Issue: https://github.com/php/php-src/issues/7796
Comment Author: cmb69
> I hereby would like to request the option to send the CData class to FFI bound functions, to
> bypass this type juggling
But why? Regarding the performance, that float <-> double casting is likely way more efficient
than constructing new CData objects. And as far as I know, a float to double conversion is lossless
(assuming IEEE 754 binary32/binary64).
In my opinion, that would only make sense, if PHP had a native float type.