Replaced ancient K&R function declarations to be C23 compatible

[seanm]

No description provided.

[seanm]

@WardF @DennisHeimbigner The codebase has a lot of old K&R function declarations. I thought it would just be just a few, so I started fixing them, but I soon realized there were many. So I stopped to share what I started, so it can be discussed. Is it just historical cruft that there are so many?

This syntax is no longer compiles with C23.

[DennisHeimbigner]

The code where this is occurring is indeed very old; from the first incarnations of the netcdf-c library.

[seanm]

@DennisHeimbigner ok so I can proceed to fixing them all?

Assuming so, one question: where do you prefer I move comments related to a specific variable, like this case for varid:

void
del_att (test, varid, iatt)	/* delete attribute iatt in the netcdf test */
     struct netcdf *test;
     int varid;			/* variable id */
     struct cdfatt *iatt;

should it end up like this maybe?

/* delete attribute iatt in the netcdf test */
void
del_att (struct netcdf *test,
         int varid, /* variable id */
         struct cdfatt *iatt)

[seanm]

@DennisHeimbigner @WardF any preference about those comment placements?

[DWesl]

Another option, if you don't mind setting up doxygen, would be:

/** delete attribute iatt in the netcdf test
 *
 * \param varid variable id
 */
void
del_att (struct netcdf *test, int varid, att *iatt)

It looks like adding \private at the end will leave the function out of the public documentation.
Alternately,

/** delete attribute iatt in the netcdf test */
void
del_att(struct netcdf *test,
            int varid /**< variable id */,
            att *iatt)

combines doxygen with the second style above

[seanm]

@DWesl almost all the affected files are test files... not sure using doxygen makes sense there...?

[WardF]

No particular preference; Doxygen would be great (even in test files), but the lack of doxygen wouldn't prevent adoption of a PR. Thanks! We're focused on working through the backlog, and making good progress.

[seanm]

No particular preference

OK, this is done then IMHO.

[edwardhartnett]

Thanks for this!