[gfortran] Followup patch: Re: revap types, step 1

Tobias Schlüter tobias.schlueter@physik.uni-muenchen.de
Fri Aug 27 13:56:00 GMT 2004


Richard Henderson wrote:
> + /* ??? These functions should go away in favor of direct access to
> +    the relevant variables.  */
> + 
> + gfc_default_integer_kind (void)
> + gfc_default_real_kind (void)
> + gfc_default_double_kind (void)
> + gfc_default_character_kind (void)
> + gfc_default_logical_kind (void)
> + gfc_default_complex_kind (void)

I did this with the help of a sed script. Patch attached, compiled and tested
with no regressions. The ChangeLog is sketchy as those changes are completely
mechanical, if you wish for a more detailed ChangeLog, I will prepare it.

Ok to commit?

- Tobi

2004-08-27  Tobias Schlueter  <tobias.schlueter@physik.uni-muenchen.de>

	* gfortran.h (gfc_default_*_kind): Remove prototypes, add extern
	variable declaration of same name.
	* arith.c, check.c, decl.c, dump_parse_tree.c, expr.c,
	intrinsic.c, io.c, iresolve.c, match.c, options.c, primary.c,
	resolve.c, simplify.c, symbol.c, trans-const.c, trans-io.c:
	Replace all calls to gfc_default_*_kind with variable accesses.
	* trans-types.c: Same as above.
	(gfc_default_*_kind_1): Rename to gfc_default_*_kind, remove
	static qualifier. Replace all occurences.
	(gfc_default_*_kind): Remove functions.
-------------- next part --------------
An embedded and charset-unspecified text was scrubbed...
Name: varpatch
URL: <http://gcc.gnu.org/pipermail/fortran/attachments/20040827/c8754c42/attachment.ksh>


More information about the Fortran mailing list