Re: [info-sather] Installing on VMWarev

[Sather User]

On Fri, 21 Sep 2012, Duke Normandin wrote:
> > The next version will handle directories, symlinks and stuff via
> > system calls.  Linux oriented, but if it worked on VMware, that would
> > be a plus.
>
> I'll have to get back to you.  I was SO pissed off at the install process,
> that I nearly nuked Sather off my HDD. I'm going to leave it alone for a week,
> and come back to it when I cool down. :D

Okay.  The next version, which I believe is now fairly close to
release, demands you to unpack the source and go: ./configure; make;
su; make install.  That is, on GNU/Linux.  Dunno about VMware.

Hope that doesn't piss you off so much.

So take a week or two to cool down, then check again.

> I was reading some old Sather installation docs, so that may account for some
> of the problems I'm having. L8r ...

Hmm.  I just overstated things.  If you're using GNU Sather version
1.2.3, all you have to do on GNU/Linux is unpack the source, set
SATHER_HOME=$(pwd), export it, unset SATHER_LIBRARY (which won't be
set, but if it was...), and type "make; make test; make testfull" and
you'll have compiled a lot of Sather source code with your new Sather
compiler.

We should probably give credit to Michael Taylor and earlier GNU
Sather maintainers for making it that easy.  One cannot today compile
ICSI Sather-1.2b without several source code changes.

Admittedly, what you get is only the compiler, you then have to
compile the optional browser if you decide to exercise that option,
and the guesses at the locations of Tcl/Tk files needed by the tcltk
platform are inflexible and may be inadequate. And there are a few
other things out of date in the Tk department, and...  But a working
compiler, able to compile Sather source code in general, shouldn't be
too difficult.  If you have difficulties they might be
VMware-specific.

I could never get the parallel/threads stuff to work and I don't think
it's ever worked under GNU/Linux.

Correct me, anyone who knows better.

That's something that won't have been fixed with the next release.

But you shouldn't have a problem with the rest, with serial Sather,
although there are bugs, e.g. INT::is_ult(INT):BOOL (i.e.,
is-less-than for integers assumed to be unsigned) has its logic back
to front.  Just one example.

-- Michael Talbot-Wilson