Re: T_CONCAT_EQUAL operator and object's write/read handlers

hi Piotr, On Fri, May 30, 2008 at 10:40 PM, Piotr Pokora <piotrek.pokora@gmail.com> wrote: > Pierre Joye writes: >> hi, > > Hi! Thanks for such quick reply! > >>> In first example only read handler is invoked ,whilst read and write >>> ones in latter. >>> I doubt it's write handler issue. Should I have another handler like >>> get_property_ptr_ptr_t? >> >> (.+-*/)=, ++, -- can't be detected from an extension point of view. I >> tried to convince internals a couple of years ago that it is amazingly >> handy to have such features available but it was affiliated to a >> complete operator overloading and was rejected. I'm not sure there is >> any chance to get it in now but who knows. > > So in theory ( if I have own read/write handlers ), I should duplicate > every property and ( at best ) register all of them as zend's class > properties? '.=' works fine for built-in handlers, but I would like to > know what should be sacrified ( and done ) to work around this issue. ++ for example will do a property read followed by a property write. .= should work too, you only don't know that a .= is being executed (unlike a->f=1). -- Pierre http://blog.thepimp.net | http://www.libgd.org