Re: [Maxima-discuss] Maxima release in sight?

[Robert D. <rob...@gm...>]

On Thu, Mar 31, 2016 at 3:27 PM, Viktor T. Toth <vt...@vt...> wrote:

> Ah, okay, that makes good sense. I wanted to test all the steps
> before messing with the actual repository, and this was going to
> be the last checkpoint, to make sure that I can actually run rpmbuild.

Well, bear in mind that any operation affects only your own repo
(i.e. anything before git push) can be undone. You can create and
destroy branches, tags, commits, whatever.

> I note that in maxima.spec, we have enable_clisp set to one,
> but enable_ccl is set to 0. Nonetheless, rpmbuild failed on my
> system with no CCL installed. I remedied this by installing CCL,
> but why was that even necessary, do you know?

I'm guessing it's due to this line in maxima.spec:

%define ccl_flags --disable-ccl --with-ccl=lx86cl

Probably if --disable-ccl is present, --with-ccl should be
omitted. If you're interested maybe you or someone can
edit maxima.spec.in accordingly.

> And why is the CLISP build the default? Isn't GCL better,

Well, GCL has IMHO more bugs than other implementations,
and I just didn't want to deal with it. I seem to recall that I tried
to change the default from Clisp to something else at one point,
and ran into some kind of idiosyncrasy. But if you'd like to
revisit the choice of the default Lisp, please go ahead.

> Lastly, is there a reason why ECL is not at all present in maxima.spec?

No good reason, just the usual laziness / disorganization / etc.
Please feel free to add it to maxima.spec.in.

> (For what it's worth, when I am testing, I usually try to build
> Maxima with all the LISPs I have installed, because occasionally,
> certain errors are only triggered in certain LISP environments.)

Yes, that's a good idea.

Thanks again,

Robert Dodier