Re: [RFC] CodingGuidelines: mark external declarations with "extern"

[Denton Liu <liu.denton@gmail.com>]

Hi Junio,

On Fri, Oct 09, 2020 at 01:33:39PM -0700, Junio C Hamano wrote:
> Jeff King <peff@peff.net> writes:
> 
> > The argument for including it is less clear to me. You say below:
> >
> >> [...]By doing so, we would also prevent a
> >> mistake of not writing "extern" when we need to (i.e. decls of data
> >> items, that are not functions) when less experienced developers try
> >> to mimic how the existing surrounding declarations are written.
> >
> > but to my recollection that has not been a big problem. And it's one
> > that's usually easily caught by the compiler. A missing "extern" on a
> > variable will usually get you a multiple-definition warning at
> > link-time (if you manage to also omit the actual definition you won't
> > see that, though "make sparse" will warn that your variable ought to be
> > static).
> 
> Not really, that is where the "common" extension comes in, to help
> us with it hurt others without it, unknowingly X-<.

I'm not really sure what you mean by the "common" extension. 

>         $ cat >a.c <<\EOF
>         #include <stdio.h>
>         #include "c.h"
> 
>         int common = 47;
> 
>         int main(int ac, char **av)
>         {
>             printf("%d\n", common + other);
>             return 0;
>         }
>         EOF
>         $ cat >b.c <<\EOF
>         #include "c.h"
> 
>         int other = 22;
>         EOF
>         $ cat >c.h <<\EOF
>         int common;
>         int other;
>         EOF
>         $ gcc -Wall -o c a.c b.c; ./c
>         59

On gcc 10.2.0, it errors out successfully. Although on clang 10.0.1, it
compiles successfully and produces "69". That being said, I think extern
variables are relatively rare in our codebase and, when it happens, they
usually come as part of lists of other extern variables so a developer
who's mimicking the surrounding code would be able to copy it
successfully. Otherwise, the decl usually pops out in header files as it
is quite unusual.

> And I have a strong preference, after thinking about it, to have
> "extern" in front in the declarations.  It gives another clue for
> patterns I feed to "git grep" to latch onto, and help my eyes to
> scan and tell decls and defns apart in the output.  The benefit
> alone is worth the extra 7 columns in front spent, which you call
> "clutter".

To be honest, I do not have any preference between having the explicit
extern or not. I do have a strong preference, however, for having a
codebase that's consistently written. When I was doing the refactor, I
wouldn't have minded introducing extern everywhere although that wasn't
suggested as an alternative.

I agree that these are all benefits of declaring functions explicitly as
extern. However, I don't think they're worth the cost of either another
huge rewrite or an inconsistent codebase.

> > IMHO the real problem here is that C's syntax for returning a function
> > pointer is so horrendous. How about this (on top of your earlier patch
> > to drop the extern from that declaration)?
> 
> In general, I like a typedef for callback function that shortens the
> decl of a function that takes such a callback, so I think
> 
> > +void set_error_routine(report_fn routine);
> > +void set_warn_routine(report_fn routine);
> > +report_fn get_error_routine(void);
> > +report_fn get_warn_routine(void);
> 
> these are good, but they are better with "extern" in front in a
> header file to make it clear they are declarations and not
> definitions when they appear in "git grep" output.

I agree that this looks a lot better, with or without the extern.