Welcome! Log In Create A New Profile

Advanced

[PHP-DEV] [RFC] [Discussion] Class Friendship in PHP

Posted by Dustin Wheeler 
Dustin Wheeler
[PHP-DEV] [RFC] [Discussion] Class Friendship in PHP
September 22, 2017 02:40AM
Hi!

Almost two years ago, I opened discussion regarding my proposal to support
class friendship through a new keyword, `friend`. For a number of reasons
(mostly work-related; some personal) I didn't have time to move the process
forward. There were several hot-topic discussions going through the mailing
list at the time that I believed deserved more attention. I feel that my
RFC was ill-timed, perhaps. That said, we're almost two years into the
future!

I would like to re-open discussion of my RFC and get it to vote
appropriately soon.

In the previous discussion, there was on-going work to introduce
package-private classes by Guilherme. As far as I can tell that work is
stalled, pending fundamental changes to how namespaces are implemented in
PHP (https://github.com/php/php-src/pull/947#issuecomment-224912697). Like
many others, I am very excited for the possibility of package-private
classes or namespace visibility features of any kind. Even so, my personal
vision of that landscape includes class friendship as a unique, explicit
and concise expression of tight-coupling between object-oriented
collaborators. I do not feel these efforts are either-or. I believe they
both contribute to a more expressive means of specifying the relationship
between collaborators with regard to visibility.

The RFC is at:

https://wiki.php.net/rfc/friend-classes

I need to rebase my implementation against `master` and send a PR. I will
have that done before voting begins as it includes several test cases and I
would appreciate additional perspectives to catch edge-cases I may not have
thought of. The implementation was quite straight-forward.

Thanks!

--
Dustin Wheeler | Software Developer
NC State University
mdwheele@ncsu.edu
"If you don't know where you're going, it's easy to iteratively not get
there."
Marco Pivetta
Re: [PHP-DEV] [RFC] [Discussion] Class Friendship in PHP
September 22, 2017 04:00AM
Hey Dustin,

Given the amount of use-cases (very low, so far) this feels much much
easier if just implemented in userland via `Closure::call()`, since the
scope of the patch complicates the class model of PHP by a whole lot.

Another simple solution is to use things such as leedavis81/altr-ego, which
already exists in userland.

On 22 Sep 2017 02:34, "Dustin Wheeler" <[email protected]> wrote:

> Hi!
>
> Almost two years ago, I opened discussion regarding my proposal to support
> class friendship through a new keyword, `friend`. For a number of reasons
> (mostly work-related; some personal) I didn't have time to move the process
> forward. There were several hot-topic discussions going through the mailing
> list at the time that I believed deserved more attention. I feel that my
> RFC was ill-timed, perhaps. That said, we're almost two years into the
> future!
>
> I would like to re-open discussion of my RFC and get it to vote
> appropriately soon.
>
> In the previous discussion, there was on-going work to introduce
> package-private classes by Guilherme. As far as I can tell that work is
> stalled, pending fundamental changes to how namespaces are implemented in
> PHP (https://github.com/php/php-src/pull/947#issuecomment-224912697). Like
> many others, I am very excited for the possibility of package-private
> classes or namespace visibility features of any kind. Even so, my personal
> vision of that landscape includes class friendship as a unique, explicit
> and concise expression of tight-coupling between object-oriented
> collaborators. I do not feel these efforts are either-or. I believe they
> both contribute to a more expressive means of specifying the relationship
> between collaborators with regard to visibility.
>
> The RFC is at:
>
> https://wiki.php.net/rfc/friend-classes
>
> I need to rebase my implementation against `master` and send a PR. I will
> have that done before voting begins as it includes several test cases and I
> would appreciate additional perspectives to catch edge-cases I may not have
> thought of. The implementation was quite straight-forward.
>
> Thanks!
>
> --
> Dustin Wheeler | Software Developer
> NC State University
> mdwheele@ncsu.edu
> "If you don't know where you're going, it's easy to iteratively not get
> there."
>
Sorry, only registered users may post in this forum.

Click here to login