On Tue, 4 Mar 2025 at 22:34, Tim Düsterhus <t...@bastelstu.be> wrote:
> Hi > > On 3/4/25 23:18, Gina P. Banyard wrote: > > Those two classes predate the Namespaces in bundled PHP extensions [1] > RFC, I do think adding a namespace is better. > > And we could also move the two classes into the new namespace and add > class aliases for the global ones, considering they are quite recent > additions to PHP. > > I agree that adding a namespace would be appropriate here. Looking at > the sockets.stub.php, it pollutes the global namespace quite a bit. I > probably miscounted, but it appears to be at least 10 different prefixes > for the global constants. > Ok I ll go the namespace route then, fair point. > > It would probably be in order to also alias all the constants into the > namespace (possibly converting them to enums where appropriate) and then > in a follow-up version deprecate the global ones. I think that would > bring quite a bit of positive impact, for minimal effort. > > -------- > > As for the classes themselves: It does not appear to be defined in our > naming policy [1], but the properties should likely use camel-case, as > that is the established convention for userland code. And I'm also > curious why the port properties are strings, not ints? Yes it is a definition mistake it means to be a int indeed. > And what is the > payload object, is that just stdClass? > Not really it's supposed to be one of the new classes. > > Best regards > Tim Düsterhus > > [1] > https://github.com/php/policies/blob/main/coding-standards-and-naming.rst >
