Re: [PATCH] gettext: do not assume '#define ... defined ...' behavior

[Bruno Haible]

Hi Paul,

> Here's a proposed patch to fix a gnulib portability bug uncovered
> with gzip 1.5.

Thanks, the patch is good. Actually when writing this code, I wondered
whether I could use '#define FOO ... defined BAR ...'. Although ISO C 99
6.10.1.(3) explicitly says that it's undefined behaviour:
  "If the token defined is generated as a result of this replacement
   process or ..., the behavior is undefined."
I tried it and was surprised to see that all compilers support it.

> +     Problem reported for gzip by Steven M. Schweda in
> +     <http://lists.gnu.org/archive/html/bug-gzip/2012-07/msg00000.html>.

Note the VMS compiler output:

  #if !_LIBGETTEXT_HAVE_VARIABLE_SIZE_ARRAYS
  .....^
  %CC-I-EXPANDEDDEFINED, Macro expansion includes the token "defined", which 
will
  be treated as an operator.  This might not be portable.

1) this is an informational message.
2) it says that the VMS compiler is treating the token 'defined' like all
   other compilers do.

> +   the C standard says
> +     it's not portable to expect that this works after macro expansion.

But real life shows that this construct _is_ portable. It's just that it
makes an assumption that the C standard does not guarantee.

Bruno