[Bug fortran/15586] [3.5? Regression] gfortran should support i18n

["tobi at gcc dot gnu dot org" <gcc-bugzilla at gcc dot gnu dot org>]

------- Additional Comments From tobi at gcc dot gnu dot org  2004-06-09 17:01 -------
(In reply to comment #2) 
> For example, this code in gcc/fortran/resolve.c:
> 
>       gfc_error ("Assumed %s array at %L must be a dummy argument",
>                 sym->as->type == AS_ASSUMED_SIZE ? "size" : "shape",
>                  &sym->declared_at);
> 

A quick check for other such lines:
[tobi@marktplatz fortran]$ grep -n "[?] [\"]" *c
error.c:430:      error_string (have_l1 ? "(2)" : "(1)");
iresolve.c:804:  name = mask ? "mmaxloc" : "maxloc";
iresolve.c:825:    gfc_get_string ("__%s_%c%d", mask ? "mmaxval" : "maxval",
iresolve.c:867:  name = mask ? "mminloc" : "minloc";
iresolve.c:887:    gfc_get_string ("__%s_%c%d", mask ? "mminval" : "minval",
iresolve.c:976:    gfc_get_string ("__%s_%c%d", mask ? "mproduct" : "product",
iresolve.c:1207:    gfc_get_string ("__%s_%c%d", mask ? "msum" : "sum",
resolve.c:3740:          sym->as->type == AS_ASSUMED_SIZE ? "size" : "shape",
scanner.c:1098:                    (filename[0] == '\0') ? "<stdin>" : filename);
trans-intrinsic.c:521:          ts->type == BT_COMPLEX ? "c" : "",
trans-intrinsic.c:523:          ts->kind == 4 ? "f" : "");

The only other line which is related to error printing is the one in error.c. It
chooses between "(1)" and "(2)", which are used for designating source locations
in error messages. They should be the same in all locales, so this poses no
problem to i18n.

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |tobi at gcc dot gnu dot org


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=15586