Re: BC break: ReflectionMethod::invoke() expects parameter1to be object, string given

From: Date: Tue, 23 Aug 2016 15:05:26 +0000
Subject: Re: BC break: ReflectionMethod::invoke() expects parameter1to be object, string given
References: 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19  Groups: php.internals 
Request: Send a blank email to internals+get-95418@lists.php.net to get a copy of this message
On Tue, Aug 23, 2016 at 4:58 PM, Levi Morrison <levim@php.net> wrote: > On Tue, Aug 23, 2016 at 8:49 AM, Julien Pauli <jpauli@php.net> wrote: >> On Tue, Aug 23, 2016 at 4:13 PM, Levi Morrison <levim@php.net> wrote: >>> On Tue, Aug 23, 2016 at 7:56 AM, Alexander Lisachenko >>> <lisachenko.it@gmail.com> wrote: >>>> >>>> 2016-08-23 16:40 GMT+03:00 Julien Pauli <jpauli@php.net>: >>>>> >>>>> My patch allows that, but I can't find a use case to it. >>>> >>>> >>>> >>>> My use case for that was decorating of static methods with additional >>>> behaviour, so it's possible to cache the result of static methods, etc. >>>> Unfortunately, reflection API doesn't provide me such an ability, so I >>>> decided to switch to the combination of special closure with >>>> forward_static_call_array() and binding it to the desired scope. >>>> >>>> >>>> 2016-08-23 16:40 GMT+03:00 Julien Pauli <jpauli@php.net>: >>>>> >>>>> Try it. >>>> >>>> >>>> Yes, sure, I will check. However as I say earlier, now I use closure >>>> extraction and scope binding as more natural way to invoke such methods. So >>>> missing scope argument for static methods and ReflectionMethod->invoke() is >>>> a nice thing to have, but not a critical one. >>> >>> Seems to me you should be calling getClosure() and using >>> Closure::bindTo before invoking it. ReflectionMethod::invoke is really >>> just a short-cut for a common case; I don't think it's reasonable to >>> have it mirror the Closure API just to save one method call to get a >>> Closure. >> >> This cannot work. >> There is a special use case forbidding that. >> "Cannot rebind scope of closure created by >> ReflectionFunctionAbstract::getClosure()" >> >> >> Julien > > I know this is a stability concern for internal classes but is there > anything really preventing this from user-land classes? I remember problems and bugs around this , but I don't remember the exact use-case why we forbid that. I guess Dmitry, Xinchen or Nikic could remember. Julien

« previous php.internals (#95418) next »