Re: 6074 exitfail "conflicting types"

[Nathan]

Attached are the stddef* files.

Some things to note:

1) ./gnulib/lib/stddef.h exists at checkout

2) after "autogen.sh" and "configure", ./gnulib/lib/stddef.h continues to exist with "make". But "make distclean" will wipe out ./gnulib/lib/stddef.h

3) after "autogen.sh" and "configure", "make -C gnulib/lib stddef.h"

    make: Entering directory '/home/user/Desktop/texinfo/gnulib/lib'

    rm -f stddef.h

    make: Leaving directory '/home/user/Desktop/texinfo/gnulib/lib'

4) $ printenv | grep GNU

    GNULIB_SRCDIR=/home/user/Desktop/gnulib/

    GNULIB_TOOL=/home/user/Desktop/gnulib/gnulib-tool

Gavin, if you would like to remote in, you may. I am just running in a temporary virtualbox snapshot.

On Thu, Jan 29, 2015 at 6:17 PM, Gavin Smith <address@hidden> wrote:

Could you post the contents of "gnulib/lib/stddef.h"?

Looking at "gnulib/lib/stddef.in.h", the definition of "max_align_t"
is conditional on the autoconf output variable "HAVE_MAX_ALIGN_T". On
line 17143 on the "config.log" file you attached, I see
"HAVE_MAX_ALIGN_T='1'". Hence I don't understand why the definition
should be visible.

If "gnulib/lib/stddef.h" looks wrong you could try running "make -C
gnulib/lib stddef.h". Because you have GL_GENERATE_STDDEF_H_TRUE being
'#' the rule in gnulib/lib/Makefile.am should simply delete the file.
In fact I don't know why the file exists in the first place.

I checked on the bug-gnulib mailing list and the check for
"max_align_t" was only added in December.
(http://lists.gnu.org/archive/html/bug-gnulib/2014-12/msg00170.html).

stddef_headers.tgz
Description: GNU Zip compressed data