On Wed, 2006-12-06 at 21:40 -0500, sen1 at math.msu.edu wrote:
> So I did a
> (%i2) stringout("funs",functions);
> (%o2) /home/maxie/sen/funs
> (%i3) sys("grep disc funs");
I guess you mean system("grep disc funs") ?
I do not have any sys function in Maxima 5.10.0cvs
> 3. In xmaxima started from the console, the output of the "sys"
> command goes to the console.
> All one gets in the tk window is "0"
The behaviour of system is identical in Maxima and Xmaxima, and as
expected from the documentation:
"Executes <command> as a separate process. The command is passed
to the default shell for execution."
Both in maxima and xmaxima, the result of system is "0".
In this case <command> is "grep disc funs" which will send some
strings to stdout. Maxima also sends its results
to stdout, where they get mixed up with the results of grep.
> This is fine for me, but, I suppose, is somewhat
> against the philosophy of xmaxima.
xmaxima shows in its text window the responses sent back from
Maxima. In this case what Maxima sends back is just "0". If
we wanted to make Xmaxima show the results of <command>, I
think the right way to go would be to change Maxima making
it capture the result from <command>.
> Wouldn't it be worthwhile to write a utility which would implement
> the "grep" utility (or similar such OS commands) into maxima itself?
Implementing several unix commands separately does not seem the best
strategy. It would be better to implement:
system(<command>,<arguments>)
which captured the output and sent it back to Maxima, so you could
simply write system("grep sync", functions). You could then also run
any Fortran code that you had. This might be related to the recent
FFI discussion (CFFI does not work for any Lisp variant).
Regards,
Jaime