From: Brandon Casey <drafnel@gmail.com>
Date: Wed, 22 Apr 2020 19:05:57 GMT
Subject: Re: [PATCH] config.mak.uname: Define FREAD_READS_DIRECTORIES for GNU/Hurd
Message-ID: <CA+sFfMdudQ-cWbXajQ_92LZLOOOnTxmQmmsUo3dY1eKCL5-Y6Q@mail.gmail.com>
In-Reply-To: <452EC7D1-BE96-4172-8F70-3E7384C87280@jrtc27.com>

On Wed, Apr 22, 2020 at 11:50 AM Jessica Clarke <jrtc27@jrtc27.com> wrote:
>
> On 22 Apr 2020, at 19:48, Brandon Casey <drafnel@gmail.com> wrote:

> > introduced
> > this feature that would make fopen() fail when opening a directory for
> > use on the platforms where fread() of a directory did not fail,
> > instead of trying to wrap fread().
>
> Then the current autoconf test is wrong and likely causing confusion:
>
> > AC_RUN_IFELSE(
> >         [AC_LANG_PROGRAM([AC_INCLUDES_DEFAULT],
> >                 [[
> >                 FILE *f = fopen(".", "r");
> >                 return f != NULL;]])],
> >         [ac_cv_fread_reads_directories=no],
> >         [ac_cv_fread_reads_directories=yes])
> > ])

Yes, we should attempt to call fread() there. If either the fopen()
fails or the fread() fails, then that should mean that
FREAD_READS_DIRECTORIES is not necessary.

-Brandon