparisons
From: nikita.ppv@gmail.com (Nikita Popov)

--0000000000007fa4fb05aa8a4f32
Content-Type: text/plain; charset="UTF-8"

On Wed, Jul 15, 2020 at 5:12 PM Arnold Daniels <arnold.adaniels.nl@gmail.com>
wrote:

> Hi Nikita,
>
> One of the problems with numeric string comparisons is that it might
> interpret a hexadecimal value as scientific notation
>
>     $red = '990000';
>     $purple = '9900e2';
>     $red == $purple; // true
>
> I suggest only interpreting a number formatted with a sign ("1e+100") or
> with decimals ("1.0e100") as scientific notation as part of this RFC. These
> are the notations that languages use when casting a float to a string,
> never "1e100"
>
> PHP "1.0E+100"
> JavaScript "1e+100"
> Python "1e+100"
> Ruby "1.0e+100"
> Java "1.0E100"
>
> Arnold
>

Hey Arnold,

I don't want to include this change in this RFC (because the proposal does
not otherwise touch string-to-string comparison at all), but I do think
this is a very good idea. I was not aware of this particular variant of the
issue (i.e., that it also occurs with prefixes other than "0e") and this
seems like a pragmatic way to resolve that particular ambiguity. A question
here would be whether to make this something specific to string-to-string
comparisons, or make a general change to numeric string recognition (i.e.,
no longer treat "1e100" as a numeric string, only "1.0e100" or "1e+100".)

Regards,
Nikita

--0000000000007fa4fb05aa8a4f32--