News for hackers. Stuff that matters.
Re:Humans should think of sizeof() as a function, says Linus... by imglorp on Saturday May 30, 2015 @08:54AM (#9630008)
Which is why it's nice to lift stuff out into typedefs. It centralizes them (DRY principle) and avoids this issue.

    typedef int thing_t;
    ...
    thing_t *foo;
    // code
    foo = malloc(sizeof(thing_t));
Re:Humans should think of sizeof() as a function, says Linus... by viraptor on Saturday May 30, 2015 @09:10AM (#9630055)
Why is that better than just getting `sizeof(* foo)`? Even with typedef, I can see this happening in the future:

    typedef int thing_t;
    ...
    thing_t *foo_internal;
    thing_wrapper_t *foo;
    // code
    foo = malloc(sizeof(thing_t));
If you do:

    foo = malloc(sizeof(*foo));
That's at least always on the same line.
Re:Humans should think of sizeof() as a function, says Linus... by userbinator on Saturday May 30, 2015 @09:11AM (#9630057)
That just makes it more verbose.

This, on the other hand, always allocates one object of foo's pointed-to-size, whatever its type:

    foo = malloc(sizeof(*foo));
Re:Humans should think of sizeof() as a function, says Linus... by dllthomas on Saturday May 30, 2015 @03:04PM (#9631367)
As an aside, I think nearly any time you want a typedef, it's worth wrapping it in a struct.

    typedef struct { int value; } thing_t;
That way the compiler catches it when you try to pass the wrong thing (at least, more of the time).
Re:Humans should think of sizeof() as a function, says Linus... by dllthomas on Saturday May 30, 2015 @03:08PM (#9631384)
DRY is good, but making the structure of your code reflect the actual semantics you want is better. What you want is to allocate space for foo. So write that.
Re:Humans should think of sizeof() as a function, says Linus... by crazy2be on Saturday May 30, 2015 @09:07AM (#9630050)
Well, in this (simple) case you can just do

    int *foo;
    foo = malloc(sizeof(*foo));
And avoid the brittleness mentioned.