Newsgroups: php.internals
Path: news.php.net
Xref: news.php.net php.internals:126468
Feedback-ID: ifab94697:Fastmail
Precedence: bulk
MIME-Version: 1.0
Date: Thu, 20 Feb 2025 20:53:04 +0100
To: "Volker Dusch" <volker@tideways-gmbh.com>
Cc: "php internals" <internals@lists.php.net>
Message-ID: <f2af2267-03cc-47fa-bc89-653160b5eda1@app.fastmail.com>
In-Reply-To: 
 <CAMzRQTda1_MMzC-SWRoxRmvssSUq3Ty5s2a8T4K8j8Bz90kY-g@mail.gmail.com>
References: <a995ac24905030e168e7f6006b996cd5@bastelstu.be>
 <CAGBsUrfFvWBcncpNorb3dmXJ_=cZLhmdaVkFHR6L5SkrhtPLcw@mail.gmail.com>
 <95a69ca549b246c3645f8c6db519f669@bastelstu.be>
 <46f5a4ab-ae85-4a15-9781-27a94eedc512@app.fastmail.com>
 <CAMzRQTda1_MMzC-SWRoxRmvssSUq3Ty5s2a8T4K8j8Bz90kY-g@mail.gmail.com>
Subject: Re: [PHP-DEV] RFC: Marking return values as important (#[\NoDiscard])
Content-Type: multipart/alternative;
 boundary=97334fef8ef24f66b75ea92a89a3e733
From: rob@bottled.codes ("Rob Landers")

--97334fef8ef24f66b75ea92a89a3e733
Content-Type: text/plain; charset=utf-8
Content-Transfer-Encoding: quoted-printable



On Thu, Feb 20, 2025, at 16:54, Volker Dusch wrote:
> > On Wed, Feb 12, 2025 at 9:57=E2=80=AFPM Rob Landers <rob@bottled.cod=
es> wrote:
> Hey Rob,
>=20
> Sorry for writing the mail again, I just noticed I forgot to include t=
he list on my first reply to you, also corrected a mistake in the second=
 paragraph.
>=20
> > There will eventually be a php 9, where BC changes will be possible.
>=20
> I don't assume PHP will change the behavior of widely used core functi=
ons to throw exceptions in PHP 9 as the BC impact will be colossal, hurt=
ing adoption. Or am I misunderstanding you here?
>=20
> The point that there might be arguments over usage in internal code is=
 why we went with an "only if it is a clear problem, especially when the=
 function has important side effects" policy to avoid this. There are no=
t a too many examples in core, more in userland.=20

I was merely pointing out that if you were to want to make BC breaks, yo=
u can. You just won=E2=80=99t get it :soon:

>=20
> > I have to stop you here. It is absolutely NOT safe and reasonable to=
 ignore the output of array_pop. You popped it for a reason. If you just=
 care about removing the last element, then you should use unset. Unset =
gives the intention. If I review code with array_pop, I=E2=80=99ll spend=
 quite a bit of time checking that it was intentionally ignoring the ret=
urn value (and leave a comment of the same, asking to use unset instead).
>=20
> There are plenty of codebases where array_pop is used for this reason.=
 Identifying the last element of a php array. `unset($foo[array_key_last=
($foo)]);` is only possible since 7.3 and not widely used. array_pop is =
also shorter and faster for the same effect (when used on unknown array =
shapes ofc). The examples are plenty https://github.com/search?q=3Drepo%=
3AWordPress%2FWordPress%20array_pop&type=3Dcode
>=20
> Regards,
> Volker

This feels like a rationalization, and not a reason. It has a return val=
ue, and there are more expressive ways to remove a value from an array t=
hat makes the intention clearer. If you aren=E2=80=99t using the return =
value, you should make it clear that is the intention. That=E2=80=99s wh=
y I said it should be included in this, for exactly the rationale you ga=
ve. It IS faster, but it clearly has a side effect that should be used o=
r intentionally discarded.=20

=E2=80=94 Rob
--97334fef8ef24f66b75ea92a89a3e733
Content-Type: text/html; charset=utf-8
Content-Transfer-Encoding: quoted-printable

<!DOCTYPE html><html><head><title></title><style type=3D"text/css">p.Mso=
Normal,p.MsoNoSpacing{margin:0}</style></head><body><div><br></div><div>=
<br></div><div>On Thu, Feb 20, 2025, at 16:54, Volker Dusch wrote:<br></=
div><blockquote type=3D"cite" id=3D"qt" style=3D""><div dir=3D"ltr"><div=
 dir=3D"ltr"><div dir=3D"ltr">&gt; On Wed, Feb 12, 2025 at 9:57=E2=80=AF=
PM&nbsp;<span class=3D"qt-gmail-il">Rob</span>&nbsp;Landers &lt;<span cl=
ass=3D"qt-gmail-il">rob</span>@bottled.codes&gt; wrote:<br></div><div di=
r=3D"ltr"><div>Hey&nbsp;<span class=3D"qt-gmail-il">Rob</span>,<br></div=
><span class=3D"qt-gmail-im" style=3D"color:rgb(80, 0, 80);"><div><br></=
div><div>Sorry for writing the mail again, I just noticed I forgot to in=
clude the list on my first reply to you, also corrected a mistake&nbsp;i=
n the second paragraph.<br></div><div><br></div><div>&gt; There will eve=
ntually be a php 9, where BC changes will be possible.<br></div><div><br=
></div></span><div>I don't assume PHP will change the behavior of widely=
 used core functions to throw exceptions in PHP 9 as the BC impact will =
be colossal, hurting adoption. Or am I misunderstanding you here?<br></d=
iv><div><br></div><div>The point that there might be arguments over usag=
e in internal&nbsp;code is why we went with an "only if it is a clear pr=
oblem, especially when the function has important side effects" policy t=
o avoid this. There are not a too many examples in core, more in userlan=
d.&nbsp;<br></div></div></div></div></blockquote><div><br></div><div>I w=
as merely pointing out that if you were to want to make BC breaks, you c=
an. You just won=E2=80=99t get it :soon:</div><div><br></div><blockquote=
 type=3D"cite" id=3D"qt" style=3D""><div dir=3D"ltr"><div dir=3D"ltr"><d=
iv dir=3D"ltr"><span class=3D"qt-gmail-im" style=3D"color:rgb(80, 0, 80)=
;"><div><br></div><div>&gt; I have to stop you here. It is absolutely NO=
T safe and reasonable to ignore the output of array_pop. You popped it f=
or a reason. If you just care about removing the last element, then you =
should use unset. Unset gives the intention. If I review code with array=
_pop, I=E2=80=99ll spend quite a bit of time checking that it was intent=
ionally ignoring the return value (and leave a comment of the same, aski=
ng to use unset instead).<br></div><div><br></div></span><div>There are =
plenty of codebases where array_pop is used for this reason. Identifying=
 the last element of a php array. `unset($foo[array_key_last($foo)]);` i=
s only possible since 7.3 and not widely used. array_pop is also shorter=
 and faster for the same effect (when used on unknown array shapes ofc).=
 The examples are plenty&nbsp;<a href=3D"https://github.com/search?q=3Dr=
epo%3AWordPress%2FWordPress%20array_pop&amp;type=3Dcode" target=3D"_blan=
k">https://github.com/search?q=3Drepo%3AWordPress%2FWordPress%20array_po=
p&amp;type=3Dcode</a><br></div><div><br></div><div>Regards,<br></div><di=
v>Volker<br></div></div></div></div></blockquote><div><br></div><div>Thi=
s feels like a rationalization, and not a reason. It has a return value,=
 and there are more expressive ways to remove a value from an array that=
 makes the intention clearer. If you aren=E2=80=99t using the return val=
ue, you should make it clear that is the intention. That=E2=80=99s why I=
 said it should be included in this, for exactly the rationale you gave.=
 It IS faster, but it clearly has a side effect that should be used or i=
ntentionally discarded.&nbsp;<br></div><div><br></div><div id=3D"sig1212=
29152">=E2=80=94 Rob<br></div></body></html>
--97334fef8ef24f66b75ea92a89a3e733--