Newsgroups: php.internals
Path: news.php.net
Xref: news.php.net php.internals:113131
To: Guilliam Xavier <guilliam.xavier@gmail.com>, Larry Garfield
	<larry@garfieldtech.com>
CC: PHP internals <internals@lists.php.net>
Thread-Topic: [PHP-DEV] [VOTE] PHP\iterable\any() and all() on iterables
Thread-Index:
 AQHW/ibG3D0K4gIX3kCegHwWEoc5IqpOVSSAgAA2OhOAAExdgIAAGauXgAAM4ACAASllgIAA8NKAgABczC4=
Date: Wed, 10 Feb 2021 14:37:48 +0000
Message-ID:
 <DM6PR07MB6618E20AC3518DB4FD0C2FA9F98D9@DM6PR07MB6618.namprd07.prod.outlook.com>
References:
 <DM6PR07MB66189594D6D3B6D820E4A5FBF98F9@DM6PR07MB6618.namprd07.prod.outlook.com>
 <77dfe9e5-a6d1-4a41-bceb-454a65cf34d0@www.fastmail.com>
 <DM6PR07MB6618999C8F1AAC35BF2F8C87F98F9@DM6PR07MB6618.namprd07.prod.outlook.com>
 <f1abc6fd-dfad-4cd4-a124-58475312fa86@www.fastmail.com>
 <DM6PR07MB66184FDB329A71B371883C54F98E9@DM6PR07MB6618.namprd07.prod.outlook.com>
 <CAAKU0u=5ypeaT7u3xi-zT=cXLJxJYXrL_LLaV+HRXHmYSJWwOA@mail.gmail.com>
 <5ed1c099-6a0b-4e92-96a7-a1f21aefe753@www.fastmail.com>,<CAC5V6gjrQeba5uMW00GC11fW1SbHiQkYU6GsFyz4bVsZ57RVLw@mail.gmail.com>
In-Reply-To:
 <CAC5V6gjrQeba5uMW00GC11fW1SbHiQkYU6GsFyz4bVsZ57RVLw@mail.gmail.com>
Accept-Language: en-CA, en-US
Content-Language: en-CA
x-ms-exchange-messagesentrepresentingtype: 1
x-tmn: [/kOCIJgX49WCt4CfnhzyVsbycdGJC+GzFBus2+MgG2SxOLQctxue6Psig1vHgPUn]
x-ms-publictraffictype: Email
x-incomingheadercount: 45
x-eopattributedmessage: 0
x-ms-office365-filtering-correlation-id: 12548a09-6b4e-4293-db59-08d8cdd17032
x-ms-traffictypediagnostic: CO1NAM04HT191:
x-microsoft-antispam: BCL:0;
x-microsoft-antispam-message-info:
 P3bmcSh221DwT+AQH05ritJLY9EjggMKCrfqqkv5YLnoBDC4SqtnCNnowh+xTu1XdJM20yAXVkXuFMw4OIAgcw9zbQLtqzXdiKgoiQVSyzoxgftnhF6z5a8Mdrwym4CGP43Uj46B8x6TGBC4ptfcGZKKyDJf5Newegty522v9rddtVpssFRQQ9zzihpXKSGCXY5ohLhbs3EJ62328jgD7qlehsHnZAIicpSUeafjdSWTT/MmUyHec3paueotvd4JyHNPfgQXZYwHLiEOkwVdE1X3FAw/nPa5+fDOjcU+KIjlaGYpBY5LoqMjfWOFHg9e5enFih6JpoDxOMop2NdZQW5wrR7hns1fw7fKJm6bdziQeg8dD/qJXACsWBX7UugSEeSUzUslfjnI6vCa5BPlu+UaQGt8v4v5kOJdbnXBA9bwlT2bHb8EB3eBZqzuCro7
x-ms-exchange-antispam-messagedata:
 TnUyCtGcQN6mhjYdT+WbZd/oPC2a40sY1pLbPAH39XUoLBWUAsNhrpHlX2tjcJCRboNK/3kCSClO4AU8Z8IShOOruFJarnH7X6OVsS5IrlaCIRlaBeRtBEgmTcZLKWTfN0HWVzFQfYWkorHz8dA8vPHZ2vV4D448+QNIZ8OM9NXNES626bHu11vDpnmFOll05aI0HOWmD9gtsunqKi0Cbg==
x-ms-exchange-transport-forked: True
Content-Type: text/plain; charset="Windows-1252"
Content-Transfer-Encoding: quoted-printable
MIME-Version: 1.0
X-OriginatorOrg: hotmail.com
X-MS-Exchange-CrossTenant-AuthAs: Anonymous
X-MS-Exchange-CrossTenant-AuthSource: CO1NAM04FT055.eop-NAM04.prod.protection.outlook.com
X-MS-Exchange-CrossTenant-RMS-PersistedConsumerOrg: 00000000-0000-0000-0000-000000000000
X-MS-Exchange-CrossTenant-Network-Message-Id: 12548a09-6b4e-4293-db59-08d8cdd17032
X-MS-Exchange-CrossTenant-originalarrivaltime: 10 Feb 2021 14:37:48.6485
 (UTC)
X-MS-Exchange-CrossTenant-fromentityheader: Internet
X-MS-Exchange-CrossTenant-id: 84df9e7f-e9f6-40af-b435-aaaaaaaaaaaa
X-MS-Exchange-CrossTenant-rms-persistedconsumerorg: 00000000-0000-0000-0000-000000000000
X-MS-Exchange-Transport-CrossTenantHeadersStamped: CO1NAM04HT191
Subject: Re: [PHP-DEV] [VOTE] PHP\iterable\any() and all() on iterables
From: tysonandre775@hotmail.com (tyson andre)

Hi Guilliam Xavier,=0A=
=0A=
> > > Hi Larry Garfield,=0A=
> > >=0A=
> > > > > Hi Larry Garfield,=0A=
> > > > >=0A=
> > > > > > > Hi internals,=0A=
> > > > > > >=0A=
> > > > > > > Voting has started on https://wiki.php.net/rfc/any_all_on_ite=
rable and=0A=
> > > > > > > ends on 2021-02-22.=0A=
> > > > > > >=0A=
> > > > > > > This RFC proposes to add the functions `PHP\iterable\any(iter=
able=0A=
> > > > > > > $input, ?callable $callback =3D null): bool` and `PHP\iterabl=
e\all(...)`=0A=
> > > > > > > to PHP's standard library's function set, using the namespace=
 preferred=0A=
> > > > > > > in the previous straw poll.=0A=
> > > > > > >=0A=
> > > > > > > [...]=0A=
> > > > > >=0A=
> > > > > >=0A=
> > > > > > Ak!=A0 I literally just finished reading it and wanted to note =
a lack of clarity on one point. :-)=0A=
> > > > > >=0A=
> > > > > > The signature of the callback is never specified explicitly.=A0=
 The ternary is a bit confusing.=A0 I assume the signature is=0A=
> > > > > >=0A=
> > > > > > callable(mixed): bool=0A=
> > > > > >=0A=
> > > > > > But that's not made explicit.=A0 It's also not made explict tha=
t omitting the callable collapses to "is truthy".=A0 That's a sensible thin=
g to do, but it's not stated explicitly anywhere, just inferred from the co=
de sample.=0A=
> > > > > >=0A=
> > > > > >=A0 [...] =0A=
> > > > >=0A=
> > > > > If there is a callable, it allows `callable(mixed): mixed`,=0A=
> > > > > and converts the callable's return value to a boolean.=0A=
> > > > > So omitting the callable is the same as passing in the callable `=
fn($x)=0A=
> > > > > =3D> $x`, which is equivalent to `fn($x) =3D> (bool)$x`.=0A=
> > > > > This is exactly what the reference implementation would do.=0A=
> > > > >=0A=
> > > > >=A0 [...] =0A=
> > > >=0A=
> > > > Oof.=A0 I'm glad I asked, because I don't like that at all.=A0 If a=
vailable, the callable should be returning bool, not "anything that may be =
truthy/falsy."=A0 If you have an explicit function, it should have an expli=
cit return type.=A0 A truthy check is a reasonable default, but not for whe=
n you're opting in to specifying the logic.=0A=
> > > >=0A=
> > > > I am in favor of the RFC, but I will have to consider if that chang=
es my vote to No.=0A=
> > > >=0A=
> > > > --Larry Garfield=0A=
> > >=0A=
> > > This was a deliberate choice and is consistent with the weak type com=
parison behavior of array_filter() and other functions that default to usin=
g weak type checks internally.=0A=
> > >=0A=
> > > I'd agree that I'd prefer to see callbacks returning booleans in code=
 I'm reviewing,=0A=
> > > but a truthiness check seems more practical and consistent with the r=
est of the language=0A=
> > > than throwing a TypeError or checking the predicate return value usin=
g `!=3D=3D true`=0A=
> > >=0A=
> > > This was made to make PHP more widely accessible and free of surprise=
s.=0A=
> > > e.g. `(bool)array_filter($arr, $predicate)` can be safely converted t=
o `any($arr, $predicate)` without introducing a TypeError or behavior chang=
e.=0A=
> > >=0A=
> > >=A0 [...] =0A=
> > =0A=
> > For what it is worth, in C++ it is fairly normal to use a convertible=
=0A=
> > to bool type. For instance, having an overload on a < b for iterators=
=0A=
> > can return whatever type it wants, as long as it is contextually=0A=
> > convertible to bool.=0A=
> =0A=
> Yet in 8.0, a non-[ 1 | 0 | -1 ] return from a comparison function as jus=
t converted to a warning.=A0 So the trend in the language seems to be the o=
ther direction.=0A=
> =0A=
> Are you referring to this frequent mistake?=0A=
> =0A=
> ```=0A=
> $list =3D [2,4,1,3];=0A=
> usort($list, fn ($a, $b) =3D> $a < $b); /* Deprecated: usort(): Returning=
 bool from comparison function is deprecated, return an integer less than, =
equal to, or greater than zero */=0A=
> echo json_encode($list); // [4,3,2,1]=0A=
> ```=0A=
> =0A=
> (That's probably because PHP user comparison functions were modelled from=
 C (e.g. strcmp()), not C++ (e.g. std::less).)=0A=
> =0A=
> Of course here the "correct" thing is to return `$a <=3D> $b` (or `$b <=
=3D> $a` for descending order), but you can also return `$a - $b` (not nece=
ssarily in [-1,0,1]), or even a string `"foo"` still without any warning in=
 8.0.2 (just a certainly wrong result)...=0A=
=0A=
A callback function called **once** (before php 8.0) on a pair mapping 2 st=
ates (true, false) onto 3 states (-1, 0, 1) only worked coincidentally=0A=
because the sort wasn't stable. The desired type had more possible values t=
han the returned type, so it did the wrong thing.=0A=
- This is different in that there's only 2 desired values.=0A=
=0A=
https://wiki.php.net/rfc/stable_sorting fixed that calling it twice and dep=
recating booleans.=0A=
It still allows you to return strings (not likely in practice) and casts th=
ose to integers.=0A=
=0A=
"Second, if boolean false is returned, PHP will automatically call the comp=
arison function again with arguments swapped.=0A=
This allows us to distinguish whether the =93false=94 stood for =93equal=94=
 or =93smaller than=94. This fallback behavior should be removed in a futur=
e version of PHP."=0A=
=0A=
> Anyway, to me it feels natural that any()/all() would "work" like array_f=
ilter().=0A=
> =0A=
> @Tyson by the way, in the any()/all() case (vs the any_value()/all_values=
() and potential any_key()/all_keys() etc.), wouldn't it be preferable to a=
dd the optional `int $flags =3D 0` (or "$mode") parameter right from the st=
art (even if not used yet), as adding it in a later release would apparentl=
y pose some BC concerns (ArgumentCountError, polyfills etc.)?=0A=
=0A=
If an iterables RFC passed, and all amendments got approved before 8.1 stab=
le was released, it wouldn't matter. There's months before the feature free=
ze.=0A=
=0A=
If a function was added with flags "just in case" in 8.1 stable and we neve=
r used the flags, we'd have to deprecate them and remove them in a subseque=
nt major release.=0A=
I don't think that's worth it.=0A=
=0A=
Cheers,=0A=
Tyson Andre=0A=