When I try to study QP/CPP code I came across below line.
QTimeEvt *t; // ... if (t == static_cast<QTimeEvt *>(0)) { Why they are doing static_cast of 0? If they want to check NULL we can do that directly right?
This source code you can find out in
http://www.state-machine.com/qpcpp/qf__time_8cpp_source.html
Yeah, that's unnecessary, though it may be mandated by some style guide for "clarity", or it may be there to silence an overzealous static analysis tool.
Of course, nowadays, we'd just write nullptr and leave it at that.
!t ?t is (accidentally or otherwise) etc etc. (I'm being a bit picky. I like to use the type system, though. Its job is to aid in ensuring program correctness, and writing out == nullptr is hardly expensive.)
The idiomatic way to write
QTimeEvt *t; // ... if (t == static_cast<QTimeEvt *>(0)) { … is
QTimeEvt* t; // ... if( !t ) { Or you can write it as
if( not t ) { … although you can also write
if( t == nullptr ) { … or, ¹C++03-style,
if( t == 0 ) { There's no need for the cast.
It's just verbiage consuming the reader's time.
Notes:
¹ If the <stddefs.h> header is included one can write NULL instead of 0 where a nullpointer is required. With a modern implementation NULL might even be defined as nullptr.
if( not t ) Is this commonplace?NULL can be nullptr since C++11 — good to find out!<iso646.h> to use not, that is, MSVC is not standard-conforming here. I use a forced include to make it more conformant.main for its most common use… Heh. :)