structure type and variable name

Question

hi all i just wanted to know whether we can declare variable name as structure name.

for example

typedef struct
{
  char c;
}t;

then in some function can i use

fun()
{
  t t;
}

is this valid? if so then how compiler differentiate between them?

Answer 1

yes because each thing lives in a different place in the compilers understanding.

t t;

The compiler is expecting a type when it encounters the first t and it has a type called t.

Edit: To address comments.

I'm not talking about scope.

But since I've not written a compiler (only interpreters) I don't know the term. The compiler is expecting a token at the first t to be a type, it also knows what type have been declared up to that point. Since it see a name that refers to a type it's happy. Whereas if a token that wasn't a type was found there then it would rightly signal an error.

Answer 2

parser first gets the data type and maintains a different table and later part as the variable name. So it works absolutely.

Answer 3

Yes, but why would you want to? If you want bugs and errors to thrive in your project, then go right ahead and name variables after types.

Answer 4

fun() { t t; }

is this valid ?

No it is not. Return type of fun() is missing and implicit int return type is deprecated.

However void fun(){ t t ;} is syntactically valid.

Answer 5

Yes, it is valid. If you do that, then the structure type is hidden in the enclosing scope and t refers only to the declared variable.