Re: [Fwd: Re: [Axiom-developer] Building Axiom on Slackware]

[root]

> >> Actually, a patch for this problem was posted a year ago (and even long 
> >> before
> >> counting the build-improvements branch).  I was (mistakenly) under the
> >> impression it was applied:
> >>
> >> http://lists.nongnu.org/archive/html/axiom-developer/2008-05/msg00000.html
> >>
> >> On Sun, Apr 12, 2009 at 01:39:45AM +0400, Igor Pashev wrote:
> >> >  # Loader flags
> >> > -LDF=" -L/usr/X11R6/lib ${XLIB}/libXpm.a "
> >> > +LDF=" -L/usr/X11R6/lib ${XLIB}/libXpm.so "
> >>
> >> LDF should likely be set as in the link above so that either the static or
> >> dynamic lib can be used (preference to shared libs IIRC):
> >>
> >>    LDF=" -L${XLIB} -lXpm"
> >>
> >
> > Steven, Igor,
> >
> > I MAY BE WRONG, but as I recall, when I applied that patch it failed
> > to build on several systems.
> >
> > Did either of you test it? How? On what system does it work?
> 
> As Stephen mentioned, a patch was applied to
> build-improvements -- picked up by both OpenAxiom
> and FriCAS.  And as far as I can tell, it works on the many
> platforms those systems build.  Solution?  Don't hardcode
> XLIB.  Get it deduced for you for free by tools that are
> specifically designed to solve that kind of problem.
> Stephen';s suggestion is correct for 99.9% of the platforms.
> 

Gaby,

Sigh.

I await your diff-Naur patch for autoconf builds of the Axiom trunk.
If you recall, I've asked for this since you first did autoconf.
That never happened. Please include the patch with your next autoconf
advertisement.

You'll admit that autoconf would have failed if libXpm was not in the
list of things to check. Exactly this kind of failure happpened with
egrep last week. FriCAS, as I recall, did NOT use autoconf to fix it.



Do we really need to relive the whole debate? I certainly don't. 
If people think autoconf is vital, there are now alternatives.
It seems like a "solved problem" to me. 



Can we please stick to the simple question at hand, which is,
did anyone actually TEST the patch and confirm that it works on slackware?
It seems like a standard of public programming to actually test a patch.

And, if the fix is so trivial, did anyone consider simply (re-)posting the
patch rather than open a diatribe about how Axiom fails everywhere?
I don't remember a public crucifiction of FriCAS for the egrep bug
despite the fact that autoconf can detect it but did not.

All systems have bugs.
Find it, fix it, TEST it, post the fix, move on.

Tim