[racket-dev] Distributing C libraries

From: Eli Barzilay (eli at barzilay.org)
Date: Thu Oct 6 14:51:23 EDT 2011

Yesterday, Neil Toronto wrote:
> On 10/05/2011 03:25 PM, John Clements wrote:
> >
> > On Oct 5, 2011, at 9:45 AM, Neil Toronto wrote:
> >
> >> I've just pushed the new 'plot' library.
> >>
> >> Eli and I both forgot that the new 'plot' still needs the old
> >> libfit, and removed it along with libplplot. I just added libfit
> >> back (the sources are in "src/fit" now instead of "src/plot/fit")
> >> along with the proper configure, Makefile.in, and get-libs.rkt
> >> changes.
> >
> > Tangential question: it looks like your distribution philosophy
> > here is to insist that users have a C compiler installed. Is this
> > correct? So people without a compiler can't use the package? (I
> > realize that if it's true of your package, it's almost certainly
> > true of the earlier one, as well; I'm just curious, because my
> > sound library has the same issue.)
> That would be true if 'plot' were a PLaneT package that includes
> libfit.  But libfit.so is compiled at the same time as the rest of
> Racket and distributed with the other binaries.

For distribution purposes, what Matthew did (which I think makes
sense) is to include binaries for platforms where a compiler is
unlikely to be present.  (=> windows and osx.)

> (Eli wants to get rid of it. So do I. You might have seen the recent
> commit I pushed that comments out check-= 'fit' tests that randomly
> fail. Out of about 50 executions they failed once for me, but DrDr
> seems more unlucky. *sigh*)

Yes, I strongly prefer that it (libfit) gets removed.  It was a
mistake to include it with plot, since it is independent of the actual
thing that plot is supposed to do.

          ((lambda (x) (x x)) (lambda (x) (x x)))          Eli Barzilay:
                    http://barzilay.org/                   Maze is Life!

Posted on the dev mailing list.