patch for acinclude.m4.in: Better test for "Qt without flags"

[Michael Ritzert]

Hi Stephan,

On Mon, Aug 04, 2003 at 02:37:47PM +0200, Stephan Kulow wrote:
> On Monday 04 August 2003 14:21, Michael Ritzert wrote:
> > To prevent this kind of error in the future I propose the attached patch to
> > have configure check whether the correct libqt-mt.so was picked up. This patch
> > is mainly intended as food for thought as I have no idea how portable the shell
> > code is. I checked that it work correctly for bash on GNU/Linux in both
> > possible cases (wrong/correct lib linked in).
> Did you test what ldd outputs on non-Debian systems? I wouldn't trust it's
> portability - especially as you name Windows yourself.

I have checked ldd on FreeBSD and HP-UX. Both versions show the full
path to the libraries linked in:
HP-UX: /usr/lib/libc.2 =>      /usr/lib/libc.2
FreeBSD: libc.so.4 => /usr/lib/libc.so.4 (0x283b1000)

Note that HP-UX doesn't have an extenstion for shared libraries. This
however should not be an issue as I used $LIBQT_GLOB in my patch and
that is exactly what is used earlier to find the libraries.

I don't have easy access to Windows, but at least ldd from Cygwin should
be fine.

> The whole test should better be removed in favor of a test if qtlib == /usr/lib,
> which was the reason the check was actually added.

Can you elaborate on that? It seems odd to me to handle one library
different from dozens others. You put that code in back in January 1998.
Maybe it's obsolete now?

Michael