Re: [Bug-gnubg] How does one set up the database

[Jim Segrave]

On Mon 28 Jun 2004 (01:42 +0200), Jim Segrave wrote:
> On Sun 27 Jun 2004 (14:15 +0000), Joern Thyssen wrote:
> > On Sun, Jun 27, 2004 at 03:59:09PM +0200, Jim Segrave wrote
> > > 
> > > I finally got around to building python support in gnubg (under
> > > FreeBSD 5.2, it finally does threading correctly and doesn't steal all
> > > the idle CPU).
> > > 
> > > A digression before the main question:
> > > 
> > >   I still can't get the autoconfigure chain to work on this system and
> > >   am forced to use autogen.sh.
> > 
> > What goes wrong?
> 
> I'll get a full dump of what happens tomorrow. Basically m4 gets
> unhappy 

The problem is an error message from autoconf regarding
_m4_divert_diversion.

googling on this leads to 
<URL:http://lists.gnu.org/archive/html/autoconf/2001-07/msg00104.html>

It appears my choices are to use autogen.sh or to revert to an older
autoconf/automake. It looks like the long term requirement will be to
rework the autoconf/automake files for gnubg, but that assumes that
updated tool chains are widely akvailable, which I assume is not the
case, as then there'd be a lot more people encountering this.

On the other issue - ftgl and autoconf - looking at the header files
for ftgl, it looks like it entirely a set of C++ class
definitions. Can an configure script simply create a static instance
of one of the classes? I assume that will fail to link if it doesn't
find the library. All of the publics in libftgl.a are C++ names with the
signature prefixes, so there's no obvious trivial way of checking that
the library has been located and linked


-- 
Jim Segrave           address@hidden