cobol: Use %td in error_msg in 3 spots

On Thu, Jul 31, 2025 at 11:33:07PM +0200, Jakub Jelinek via Gcc wrote:
> > this was all described in excruciating detail in the patch submission
> >
> >     https://gcc.gnu.org/pipermail/gcc-patches/2025-June/687385.html
> >
> > and the commit message.
>
> Looking at that patch, the dbgmsg change looks correct (dbgmsg is
> ATTRIBUTE_PRINTF_1), while the last 3 hunks are suboptimal, they should
> really use %td and keep the ptrdiff_t arguments without casts.

Here it is in patch form.
I couldn't find other similar casts in calls to ATTRIBUTE_GCOBOL_DIAG
functions.

2025-08-02  Jakub Jelinek  <[email protected]>

	* parse.y (intrinsic): Use %td format specifier with no cast on
	argument instead of %ld with cast to long.
	* scan_ante.h (numstr_of): Likewise.
	* util.cc (cbl_field_t::report_invalid_initial_value): Likewise.

Author: jakubjelinek

gcc/cobol/parse.y

@@ -10336,8 +10336,8 @@ intrinsic:      function_udf
                   if( p != NULL ) {
 		    auto loc = symbol_field_location(field_index(p->field));
                     error_msg(loc, "FUNCTION %qs has "
-                              "inconsistent parameter type %ld (%qs)",
-                              keyword_str($1), (long)(p - args.data()), name_of(p->field) );
+                              "inconsistent parameter type %td (%qs)",
+                              keyword_str($1), p - args.data(), name_of(p->field) );
                     YYERROR;
                   }
                   $$ = is_numeric(args[0].field)?

gcc/cobol/scan_ante.h

@@ -149,7 +149,7 @@ numstr_of( const char string[], radix_t radix = decimal_e ) {
     }
     auto nx = std::count_if(input, p, fisdigit);
     if( 36 < nx ) {
-      error_msg(yylloc, "significand of %s has more than 36 digits (%ld)", input, (long)nx);
+      error_msg(yylloc, "significand of %s has more than 36 digits (%td)", input, nx);
       return NO_CONDITION;
     }
 

gcc/cobol/util.cc

@@ -1049,8 +1049,8 @@ cbl_field_t::report_invalid_initial_value(const YYLTYPE& loc) const {
                                                  return TOUPPER(ch) == 'E';
                                                } );
               if( !has_exponent && data.precision() < pend - p ) {
-                error_msg(loc, "%s cannot represent VALUE %qs exactly (max %c%ld)",
-                          name, data.initial, '.', (long)(pend - p));
+                error_msg(loc, "%s cannot represent VALUE %qs exactly (max %c%td)",
+                          name, data.initial, '.', pend - p);
               }
             }
           }