Jon Buller wrote:

> In message <3DC73762.6BC4A7C2@bom.gov.au>, Paul NCC/CS writes:
>
> >I'm still trying to get clisp to build but I'm stumped so
> >could someone please give me some help and just point
> >me in the direction of somewhere where I could help
> >myself please?  I don't know what else to do at the moment.
>
> I tried this a few months ago, also without much luck.



Ohh thanks for letting me know. Thats a help.
I tried to hack a makefile ( see previous email
to this list) and got it going further but then it
core dumped anyway.

But I did manage to get the clisp version from pkgsrc
1.5.2 working. I just had to increase the stacksize
using the shells limit command and it built!:))).

Pity about the more current version though.
Like you say it builds on i386 but not sparc.

cheers,
Paul.







>
>
> >Because pkgsrc failed:
> >I tried to download the raw tarball from "clisp.cons.org"
> >and install it using their instructions. But it bombed in
> >a similar fashion that it does when I try to build it using
> >pkgsrc.    It's basically during the configure script in the
> >raw tarball build.
>
> Yup. It works on i386 (and I think macppc) but not on a sparc.  I
> went a little farther and asked for help on one of the CLISP mail
> lists.  IIRC, it has something to do with some structures, their
> alignment/padding, and the way the CLISP code wants to deal with
> it.
>
> They did not like the fact that "cc --version" reported egcs 1.1.2
> (or whatever the exact version was.)  I got two recommendations:
> 1) get a newer compiler and tell us what that does. and 2) use the
> latest version (2.30?).  Neither were much help at all, building
> a new version of gcc just for that was too much trouble, and the
> latest release wasn't any better than the pkgsrc stuff.  I haven't
> pestered them since, since the attitude seemed to be "We don't care
> about junk like that, if you use a real compiler/machine we probably
> will."
>
> I've since upgraded to 1.6 and had the same problem.  (At least I
> think I did, it's been a few days, and I have too many higher
> priority things right now to give it very many resources.)  When
> I get a little spare time, I plan to tell them their stuff is broken
> with current compilers too, and see what happens.
>
> Executive summary seems to be: 1) The CLISP stuff is broken, 2)
> they don't want to believe it.  3) They may believe it if you use
> a new enough compiler and gripe enough.
>
> Jon
>
> PS When does GCC 3.2 (or newer) get imported so I can try building
> -current (and maybe CLISP too) on my pc532? 8^)

--

____________________________________________

Paul
Australian Bureau Of Meteorology
____________________________________________