On Wed, Jun 10, 2015 at 06:59:54PM +0200, Rafael Sadowski wrote:
> On Mon Jun 08, 2015 at 06:19:55PM +0200, Ingo Feinerer wrote:
> > On Thu, May 28, 2015 at 05:46:12PM +0200, Ingo Feinerer wrote:
> > > Update of math/maxima from 5.29.1 to 5.36.1:
> > >
> > > - patches/patch-src_init-cl_lisp is no longer needed as already included
> > > upstream.
> > > - patches/patch-share_Makefile_in and patches/patch-src_Makefile_in
> > > removed as maxima builds, installs, and runs fine for me without these
> > > two as well.
> >
> > Based on the received feedback an updated diff:
> >
> > - Take maintainership
> > - Minor rearrangements to follow the order as shown in Makefile.template
> > - Consistent whitespace ("RUN_DEPENDS =" vs. "RUN_DEPENDS=")
> > - Consistent lower/upper case (Yes vs. yes)
> > - Do not ship abs_integrate.info, kovacicODE.info, drawutils.info, and
> > logic.info as there are other help formats available and install-info
> > has problems with them ("install-info: warning: no entries found for")
> > - pre-configure is no longer necessary
> > - make test:
> >
> > 1 test failed out of 10,185 total tests.
> >
> > The single failing test is a rounding error / different notation,
> > e.g., 12.344321123443212e-5 vs. 1.234432112344321E-4
> >
>
> hmm, less tests but no one fail in the current 5.29.1 version:
>
> No unexpected errors found out of 9,519 tests.
> real time : 519.432 secs
> run time : 514.450 secs
> gc count : 1651 times
> consed : 27747929638208 bytes
>
> Which one fails?
It is "Problem 50" in "rtest16". It is a known issue
(https://bugzilla.redhat.com/show_bug.cgi?id=1217814), the problem is on
ecl's side (but appears to be already fixed upstream):
https://www.mail-archive.com/ecls-list@lists.sourceforge.net/msg03213.html
So the maxima update should be fine.
Best regards,
Ingo
