The problem is that first class callables actually makes a new object wrapper for each reference, so trim(…) != trim(…). (It can be true in some cases, it depends if the memory is freed for the first reference).
I think the point of introducing the (...) syntax that for some reason this article apparently denigrates (if I even understand it correctly) is to eventually deprecate that.
Why? The article fails to explain.
> The introduction of first-class callables simplifies callback handling. You no longer have an excuse to define your callbacks like this:
> <?php
> $data = array_map(trim(...), [' x', 'z ']);
“First-class callables” is that exact syntax.
Exactly. The article kind of circles around the idea that it's bad, but never really lands on why.
My experience is with modern annotations, callable and iterable are pretty powerful, convenient, and safe.
- https://phpstan.org/writing-php-code/phpdoc-types#callables
- https://phpstan.org/writing-php-code/phpdoc-types#iterables
The problem is that first class callables actually makes a new object wrapper for each reference, so trim(…) != trim(…). (It can be true in some cases, it depends if the memory is freed for the first reference).
And how is Closure::fromCallable('trim') recommended by the article any better in that respect?
It is not better, this is a problem with the first-class callable syntax altogether.
Stop using languages where strings are callable :-p
I think the point of introducing the (...) syntax that for some reason this article apparently denigrates (if I even understand it correctly) is to eventually deprecate that.