BINGO! was: aldor fricas connection - which ao files should be generated?

Martin Rubey

2008-08-01 06:47:34 UTC

IMPORTANT UPDATE: the error message complaining about missing MODULE.ao was
*my* fault! "make clean" does not remove src/libcombinatax.a*, and that
caused the problem.

Interesting that you think that FriCAS's "make clean" should remove
libcombinatax.a*. It doesn't know about the combinat files.
And if you actually meant "make clean" of aldor-combinat, then say so. You
are mathematician, so be precise!

I'm a mathematician, and I'm imprecise and chaotic. I know mathematicians of
all sorts... I meant make clean of aldor-combinat.

Post by Ralf Hemmecke
AC's "make clean" does not remove the lib*.* files, because they cost time to
build them. Currently use "make distclean" to remove any file that was
generated after you unpacked an aldor-combinat distribution. It will
basically leave you with the files that are in the SVN repository.

OK. I will probably never remember how Makefiles and their traditional targets
work. I'm afraid that you told me exactly this already once, sorry about that.

Post by Ralf Hemmecke

I can now compile aldor-combinat at least with fricas rev. 11, and very very
likely also with the current version. I'll check this tomorrow. Also the test
(as on MathAction) passes now!

Ooops, could you please, please tell me whether AC actually already runs in
FriCAS on your computer? I am currently looking at Peter's Makefile and it
seems there is some runtime stuff not yet compiled in
branches/aldor-interface.

Yes, it does (well, FriCAS rev. 11). But note that I was using the very old
aldor-interface (src_aldor2.tgz). Still, I think that this is very good,
because now I have something I can compare your new interface with.

Post by Ralf Hemmecke

--OBSOLETE--------------------------------------------------------------------
I just looked into Ralf's new set of Makefiles and programs, to find out which
ao files are actually generated. Unfortunately, I did not understand.
It seems that for each clique, exactly one file is generated, namely the one
which happens to be first in the list. For example, from the clique
Clq(15-[ABELGRP,ALGEBRA,BMODULE,COMRING,ENTIRER,INTDOM,LMODULE,MODULE,RING,RMODULE,RNG,init_INT])
ABELGRP.ao is generated, the others seem to be ignored. Or am I mistaken here?

Yes, you are wrong. The exports of *all* the domains in the clique *are* in the
corresponding .ao file. Just look into src/aldor/ap/APELGRP.ap.

OK. Maybe it would make sense to add a note to the documentation (I'm not sure
where though), or rename the files so that it's apparent that the contain the
whole clique.

Maybe:
Index: libaxiom.tex
===================================================================
--- libaxiom.tex (revision 322)
+++ libaxiom.tex (working copy)
@@ -260,6 +260,10 @@

\item Call \file{cliques} on \file{libaxiom.lst} and all the
dependency files and thus compute \file{cliques.mk}.
+
+\item For each clique produce an \texttt{.ao} and then an \texttt{.ap} file.
+ For example, the file \file{ABELGRP.ap} could contain code for
+ \texttt{ABELGRP}, \texttt{ALGEBRA}, \texttt{BMODULE}, etc.
\end{enumerate}

-------------------------------------------------------------------------
This SF.Net email is sponsored by the Moblin Your Move Developer's challenge
Build the coolest Linux based applications with Moblin SDK & win great prizes
Grand prize is a trip for two to an Open Source event anywhere in the world
http://moblin-contest.org/redirect.php?banner_id=100&url=/