Re: [patch/minor] fix compile warning in linux-proc.c

[Andrew Cagney <ac131313 at redhat dot com>]

> On Wed, 20 Aug 2003 22:11:49 -0700, Randolph Chung <randolph@tausq.org> said:
>
>
> > While trying to get hppa-linux building with -Werror, I came across one
> > warning in linux-proc.c that's triggered by -Wformat-nonliteral. This 
> > patch fixes it if anybody cares :)
>
>
> I'm seeing this on i686-pc-linux-gnu, too.  The patch looks fine to me
> (though I'm not in a position to approve it); if I were in a
> hyper-refactoring mode, I suppose I'd be tempted to create new
> functions print_header_format_32, print_data_fmt_32,
> print_header_format_64, print_data_fmt_64 defined along these lines:

David, feel free to check it in.  A comment mentioning that this should 
use a generic (and not yet existing) local_address_string() method would 
be useful.

The casts don't phase me.  The conflicting "%#10lx" and "%#18lx" do but 
again not desperatly.  That local_address_string custom method.

Andrew


> static void
> print_data_fmt_64 (unsigned long addr, unsigned long endaddr,
>                    int size, unsigned int offset, const char *filename) {
>   printf_filtered ("  %#181x %#181x %#10x %#10x %7s\n",
>                    addr, endaddr, size, offset, filename);
> }
>
> but I don't think it's really urgent to do so.  Hmm; maybe we really
> should do that, at least for the data strings, because it gets rid of
> casts?  Andrew, what do you think?