Newsgroups: php.internals
Path: news.php.net
Xref: news.php.net php.internals:93728
Mailing-List: contact internals-help@lists.php.net; run by ezmlm
Received-SPF: pass (pb1.pair.com: domain hotmail.com designates 65.55.111.166 as permitted sender)
Message-ID: <BLU436-SMTP581BDD79356EC215F3FAF3E2580@phx.gbl>
Content-Type: text/plain; charset="us-ascii"
MIME-Version: 1.0 (Mac OS X Mail 9.2 \(3112\))
In-Reply-To: <64cafa7a-2426-6c54-8e48-f060ab1edaf9@fleshgrinder.com>
Date: Thu, 2 Jun 2016 20:45:35 +0200
CC: Andrea Faulds <ajf@ajf.me>
Content-Transfer-Encoding: quoted-printable
References: <BLU437-SMTP785CB095BE08FBC1CC2FAFE2580@phx.gbl> <0A.C5.62101.1C860575@pb1.pair.com> <BLU436-SMTP143ECD8F14887F564D20E0E2580@phx.gbl> <64cafa7a-2426-6c54-8e48-f060ab1edaf9@fleshgrinder.com>
To: PHP internals <internals@lists.php.net>
Subject: Re: [PHP-DEV] Re: [RFC] [PRE-VOTE] Union types
From: bobwei9@hotmail.com (Bob Weinand)


> Am 02.06.2016 um 19:52 schrieb Fleshgrinder <php@fleshgrinder.com>:
>=20
> On 6/2/2016 7:37 PM, Bob Weinand wrote:
>> Hey,
>>=20
>> I think this is more of a presentation problem.
>> As you say, there's not much a better way to do that.
>>=20
>> It's basically our weak casting rules, just applied to the most =
lossless type available.
>> It's out weak casting rules which are so complex; the RFCs =
combination is not particularly complex.
>> It needs to cover everything, but it's particularly important for the =
trivial cases to be handled here. Would be quite WTF if you could not =
pass a float to something accepting int (as one of its types).
>>=20
>> I'd welcome every suggestion to make the rules look less scary, =
because they actually aren't.
>>=20
>> Bob
>>=20
>=20
> Don't you think this matrix would be a better presentation and =
simpler?
>=20
> |--------|--------|--------|--------|
> | type   |   #1   |   #2   |   #3   |
> |--------|--------|--------|--------|
> | bool   | string | int    |        |
> | int    | string | float  | bool   |
> | float  | string | int    | bool   |
> | object | string |        |        |
> | string | float  | int    | bool   |
> |--------|--------|--------|--------|
>=20
> Everything else is a TypeError.

Done that, does it look better now? :-)

> Note that I try string first all the time. Simply because it makes no
> difference for PHP afterwards, e.g.:
>=20
>  function f(string|int $a) {
>    if ($a =3D=3D 1) echo 'foo';
>  }
>=20
>  f(true); // true ~> '1' ~> 1 =3D foo
>=20
> Even if the change I applied is not intelligent and is not a good idea
> (haven't thought about it deeply), the presentation as a table should
> help a lot.

There is one reason, it must not be inconsistent with strict types. =
Switching from strict_types=3D1 to strict_types=3D0 shall always work =
without semantics changes (as long as you don't catch TypeErrors).

> --=20
> Richard "Fleshgrinder" Fussenegger


Thanks,
Bob=