Bug 622 – There should be a warning for unininitalized class reference

This error is very common for people with C++ background (I've already seen people asking questions about similar code several times): ---------------------------------------------- class Foo{ public int data; } void main() { Foo f; // obviously they try do this way as they got used to it in C++ // should be a warning here! f.data = 1;// Access violation } ----------------------------------------------

Yes please! I make this mistake all the time, and every time it takes a few minutes tracking down why and where my program is crashing.

[email protected] escribi

This bug is half-fixed: the compiler is now able to catch this bug at compile-time, but only if the code is compiled with -O. I think the compiler has to catch this bug when not optimized mode is used too.

This issue would may be solved in 2 ways: 1. Issue a warning whenever you have an initialized class reference. This is not a valid solution because you may want an initialized class reference that you will use later to point to another class instance. 2. Implement a dataflow analysis mechanism that will track all uses of f. This is against dmd's principle of not doing complex dataflow analysis. Even it we were to implement this, there would still be situations were you cannot be sure whether you are dereferencing a null pointer: class Foo { int a; } void fun(Foo a); void main() { Foo f; fun(f); f.a; // => cannot know what fun did to a. } In addition, this sort of access violation is extremely easy to solve because you get a stack trace. Having a warning pointing it out is really redundant. Closing as WONTFIX.