Re: [AMBER-Developers] Multiple definition of cifpin in cifparse

From: David A Case via AMBER-Developers <amber-developers.ambermd.org>
Date: Tue, 28 Jun 2022 21:39:12 -0400

On Thu, Jun 23, 2022, Åke Sandgren via AMBER-Developers wrote:
>
>Since cifparse.y and cifparse.l uses %option prefix="cifp", yyin is
>defined as cifpin.
>This causes a multiply defined cifpin from cifparse.c and lex.cif.c

Can you say what code version you are using? cifparse.y has used
'%name-prefix="cifp"' for some time now.

>I guess that the correct solution would be to change cifparse.h to
>always do extern FILE *cifpin ?

I don't think so. The first step is to try to figure out why you are seeing
these problems and (as far as I can tell) others are not. What is your OS
and compiler and flex/bison versions?

[Aside: I don't think that CIF_GLOBAL is actually defined anywhere; if so,
then "extern FILE *cifpin" is already present in cifparse.h.]

You can see the OS and compiler combinations we have tested here:

   https://gitlab.ambermd.org/amber/amber/-/wikis/Singularity-Amber22

Actually, a lot of effort has gone into trying to make cifparse work in a
variety of environments. So seeing the exact error messages would also
be helpful. cifparse.h hasn't changed in a long time, and the original code
comes from John Westbrook and the PDB, so I'm a bit reluctant to make
changes. But we can try to to come up with workarounds for corner cases.

....thx...dac


_______________________________________________
AMBER-Developers mailing list
AMBER-Developers.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber-developers
Received on Thu Aug 04 2022 - 13:34:17 PDT
Custom Search