On 11/3/11 3:09 PM, Jeroen Demeyer wrote:
On 2011-11-03 22:24, William Stein wrote:
Also, any longterm strategy for getting Sage to really
work on Windows will require getting rid of any use of pexpect
Good riddance! Why not replace pexpect by somthing much simpler? I
have never believed that we really need pseudo-terminals for what we
want to do. We only need to start a program, monitor its stdout/stderr
and send stuff to its stdin. Which terminal features do we need?
That's not what I meant at all, though if you have some magic way to
control arbitrary command line subprocesses without using a
pseudo-terminal, I'd love to see it.
What I mean (I mentioned libGAP and libECL) is to completely get rid of
any use of separate subprocesses in most routine use of Sage, and
instead use C library interfaces. The best work in this direction
has been by Nils Bruin et al. on libECL, and also some work by John
Cremona and Martin Raum on getting rid of using gp via pexpect, and also
work of Volker Braun on libGAP (which hasn't actually gotten into Sage
at all yet unfortunately).
This is not to say that we shouldn't continue to support being able to
control other programs (like scilab) from within Sage. However, we
shouldn't rely on separate processes for core functionality (e.g.,
computing the order of a group).
-- William
--
To post to this group, send an email to sage-devel@googlegroups.com
To unsubscribe from this group, send an email to
sage-devel+unsubscr...@googlegroups.com
For more options, visit this group at http://groups.google.com/group/sage-devel
URL: http://www.sagemath.org
