Leo Butler wrote:
>
> < And most of the newly expanded options aren't defined. I guess that
> < clause is meant to handle the case of -abc where -a, -b, and -c are
> < options. But this doesn't work well if the value of an option looks
> < like an option. Expand-args seems too greedy.
> <
> < Can we live without being able to recognize -abc as three separate
> < options all smashed into one?
> <
> < Or maybe we need a new command line parser? (I'd rather not write a
> < new one.)
>
> Getopt_long has been ported to cl by K Rosenberg.
>
That sounds interesting. However, it will have to wait until after Aug 1
when the 5.19 branch is started. In fact I think this work will wait
until after Aug 1.
>
> Alternatively, why not make -o greedy? That is, everything that occurs
> after -o is passed to the lisp.
>
That would be incorrect. We don't want -o to gobble up everything
because there's no expectation that -o must be the last option.
And besides, given the current scheme, I don't even know how to do that. :-)
> <
> < Also, there seems to be a limitation on the number of command line
> < args that can be handled. In maxima, we gather up args 1-9 and pass
> < them to the underlying lisp. But what if there were more?
>
> I think this is a bug in the shell script. It looks like the author is
> trying to avoid arrays in bash, but I have no idea why.
>
It's not necessarily a bash script, and definitely isn't on Solaris.
Ray