Re: [PHP-DEV] [RFC] Function Autoloading (mark 5)

Hi Tim,=20

I posted this in the other thread, but I'm not sure it got through, so I'l=
l repost here; apologies if both show up at once=2E

On 31 July 2026 18:12:01 BST, "Tim D=C3=BCsterhus" <tim@bastelstu=2Ebe> wr=
ote:
>> Without giving it detailed thought, I believe =E2=80=9CNamespaced funct=
ion, Autoload namespaced function, Global fallback=E2=80=9D (i=2Ee=2E swap =
steps 2 and 3) would result in much more predictable semantics already - th=
e global fallback would work just as before, it just wouldn't have autoload=
ing support=2E


Semantically, that makes most sense, but it leads to terrible performance =
for the global fallback case, unless there's some way to avoid multiple che=
cks for the same namespaced function=2E That's why we keep going round in c=
ircles on the whole issue=2E=20

I think Michael's "namespace autoloader" idea is a much cleaner way forwar=
d - it avoids the fallback issue without creating lots of extra special cas=
es, and in reality "load a file with all functions for namespace Foo\Bar" i=
s likely to be more common than "load a file which specifically defines fun=
ction Foo\Bar\baz()"=2E

Regards,

Rowan Tommins
[IMSoP]