Update of ECL's sysfun.lsp resolves both elliptic integral issues

Juanjo, the main ECL developer has found the bugs in sysfun.lsp causing 
the problems I noticed earlier on a SPARC box. After copying sysfun.lsp 
from CVS, bugs affecting both

* ellipic_e http://sagetrac.org/sage_trac/ticket/6716
* elliptic_eu http://sagetrac.org/sage_trac/ticket/6719

have been resolved. As such, both of these can be considered closed as 
far as I am concerned.

A new ECL .spkg file can be found here.

http://sagetrac.org/sage_trac/ticket/6564

As noted, in the above, it is necessary to update Maxima too, since 
Maxima has a bug which ECL found. The trac to update Maxima is 
http://sagetrac.org/sage_trac/ticket/6699

I'm aware the update of Maxima will cause some doctest failures in Sage, 
but many actually look like improvements in the output from Maxima which 
are causing the failures. For example, the following doctest fails, but 
clearly the output is better that what Sage expected. Things like '?%' 
have been removed in many places.


"/export/home/drkirkby/sage/sage-4.1.1.rc0/devel/sage/doc/en/constructions/interface_issues.rst", 

line 478:
      sage: maxima.eval("f:bessel_y (v, w)")
Expected:
      '?%bessel_y(v,w)'
Got:
      'bessel_y(v,w)'




Dave