On Thu, 19 Jun 2025 13:53:06 +0200
Jakub Jelinek <ja...@redhat.com> wrote:
> On Thu, Jun 19, 2025 at 01:38:06PM +0200, Rainer Orth wrote:
> > --- a/gcc/cobol/genapi.cc
> > +++ b/gcc/cobol/genapi.cc
> > @@ -957,7 +957,7 @@ parser_compile_ecs( const std::vector<ui
> > {
> > SHOW_PARSE_HEADER
> > char ach[64];
> > - snprintf(ach, sizeof(ach), " Size is %ld; retval is %p",
> > + snprintf(ach, sizeof(ach), " Size is %zu; retval is %p",
> > ecs.size(), as_voidp(retval));
> > SHOW_PARSE_TEXT(ach)
> > SHOW_PARSE_END
>
> Please don't add %zu/%zd uses in *printf family functions, not
> all hosts support that unfortunately (I think HPUX among others).
>
> There are HOST_SIZE_T_PRINT_UNSIGNED etc. macros in hwint.h and
> fmt_size_t type the argument should be cast to (in most cases it is
> actually same type as size_t, but not in all).
There may still be errors, but I went to great pains to correct the
message format strings over the last few weeks.
For messages that use the diagnostics framework, %z is used because
it's documented to work in gcc/pretty-print.cc.
For messages that rely on printf(3) and friends from the C standard
library ("standard" should perhaps be in quotes), I'm learning to use
the hwint.h macros, even though they give "ungainly" a bad name. The
things we do for portability to antique proprietary systems.
It is unfortunately not obvious which is which; the function name
yields little clue to which library it relies on. When I need to
remind myself, I just look at the function definition.
Thank you for the report, Rainer. I'll attempt to correct
immediately.
--jkl
