Constructor Destructor in Structure

[mof]

Is it possible to have a structure say initialise itself without defining it as a class? I'm probably being pedantic, so let me know if I am.

e.g.

typedef struct a
{
   char *ch;
} a_t;

Something that would initialise ch to null.

[D80DSL3A]

A structure can have a constructor, just like a class.
Why typedef?

struct a
{
   char *ch;
   a(): ch(NULL){}// constructor initializes ch = NULL
} a_t;

[mof]

Thankyou, that will be handy to know.
'typedef' so that a_t becomes a variable type.

[D80DSL3A]

You're welcome. I'm still confused about using typedef for a structure. The structure a IS it's own variable type:
a a1, a2, a3;// 3 instances of the variable type a are declared here
Could you show an example where the typedef is helpful?

[1yR4jE8b]

typedef struct a
{
   char *ch;
} a_t;

This is standard C.

struct a
{
   char *ch;
   a(): ch(NULL){}// constructor initializes ch = NULL
} a_t;

This is the C++ way of doing it, both ways are valid in C++ because compilers need to be backward compatible with C. However typically, you ignore the second typename at the end of the C++ style struct definition...

struct a
{
   char *ch;
   a(): ch(NULL){}// constructor initializes ch = NULL
};

Semantically, there is no difference.

[1yR4jE8b]

EDIT: removed a double post

wtf, there's like 3 copies of this thread?

[D80DSL3A]

@darkestfright. Thanks for the explanation.

This is standard C.

That explains why I didn't recognize it.

[mof]

Thanks also, there's a lot of C++ syntax I'm not familiar with.

[sohguanh]

All you need to know is C++ is created to be backward-compatible to old existing C code and with such baggage you can have such "cross-breed" code.

All new C++ programs should strive to do it the "C++ syntax way" and for legacy C program maintenance, we revert back to "C syntax way".

I don't really recommend "cross-breed" code but that is entirely my personal opinion of cuz.