Newsgroups: php.internals
Path: news.php.net
Xref: news.php.net php.internals:75099
Mailing-List: contact internals-help@lists.php.net; run by ezmlm
Received-SPF: pass (pb1.pair.com: domain ajf.me designates 192.64.116.199 as permitted sender)
Content-Type: text/plain; charset=windows-1252
Mime-Version: 1.0 (Mac OS X Mail 7.3 \(1878.2\))
In-Reply-To: <53AC597E.9090904@gmail.com>
Date: Thu, 26 Jun 2014 18:40:58 +0100
Cc: PHP Internals <internals@lists.php.net>,
 Chris Wright <cw@daverandom.com>
Content-Transfer-Encoding: quoted-printable
Message-ID: <E5F0E485-B14A-4409-B3F7-449E61167B9E@ajf.me>
References: <EFEB30F4-3F52-44B3-98CE-9DD29ED08E74@ajf.me>	<1CA68980-7451-4E28-AD4F-ED120393E131@ajf.me>	<FFFE60C4-41F1-47A3-80E8-CE5BC5A1BF79@gmail.com>	<7807D11E-BE36-43AF-9E6E-854BE2B0C08D@ajf.me>	<53AC0E2B.1070002@gmail.com> <CAGAGxbZhiJbbTmoT0cqK=rgXOhvcCpSKDickDTsXfvSDSGbJGw@mail.gmail.com> <53AC597E.9090904@gmail.com>
To: Rowan Collins <rowan.collins@gmail.com>
Subject: Re: [PHP-DEV] [DRAFT][RFC] Big Integer Support
From: ajf@ajf.me (Andrea Faulds)


On 26 Jun 2014, at 18:33, Rowan Collins <rowan.collins@gmail.com> wrote:

> Chris Wright wrote (on 26/06/2014):
>>> I'm assuming if bigint keys were available, PHP_INT_MAX would still =
have the
>>> same value, but PHP_INT_MAX + 1 would become a valid key, making =
this work
>>> transparently.
>> PHP_INT_MAX would probably need to be renamed (i.e. deprecated and
>> replaced) because it would become a misnomer - the max*PHP*  int is =
no
>> longer known because it is limited only by memory.
>=20
> An interesting thought. Is there really much value in renaming it =
though? It would cause a compatibility break (once the deprecation =
period expired) for something which could easily be covered by =
documentation. If done right, the BigInt support would simply make the =
constant unnecessary in the majority of cases.

Right. Myself, I=92m in favour of keeping PHP_INT_MAX around. The name =
might be misleading, but it=92s still important in a few places. So long =
as the documentation makes this clear, it=92s not a problem.

Actually, I=92d like to add a new constant perhaps, PHP_INT_MIN. It =
seems silly that we have a _MAX yet no _MIN despite having signed =
integers, especially since your minimum can=92t be typed as a literal =
(longstanding bug, one that bigints happen to =93fix=94 as it would be a =
bigint, but it should really be a long), and it=92s -(PHP_INT_MAX - 1) =
not -PHP_INT_MAX.

> It's also not entirely clear to me what it could be renamed *to* which =
was any more appropriate.

PHP_LONG_MIN perhaps, but so far as user land cares, long and int are =
synonyms. I don=92t want to rename it, anyway.

--
Andrea Faulds
http://ajf.me/