On Apr 28, 9:06 pm, Gonzalo Tornaria <torna...@math.utexas.edu> wrote:
> On Wed, Apr 29, 2009 at 12:41 AM, mabshoff <mabsh...@googlemail.com> wrote:
> >> 5. Switch from Clisp to ECL:
> >> http://trac.sagemath.org/sage_trac/ticket/5929
Hi,
> > Yeah, the main issue is not so much getting ecl and then Maxima to
> > build, the main problem will be dealing with potential pexpect issues
> > that might crop up. The pynac switch will greatly lessen the use of
> > Maxima, so that ought to help a lot :)
>
> On the contrary, I'd argue that the pynac switch will greatly lessen
> the use of Maxima, making it much harder to catch bugs with the
> pexpect interface to maxima which may potentially appear with the
> switch from Clisp to ECL.
Well, I don't think the Maxima interface will ever reliably work.
> See the reason why we were able to track down the bug with
> pexpect+maxima/clisp was because on particular hardware it caused a
> failure 2 or 3 times each run of the testsuite... If sage were using
> pynac at the time, that number could had been one failure every 10,
> 20, or more runs of the testsuite. We would still experience lock ups,
> just less often and more mysterious...
You can always run the relevant doctests that still use Maxima in a
loop. The interface is not going away anytime soon.
> Is it possible to switch to ECL in one release, have this new
> arrangement for maxima thorougly tested in real life, and switch to
> pynac in a later release?
Not really if we want Solaris and the pynac switch in 4.0 - which is
the case - we need ecl since clisp does not work with any even
remotely current gcc release. What I can see is to switch to ecl and
cut an alpha before switching to pynac in the next release.
> Gonzalo
Cheers,
Michael
--~--~---------~--~----~------------~-------~--~----~
To post to this group, send email to sage-devel@googlegroups.com
To unsubscribe from this group, send email to
sage-devel-unsubscr...@googlegroups.com
For more options, visit this group at http://groups.google.com/group/sage-devel
URLs: http://www.sagemath.org
-~----------~----~----~----~------~----~------~--~---
