Newsgroups: php.internals
Path: news.php.net
Xref: news.php.net php.internals:82497
Mailing-List: contact internals-help@lists.php.net; run by ezmlm
Received-SPF: pass (pb1.pair.com: domain ajf.me designates 192.64.116.200 as permitted sender)
Content-Type: text/plain; charset=utf-8
Mime-Version: 1.0 (Mac OS X Mail 8.2 \(2070.6\))
In-Reply-To: <203e611c8e0b03568a868b8d931aec37@mail.gmail.com>
Date: Wed, 11 Feb 2015 21:46:30 +0000
Cc: Rasmus Lerdorf <rasmus@lerdorf.com>,
 PHP Internals <internals@lists.php.net>
Content-Transfer-Encoding: quoted-printable
Message-ID: <F79D6CDF-20DD-4E24-B2F4-F12637368242@ajf.me>
References: <8703B53E-2C4A-4AC6-95C4-D4F19C6D5221@ajf.me> <CABBJUpf=wKvGY9+xORRv6WVhCb5+LfRZXGYNasHFS=fz_mwaZQ@mail.gmail.com> <CABBJUpeceN1uDK870eV2=znC9f3eaxqYCmaEk0ErH2cbFxTNEw@mail.gmail.com> <CABBJUpdYnaR2nztPwFUdDAiAneRrAWzGo-36-DA9hqiBjEHXNA@mail.gmail.com> <54DAF884.7000508@lerdorf.com> <203e611c8e0b03568a868b8d931aec37@mail.gmail.com>
To: Zeev Suraski <zeev@zend.com>
Subject: Re: [PHP-DEV] [VOTE] Scalar Type Hints
From: ajf@ajf.me (Andrea Faulds)


> On 11 Feb 2015, at 21:30, Zeev Suraski <zeev@zend.com> wrote:
>=20
>> -----Original Message-----
>> From: Rasmus Lerdorf [mailto:rasmus@lerdorf.com]
>> Sent: Wednesday, February 11, 2015 8:37 AM
>> To: Xinchen Hui; Andrea Faulds
>> Cc: PHP Internals
>> Subject: Re: [PHP-DEV] [VOTE] Scalar Type Hints
>>=20
>> On 02/10/2015 07:57 PM, Xinchen Hui wrote:
>>>> am I wrong?!
>>> seems I am wrong with this, it's a false alarm...  it can restore
>> automatically.
>>=20
>> Yeah, declare() doesn't span files so that isn't a problem.
>>=20
>> My worry is still the lack of type coercion for internal calls.
>=20
> There's also some pretty interesting analysis on this proposal - based =
on
> some real world experimentation - from Benjamin Eberlei - =
bit.ly/1vl0UED

I disagree with some aspects of the blogpost:

> enabling strict mode completly defeats the purpose, because now we are =
forced to convert manually, reimplementing weak type hinting in our own =
code

This isn=E2=80=99t true at all. The key difference between explicit =
conversions and implicit conversions is that they=E2=80=99re, well, =
*explicit*. You can see them, they are blatantly obvious, if you use =
grep you can find them. Implicit conversions, while they can sometimes =
be great, happen silently and automatically - if they screw up, it can =
be hard to track down where the conversion is taking place. With =
explicit conversions, you=E2=80=99re converting once and you know =
you=E2=80=99ve converted. With implicit conversions, you have no idea =
where and how often values are being converted.

> We write code with casts already, the scalar type hints patch is not =
necessary for that! Only a superficial level of additional safety is =
gained, one additional check of something we already know is true!

Yet in the previous part he was explaining how existing code doesn=E2=80=99=
t use casts and would need them added=E2=80=A6 so therefore, this thing =
*isn=E2=80=99t* already known to be true, and the blogpost is =
self-contradictory.

> strict mode is useless for library developers, because I always have =
to assume weak mode anyways.

I don=E2=80=99t really understand at all=E2=80=A6 either he doesn=E2=80=99=
t understand the proposal (which guarantees you get the types you ask =
for), or it=E2=80=99s a reference to the claim from the previous =
paragraph, which I=E2=80=99ve already responded to.

> In a well designed application or library, the developer can already =
trust the types of his variables today, 95% of the time, without even =
having type hints, using carefully designed abstractions (example =
Symfony Forms and Doctrine ORM): No substantial win for her from having =
strict type hints.

Yet he was previously explaining how, actually, data coming in from web =
requests isn=E2=80=99t the correct types.

> In a badly designed application, the developer is uncertain about the =
types of variables. Using strict mode in this scenario she needs to =
start casting everywhere just to be sure.

In an application using strict mode everywhere with scalar type hints =
everywhere, you can=E2=80=99t possibly be uncertain about the types of =
variables.

=E2=80=A6 there=E2=80=99s a lot more in that blog post I could disagree =
with, but I won=E2=80=99t bother. Benjamin has his reasons for not =
liking strict mode. That=E2=80=99s fine. But if PHP adds strict mode, it =
has zero effect on him.

> If you've voted on this RFC or intend to vote - please spend the few =
minutes
> to read through his blog post.  It does a good job at actually =
illustrating
> what life would look like if this rubber ends up meeting the road.

Anthony Ferrara also had his own analysis which showed some of the =
problems with weak type hinting, and where strict types can be =
beneficial: http://blog.ircmaxell.com/2015/02/scalar-types-and-php.html

I would suggest people read that blog post before voting, as well. It =
points out some real-world examples of where weak typing in PHP can be =
insufficient (div), or even downright dangerous (curlopt).

--
Andrea Faulds
http://ajf.me/