Inline conditional that returns null if falsy

3 years ago by Ben Ramsey — view source

unread

The => is just a suggestion, other options using existing keywords is:
return expr() then output();
return expr(): output();

I do not know if other languages support something like that.

I think it might be a good idea to check other languages to see if they support something like this. We could use their examples as points of reference for discussing whether to include this functionality in PHP.

Cheers,
Ben

3 years ago by Michael Morris — view source

unread

Javascript has this now though support isn't widespread.

https://developer.mozilla.org/en-US/docs/Web/JavaScript/Reference/Operators/Optional_chaining

The most similar way to do it in PHP would be ?->

The => is just a suggestion, other options using existing keywords is:
return expr() then output();
return expr(): output();

I do not know if other languages support something like that.

I think it might be a good idea to check other languages to see if they
support something like this. We could use their examples as points of
reference for discussing whether to include this functionality in PHP.

Cheers,
Ben

3 years ago by David Rodrigues — view source

unread

Em qua., 24 de fev. de 2021 às 12:29, Michael Morris tendoaki@gmail.com
escreveu:

Javascript has this now though support isn't widespread.

https://developer.mozilla.org/en-US/docs/Web/JavaScript/Reference/Operators/Optional_chaining

The most similar way to do it in PHP would be ?->

Optional chaining is not the same.

My suggestion is similar to:

$variable = $user->exists ? $user->fullname; // suggestion (will return
null if expression is falsy)
$variable = $user->exists ? $user->fullname : null; // same as

Optional chaining is:

$variable = $user?->exists; // PHP 8
$variable = $user && $user->exists ? true : null; // same as

3 years ago by Marco Pivetta — view source

unread

Hey David,

Hello!

It is just a suggestion to be discussed.

A lot of places on my projects I have codes like:

$companies = $user->companies->count()
? new Collection($user->companies)
: null;

Even upfront, this looks like a quite bad idea: either you have an empty
collection, or you have a collection with elements in it, but null is a
very different concept that isn't homogeneous with the resulting type.

I've updated multiple projects in the past where array|null was used as
return type for a collection-alike endpoint, removing null and replacing
it with an empty iterable instead.

3 years ago by David Rodrigues — view source

unread

Em sáb., 13 de fev. de 2021 às 00:11, Marco Pivetta ocramius@gmail.com
escreveu:

Hey David,

Even upfront, this looks like a quite bad idea: either you have an empty
collection, or you have a collection with elements in it, but null is a
very different concept that isn't homogeneous with the resulting type.

I've updated multiple projects in the past where array|null was used as
return type for a collection-alike endpoint, removing null and replacing
it with an empty iterable instead.

This was just one of the examples I have, but I still think it is valid.
Creating an empty collection costs resources that could be avoided in
situations where null could be used (as is most of my real use cases).

But there are other situations where we can use different situations to use
this type of resource.

One of the examples I have is when I need to include a class in an HTML
element conditionally. There are probably better ways, but this is a quick
one for simple situations:

<div class="user {{ $user => "online" }}">...</div> vs. <div class="user {{ $user ? "online" : null }}">...</div>

It is much easier to read, and I believe that the cost for language would
be minimal, and perhaps even more optimized than the use of ?:.

Atenciosamente,
David Rodrigues

Em sáb., 13 de fev. de 2021 às 00:11, Marco Pivetta ocramius@gmail.com
escreveu:

Hey David,

On Fri, Feb 12, 2021, 20:24 David Rodrigues david.proweb@gmail.com
wrote:

Hello!

It is just a suggestion to be discussed.

A lot of places on my projects I have codes like:

$companies = $user->companies->count()
? new Collection($user->companies)
: null;

Even upfront, this looks like a quite bad idea: either you have an empty
collection, or you have a collection with elements in it, but null is a
very different concept that isn't homogeneous with the resulting type.

I've updated multiple projects in the past where array|null was used as
return type for a collection-alike endpoint, removing null and replacing
it with an empty iterable instead.

3 years ago by Nikita Popov — view source

unread

On Fri, Feb 12, 2021 at 8:25 PM David Rodrigues david.proweb@gmail.com
wrote:

Hello!

It is just a suggestion to be discussed.

A lot of places on my projects I have codes like:

$companies = $user->companies->count()
? new Collection($user->companies)
: null;

So $companies will be null except if the user has companies.

My suggestion is create some kind of inline conditional to works like that:

$companies = $user->companies->count() => new Collection($user->companies);

If the conditional ($user->companies->count()) is true, then will return
the value (after =>), else will be null.

In my current work project, I have more than 100+ occurrences like that.

So some more examples:

$userCategory ? $userCategory->getTitle() : null
-> It could be optimized to the new nullsafe operator
$userCategory?->getTitle()

return $languageFirst instanceof LanguageExperience
? $languageFirst->title : null;
-> This not, with my suggestion we have:
return $languageFirst instanceof LanguageExperience =>
$languageFirst->title;

The => is just a suggestion, other options using existing keywords is:
return expr() then output();
return expr(): output();

I do not know if other languages support something like that.

Thanks!

There's a limited budget for this kind of syntax sugar, and we already have
?:, ??, ??=, ?->. I don't think there's space for yet another shorthand
conditional syntax.

Note that => cannot be used for this purpose, as it is already used for
array literals.

Regards,
Nikita

3 years ago by Guilliam Xavier — view source

unread

Hi,

I think it might be a good idea to check other languages to see if they
support something like this. We could use their examples as points of
reference for discussing whether to include this functionality in PHP.

On Sat, Feb 13, 2021 at 5:22 PM David Rodrigues david.proweb@gmail.com
wrote:

One of the examples I have is when I need to include a class in an HTML
element conditionally. There are probably better ways, but this is a quick
one for simple situations:
<div class="user {{ $user => "online" }}">...</div> vs. <div class="user {{ $user ? "online" : null }}">...</div>

There's a limited budget for this kind of syntax sugar, and we already have
?:, ??, ??=, ?->. I don't think there's space for yet another shorthand
conditional syntax.

Note that => cannot be used for this purpose, as it is already used for
array literals.

In Twig https://en.wikipedia.org/wiki/Twig_(template_engine), the "else"
part of a ternary expression is optional (colon included) and defaults to
an empty string. So, the following two lines are equivalent:

 <div class="user {{ user ? 'online' }}">...</div>
 <div class="user {{ user ? 'online' : '' }}">...</div>

Maybe PHP could do the same (with null rather than an empty string for
the default), i.e.:

var_dump(true ? 'foo'); // string(3) "foo"
var_dump(false ? 'foo'); // `NULL`

but I'm not sure if the added complexity (in parser implementation and in
the language) would be worth it, to avoid typing " : null", nor if everyone
would actually find the shorthand more readable?

--
Guilliam Xavier

3 years ago by Chase Peeler — view source

unread

On Fri, Feb 12, 2021 at 8:25 PM David Rodrigues david.proweb@gmail.com
wrote:

Hello!

It is just a suggestion to be discussed.

A lot of places on my projects I have codes like:

$companies = $user->companies->count()
? new Collection($user->companies)
: null;

So $companies will be null except if the user has companies.

My suggestion is create some kind of inline conditional to works like
that:

$companies = $user->companies->count() => new
Collection($user->companies);

If the conditional ($user->companies->count()) is true, then will return
the value (after =>), else will be null.

In my current work project, I have more than 100+ occurrences like that.

So some more examples:

$userCategory ? $userCategory->getTitle() : null
-> It could be optimized to the new nullsafe operator
$userCategory?->getTitle()

return $languageFirst instanceof LanguageExperience
? $languageFirst->title : null;
-> This not, with my suggestion we have:
return $languageFirst instanceof LanguageExperience =>
$languageFirst->title;

The => is just a suggestion, other options using existing keywords is:
return expr() then output();
return expr(): output();

I do not know if other languages support something like that.

Thanks!

There's a limited budget for this kind of syntax sugar, and we already have
?:, ??, ??=, ?->. I don't think there's space for yet another shorthand
conditional syntax.

I don't really have any strong feelings on this either way, but ?! seems
like a logical choice if it was to be implemented.

Note that => cannot be used for this purpose, as it is already used for
array literals.

Regards,
Nikita

--
Chase Peeler
chasepeeler@gmail.com

3 years ago by Albert Casademont — view source

unread

Another example is when a scalar input needs to be either left null or
converted to a Value Object:

$color = $data['color'] ? new Color($data['color']) : null;

This would become;

$color = $data['color'] ? new Color($data['color']); //the ": null" part is
implicit

It's actually kinda the inverse of the "?:" operator.

As Guilliam said, Twig already implements such behaviour, it's quite handy
when dealing with nullable variables,

On Fri, Feb 12, 2021 at 8:25 PM David Rodrigues david.proweb@gmail.com
wrote:

Hello!

It is just a suggestion to be discussed.

A lot of places on my projects I have codes like:

$companies = $user->companies->count()
? new Collection($user->companies)
: null;

So $companies will be null except if the user has companies.

My suggestion is create some kind of inline conditional to works like
that:

$companies = $user->companies->count() => new
Collection($user->companies);

If the conditional ($user->companies->count()) is true, then will
return
the value (after =>), else will be null.

In my current work project, I have more than 100+ occurrences like
that.

So some more examples:

$userCategory ? $userCategory->getTitle() : null
-> It could be optimized to the new nullsafe operator
$userCategory?->getTitle()

return $languageFirst instanceof LanguageExperience
? $languageFirst->title : null;
-> This not, with my suggestion we have:
return $languageFirst instanceof LanguageExperience =>
$languageFirst->title;

The => is just a suggestion, other options using existing keywords is:
return expr() then output();
return expr(): output();

I do not know if other languages support something like that.

Thanks!

There's a limited budget for this kind of syntax sugar, and we already
have
?:, ??, ??=, ?->. I don't think there's space for yet another shorthand
conditional syntax.

I don't really have any strong feelings on this either way, but ?! seems
like a logical choice if it was to be implemented.

Note that => cannot be used for this purpose, as it is already used for
array literals.

Regards,
Nikita

--
Chase Peeler
chasepeeler@gmail.com

3 years ago by Larry Garfield — view source

unread

Another example is when a scalar input needs to be either left null or
converted to a Value Object:

$color = $data['color'] ? new Color($data['color']) : null;

This would become;

$color = $data['color'] ? new Color($data['color']); //the ": null" part is
implicit

It's actually kinda the inverse of the "?:" operator.

As Guilliam said, Twig already implements such behaviour, it's quite handy
when dealing with nullable variables,

Please don't top post.

The advantage of ?: over long-ternary is that the part it lets you omit is of variable size, and is often verbose (nested array elements). That's not the case here, as the omitted portion is a fixed length short constant value (": null"). So the value of the abbreviation is much less.

I am also not a fan of null being commonly used, as it is a dangerous value. More often I would not want null if the color were missing but some default color value, which would be represented by something other than null anyway. And "falsy" is, as we've been reminded numerous times, a dangerous and tricky thing. (I just had several PRs against one of my OSS libraries because it relied on falsy, which had all sorts of incorrect failure conditions.)

I'd be a -1 here.

--Larry Garfield

3 years ago by Guilliam Xavier — view source

unread

On Tue, Feb 23, 2021 at 6:28 PM Larry Garfield larry@garfieldtech.com
wrote:

Another example is when a scalar input needs to be either left null or
converted to a Value Object:

$color = $data['color'] ? new Color($data['color']) : null;

This would become;

$color = $data['color'] ? new Color($data['color']); //the ": null" part
is
implicit

It's actually kinda the inverse of the "?:" operator.

As Guilliam said, Twig already implements such behaviour, it's quite
handy
when dealing with nullable variables,

Please don't top post.

The advantage of ?: over long-ternary is that the part it lets you omit is
of variable size, and is often verbose (nested array elements).

For completeness: it's also important when implying side effects, e.g.
foo() ?: bar() calls foo() only once, while foo() ? foo() : bar() may
twice.

That's not the case here, as the omitted portion is a fixed length short

constant value (": null"). So the value of the abbreviation is much less.

Indeed, and (as I said too) I'm not even sure whether it improves
readability or instead worsens it.
(By the way, I mentioned Twig because David's HTML example (Smarty maybe?)
reminded me of it, but that's a "template engine"; I actually don't know a
"general purpose" programming language with a C-like ternary operator that
allows omitting the ": null" part.)

I am also not a fan of null being commonly used, as it is a dangerous
value. More often I would not want null if the color were missing but some
default color value, which would be represented by something other than
null anyway.

You seem to join Marco here ;)

And "falsy" is, as we've been reminded numerous times, a dangerous and

tricky thing. (I just had several PRs against one of my OSS libraries
because it relied on falsy, which had all sorts of incorrect failure
conditions.)

I would just add that it's also true for "?:" (i.e. not specific to the
suggestion at hand, although the lack of a non-controversial example so far
may be a hint)

I'd be a -1 here.

--Larry Garfield

--

To unsubscribe, visit: https://www.php.net/unsub.php

--
Guilliam Xavier

3 years ago by Albert Casademont — view source

unread

On Tue, Feb 23, 2021 at 6:28 PM Larry Garfield larry@garfieldtech.com
wrote:

Please don't top post.

Sorry for that!

The advantage of ?: over long-ternary is that the part it lets you omit is
of variable size, and is often verbose (nested array elements). That's not
the case here, as the omitted portion is a fixed length short constant
value (": null"). So the value of the abbreviation is much less.

Sure, it's not a big deal having to write the ": null" but it doesn't add
any value

I am also not a fan of null being commonly used, as it is a dangerous
value. More often I would not want null if the color were missing but some
default color value, which would be represented by something other than
null anyway. And "falsy" is, as we've been reminded numerous times, a
dangerous and tricky thing. (I just had several PRs against one of my OSS
libraries because it relied on falsy, which had all sorts of incorrect
failure conditions.)

Agreed, it was just the first example that came to my mind, probably not
the best one. A better example would creating UUID VO's from string|null
input, there's little room for a default here.

3 years ago by Rowan Tommins — view source

unread

Sure, it's not a big deal having to write the ": null" but it doesn't add
any value

On the contrary, it adds an important piece of information: that the
default value is "null", rather than "false", or "0", or "new EmptyValue()".

For instance, it doesn't seem at all obvious to me that this code should
produce a null:

$items = [];
$itemCount = $items ? count($items);

I might be more convinced that "null" is the "natural" value if the
left-hand operand was only checked against null, not "falsiness": in
that case, you can read it as "if it's null, leave it alone". I'd still
be inclined towards "too specific to use up more syntax", though.

A better example would creating UUID VO's from string|null
input, there's little room for a default here.

I'm not quite sure what process you're describing, but any of the
following seem equally reasonable:

$nullableUuid = is_null($input) ? UUID::for($input) : null;
$maybeNilUuid = is_null($input) ? UUID::for($input) : UUID:nilValue();
$definitelyUuid = is_null($input) ? UUID::for($input) : UUID:random();

Syntax for skipping the is_null/isset on the left would be higher up my
wish list than syntax for skipping the :null on the right.

Regards,

--
Rowan Tommins
[IMSoP]

3 years ago by Sara Golemon — view source

unread

On Tue, Feb 23, 2021 at 1:05 PM Rowan Tommins rowan.collins@gmail.com
wrote:

Sure, it's not a big deal having to write the ": null" but it doesn't
add
any value

On the contrary, it adds an important piece of information: that the
default value is "null", rather than "false", or "0", or "new
EmptyValue()".

For instance, it doesn't seem at all obvious to me that this code should
produce a null:

$items = [];
$itemCount = $items ? count($items);

I might be more convinced that "null" is the "natural" value if the
left-hand operand was only checked against null, not "falsiness": in
that case, you can read it as "if it's null, leave it alone". I'd still
be inclined towards "too specific to use up more syntax", though.

This sums up my feelings on the topic very well. The proposal seeks to
remove clarity of intent to save four characters. Hard nope as presented.

-Sara

3 years ago by Mike Schinkel — view source

unread

Sure, it's not a big deal having to write the ": null" but it doesn't add
any value

On the contrary, it adds an important piece of information: that the default value is "null", rather than "false", or "0", or "new EmptyValue()".
For instance, it doesn't seem at all obvious to me that this code should produce a null:

$items = [];
$itemCount = $items ? count($items);

I might be more convinced that "null" is the "natural" value if the left-hand operand was only checked against null, not "falsiness": in that case, you can read it as "if it's null, leave it alone". I'd still be inclined towards "too specific to use up more syntax", though.

If you look at it from a software engineering perspective — which is how I assume most on this thread have been looking at it — being more explicit in the code is probably a better practice than saving a few keystrokes.

OTOH, if you view it from the perspective of a templating language — which is what PHP was initially created to be, for the web — then the reduction in visual noise from omitting ": null" would be a nice plus. And in the case of templating, the distinction between null vs. false vs. "" would be completely moot.

Repeating code similar to that from the message sent by David Rodrigues you can see that eliminating the ": null" is less noisy:

<div class="user &lt;?php user ? 'online' : null ?>">...</div>
<div class="user &lt;?php user ? 'online' ?>">...</div>

Although one ":null" eliminated is not that significant, PHP templates for generating HTML are notorious for having conditionals littered throughout, with tens if not hundreds of cases per file. So being able to drop the ":null" would be a really nice addition for templating.

To emphasize this feature address templating use-cases one might argue that dropping the "or" case of the trinary might only be supported when between single-line "<?php" and "?>" delimiters.

-Mike

P.S. Of course making it work differently between single-line delimiters and elsewhere would create inconsistency in the language so I probably would not actually argue that, I was just trying to make a rhetorical point.

3 years ago by Chase Peeler — view source

unread

On Feb 23, 2021, at 2:05 PM, Rowan Tommins rowan.collins@gmail.com
wrote:

Sure, it's not a big deal having to write the ": null" but it doesn't
add
any value

On the contrary, it adds an important piece of information: that the
default value is "null", rather than "false", or "0", or "new EmptyValue()".
For instance, it doesn't seem at all obvious to me that this code should
produce a null:

$items = [];
$itemCount = $items ? count($items);

I might be more convinced that "null" is the "natural" value if the
left-hand operand was only checked against null, not "falsiness": in that
case, you can read it as "if it's null, leave it alone". I'd still be
inclined towards "too specific to use up more syntax", though.

If you look at it from a software engineering perspective — which is how I
assume most on this thread have been looking at it — being more explicit in
the code is probably a better practice than saving a few keystrokes.

OTOH, if you view it from the perspective of a templating language —
which is what PHP was initially created to be, for the web — then the
reduction in visual noise from omitting ": null" would be a nice plus. And
in the case of templating, the distinction between null vs. false vs. ""
would be completely moot.

But PHP isn't just a templating language anymore. If PHP currently had
support for omitting the third argument on the ternary operator, then I'd
be against changes that required it given the BC implications. I think just
dropping the need for a third argument on the existing ternary operator is
a bad idea.

I'm ambivalent when it comes to creating a new symbol to support this
behavior (e.g. ?!). I don't really think it's necessary, but, I don't think
it hurts anything either.

Repeating code similar to that from the message sent by David Rodrigues
you can see that eliminating the ": null" is less noisy:
<div class="user &lt;?php user ? 'online' : null ?>">...</div>
<div class="user &lt;?php user ? 'online' ?>">...</div>
Although one ":null" eliminated is not that significant, PHP templates
for generating HTML are notorious for having conditionals littered
throughout, with tens if not hundreds of cases per file. So being able to
drop the ":null" would be a really nice addition for templating.

To emphasize this feature address templating use-cases one might argue
that dropping the "or" case of the trinary might only be supported when
between single-line "<?php" and "?>" delimiters.

-Mike

P.S. Of course making it work differently between single-line delimiters
and elsewhere would create inconsistency in the language so I probably
would not actually argue that, I was just trying to make a rhetorical point.

--

To unsubscribe, visit: https://www.php.net/unsub.php

--
Chase Peeler
chasepeeler@gmail.com

3 years ago by Mike Schinkel — view source

unread

Sure, it's not a big deal having to write the ": null" but it doesn't add
any value

On the contrary, it adds an important piece of information: that the default value is "null", rather than "false", or "0", or "new EmptyValue()".
For instance, it doesn't seem at all obvious to me that this code should produce a null:

$items = [];
$itemCount = $items ? count($items);

I might be more convinced that "null" is the "natural" value if the left-hand operand was only checked against null, not "falsiness": in that case, you can read it as "if it's null, leave it alone". I'd still be inclined towards "too specific to use up more syntax", though.

If you look at it from a software engineering perspective — which is how I assume most on this thread have been looking at it — being more explicit in the code is probably a better practice than saving a few keystrokes.

OTOH, if you view it from the perspective of a templating language — which is what PHP was initially created to be, for the web — then the reduction in visual noise from omitting ": null" would be a nice plus. And in the case of templating, the distinction between null vs. false vs. "" would be completely moot.

But PHP isn't just a templating language anymore.

It is not? If it is not I think a lot of userland PHP developers never got that decree.

If PHP currently had support for omitting the third argument on the ternary operator, then I'd be against changes that required it given the BC implications.

Can you elaborate on those?

I think just dropping the need for a third argument on the existing ternary operator is a bad idea.

I'm ambivalent when it comes to creating a new symbol to support this behavior (e.g. ?!). I don't really think it's necessary, but, I don't think it hurts anything either.

Repeating code similar to that from the message sent by David Rodrigues you can see that eliminating the ": null" is less noisy:
<div class="user &lt;?php user ? 'online' : null ?>">...</div>
<div class="user &lt;?php user ? 'online' ?>">...</div>
Although one ":null" eliminated is not that significant, PHP templates for generating HTML are notorious for having conditionals littered throughout, with tens if not hundreds of cases per file. So being able to drop the ":null" would be a really nice addition for templating.

To emphasize this feature address templating use-cases one might argue that dropping the "or" case of the trinary might only be supported when between single-line "<?php" and "?>" delimiters.

-Mike

P.S. Of course making it work differently between single-line delimiters and elsewhere would create inconsistency in the language so I probably would not actually argue that, I was just trying to make a rhetorical point.

--

To unsubscribe, visit: https://www.php.net/unsub.php https://www.php.net/unsub.php

--
Chase Peeler
chasepeeler@gmail.com <mailto:chasepeeler@gmail.com

3 years ago by Chase Peeler — view source

unread

On Tue, Feb 23, 2021 at 11:27 PM Mike Schinkel mike@newclarity.net
wrote:

On Feb 23, 2021, at 2:05 PM, Rowan Tommins rowan.collins@gmail.com
wrote:

Sure, it's not a big deal having to write the ": null" but it doesn't
add
any value

On the contrary, it adds an important piece of information: that the
default value is "null", rather than "false", or "0", or "new EmptyValue()".
For instance, it doesn't seem at all obvious to me that this code
should produce a null:

$items = [];
$itemCount = $items ? count($items);

I might be more convinced that "null" is the "natural" value if the
left-hand operand was only checked against null, not "falsiness": in that
case, you can read it as "if it's null, leave it alone". I'd still be
inclined towards "too specific to use up more syntax", though.

If you look at it from a software engineering perspective — which is how
I assume most on this thread have been looking at it — being more explicit
in the code is probably a better practice than saving a few keystrokes.

OTOH, if you view it from the perspective of a templating language —
which is what PHP was initially created to be, for the web — then the
reduction in visual noise from omitting ": null" would be a nice plus. And
in the case of templating, the distinction between null vs. false vs. ""
would be completely moot.

But PHP isn't just a templating language anymore.

It is not? If it is not I think a lot of userland PHP developers never
got that decree.

I didn't say it COULDN'T be used for templating (although even then I'd
suggest utilizing a library like smarty or twig) but that it is a lot more
than that. I'd also say that, good or bad, the direction of the language in
the last 15+ years has been to make it a robust language that can be used
for more than just templating.

If PHP currently had support for omitting the third argument on the
ternary operator, then I'd be against changes that required it given the BC
implications.

Can you elaborate on those?

Well it's pretty obvious. If php currently supported a statement like
"$user ? 'online'" then a change that requires the third argument would
break every single instance of code that didn't have one and wouldn't
provide any functional gains as nothing prevents people from adding on the
": null" if they want to. It's a moot point, though, because PHP doesn't
support it. I just included that statement to give context to my views -
that I'm not automatically in favor of changes just because they enforce
better development practices.

I think just dropping the need for a third argument on the existing
ternary operator is a bad idea.

I'm ambivalent when it comes to creating a new symbol to support this
behavior (e.g. ?!). I don't really think it's necessary, but, I don't think
it hurts anything either.
Repeating code similar to that from the message sent by David Rodrigues
you can see that eliminating the ": null" is less noisy:
<div class="user &lt;?php user ? 'online' : null ?>">...</div>
<div class="user &lt;?php user ? 'online' ?>">...</div>
Although one ":null" eliminated is not that significant, PHP templates
for generating HTML are notorious for having conditionals littered
throughout, with tens if not hundreds of cases per file. So being able to
drop the ":null" would be a really nice addition for templating.

To emphasize this feature address templating use-cases one might argue
that dropping the "or" case of the trinary might only be supported when
between single-line "<?php" and "?>" delimiters.

-Mike

P.S. Of course making it work differently between single-line delimiters
and elsewhere would create inconsistency in the language so I probably
would not actually argue that, I was just trying to make a rhetorical point.

--

To unsubscribe, visit: https://www.php.net/unsub.php
--
Chase Peeler
chasepeeler@gmail.com

--
Chase Peeler
chasepeeler@gmail.com

3 years ago by Mike Schinkel — view source

unread

Sure, it's not a big deal having to write the ": null" but it doesn't add
any value

On the contrary, it adds an important piece of information: that the default value is "null", rather than "false", or "0", or "new EmptyValue()".
For instance, it doesn't seem at all obvious to me that this code should produce a null:

$items = [];
$itemCount = $items ? count($items);

I might be more convinced that "null" is the "natural" value if the left-hand operand was only checked against null, not "falsiness": in that case, you can read it as "if it's null, leave it alone". I'd still be inclined towards "too specific to use up more syntax", though.

If you look at it from a software engineering perspective — which is how I assume most on this thread have been looking at it — being more explicit in the code is probably a better practice than saving a few keystrokes.

OTOH, if you view it from the perspective of a templating language — which is what PHP was initially created to be, for the web — then the reduction in visual noise from omitting ": null" would be a nice plus. And in the case of templating, the distinction between null vs. false vs. "" would be completely moot.

But PHP isn't just a templating language anymore.

It is not? If it is not I think a lot of userland PHP developers never got that decree.

I didn't say it COULDN'T be used for templating

It would benefit when PHP is used for templating which is still not an insignificant use of PHP, and this one simple thing would make templates easier to read.

(although even then I'd suggest utilizing a library like smarty or twig)

A long time ago I read an article on SitePoint that argued that using templating languages like Smarty or Twig made no sense because PHP is a templating language.

That stuck with me and for many (most?) use-cases I still agree with it.

but that it is a lot more than that. I'd also say that, good or bad, the direction of the language in the last 15+ years has been to make it a robust language that can be used for more than just templating.

I think it is specious to imply allowing $foo ? $bar to return null when $foo==nil would make PHP a less robust language.

Well it's pretty obvious. If php currently supported a statement like "$user ? 'online'" then a change that requires the third argument would break every single instance of code that didn't have one and wouldn't provide any functional gains as nothing prevents people from adding on the ": null" if they want to. It's a moot point, though, because PHP doesn't support it. I just included that statement to give context to my views - that I'm not automatically in favor of changes just because they enforce better development practices.

So 1.) if PHP allowed "$user ? 'online'" and then 2.) PHP later disallowed it then 3.) the second action would created BC concerns?

Yes, that is a moot point with respect to the proposal posed by this thread.

-Mike

3 years ago by Rowan Tommins — view source

unread

Repeating code similar to that from the message sent by David Rodrigues you can see that eliminating the ": null" is less noisy:
 <div class="user &lt;?php user ? 'online' : null ?>">...</div>
 <div class="user &lt;?php user ? 'online' ?>">...</div>

That's an interesting use case. Of course, the logical default there is
actually '' (empty string) rather than null, since "echo null;" doesn't
actually make much sense, but null works under current type juggling rules.

Then again, perhaps this is a good example of why PHP isn't the best
choice any more if you want a templating language. Twig, for instance,
has exactly this facility
[https://twig.symfony.com/doc/3.x/templates.html#other-operators]:

{{ foo ? 'yes' : 'no' }}
{{ foo ?: 'no' }}is the same as {{ foo ? foo : 'no' }}
{{ foo ? 'yes' }}is the same as {{ foo ? 'yes' : '' }}

If we wanted to close the gap between Twig and plain PHP for templating,
this might be somewhere on the list, but after more significant features
like automatic escaping, filter / pipe syntax, etc.

Regards,

--
Rowan Tommins
[IMSoP]

3 years ago by Kingsquare.nl - Robin Speekenbrink — view source

unread

Op do 25 feb. 2021 om 09:55 schreef Rowan Tommins rowan.collins@gmail.com:

Repeating code similar to that from the message sent by David Rodrigues
you can see that eliminating the ": null" is less noisy:
 <div class="user &lt;?php user ? 'online' : null ?>">...</div>
 <div class="user &lt;?php user ? 'online' ?>">...</div>

Also just to have a somewhat apples to apples (working) comparission, it
would read something like:

<div class="user <?php echo user ? 'online' : null ?>">...</div> <div class="user <?php echo user ? 'online' ?>">...</div> or one of the other possibilities (print / <?= ) and with the addition of those extra strokes i fail to see the cleanliness over the added obtrusion of `the one way`: <div class="user <?= user ? 'online' : '' ?>">...</div>

this atleast is consistent, doesn't provide yet another way of doing the
same thing and even allows for easier refactoring (i.e. if the falsy
situation has to change from '' to 'offline')

to summarize from my 2cent userland developer standpoint: Hope this doesn't
make this in... (or is at least strongly discouraged)

Thanks,
Robin

3 years ago by Larry Garfield — view source

unread

Repeating code similar to that from the message sent by David Rodrigues you can see that eliminating the ": null" is less noisy:
 <div class="user &lt;?php user ? 'online' : null ?>">...</div>
 <div class="user &lt;?php user ? 'online' ?>">...</div>
That's an interesting use case. Of course, the logical default there is
actually '' (empty string) rather than null, since "echo null;" doesn't
actually make much sense, but null works under current type juggling rules.

Then again, perhaps this is a good example of why PHP isn't the best
choice any more if you want a templating language. Twig, for instance,
has exactly this facility
[https://twig.symfony.com/doc/3.x/templates.html#other-operators]:

Given the security implications of plain PHP templating (you're on your own for absolutely all filtering and escaping, good luck), I'd say this isn't even a controversial statement. PHP may have begun life as a templating language for the 90s, but in the 2020s it's a fairly mediocre option at best compared to what else is readily available.

--Larry Garfield