On Fri, Nov 07, 2008, Scott Brozell wrote:
>
> On Thu, 9 Oct 2008, David A. Case wrote:
>
> > On Thu, Oct 09, 2008, Roberto Gomperts wrote:
> > >
> > > Immediately I encounter several problems, starting with how configure_at
> > > generates config.h for irix_cc. For example it defines a "ranlib"
> > > function not present in irix.
> >
> > will fix.
>
>
> I didnt see a fix; so i added one while cleaning up the irix_cc
> section. see configure_at revision 1.67
I didn't post a fix because I couldn't find the problem. I don't understand
why revision 1.67 fixes anything related to ranlib. Roberto did not really
say what the problem was, so I gave up.
> sgi's f77 has worked
OK, we can add that to the list of compilers that seem to work. I've hoped
that we could remove all fortran code from AmberTools, and hence all need to
specify which fortran code to use...
>
> > > reduce/reduce_src/Makefile has a bug in defining the OBJLIST:
This was fixed a while ago--Roberto had been testing an outdated AmberTools
(if I remember correctly).
> > > cc-1028 cc: ERROR File = lex.dg_options.c, Line = 45
> > > The expression used must have a constant value.
> > >
> > > FILE *dgoin = {stdin}, *dgoout = {nabout};
> >
>
> It may be premature since Ive asked for two bug clarifications
> (and aside from bug 93), but all the bugs above seem to be programmer errors
> not directly OS, make, lex or compiler issues.
This is perhaps not a lex error, per se, but a failure of portability -- we
are assuming that lex code can be handled in a way that doesn't work for SGI's
lex. The only simple solution I can see is to require flex, even for SGI.
...dac
Received on Fri Dec 05 2008 - 11:21:05 PST
