Re: [Caml-list] cmdliner difficulties

[Gabriel Scherer <gabriel.scherer@gmail.com>]

I would write a dedicated parser (typically using (ocamllex or sedlex)
and Menhir), and I don't think this is a problematic approach at all.
It would also turn out to be very useful if you later decided to also
be able to read these set expressions from other inputs, such as
configuration files.
On Fri, Jul 27, 2018 at 8:11 PM Kenneth Adam Miller
<kennethadammiller@gmail.com> wrote:
>
> Hello,
>
>
> I have a task where I need to have command liner allow a sequence of various different types compose a certain way without having to write a small language for what belongs on the command line.
>
> Basically, I am trying to color a graph with different operations on the command line in such a way that different colors can be specified as the target of results of set operations against names of sets. For example:
>
> --color Red=(Feature1 MINUS Feature2)
>
> There are a predefined set of colors and features, and some operations that can be performed. However, because it's a grammar just like any language, there might be arbitrarily deep expressions, and because the command line won't have me constructing this in a straight forward way without writing a dedicated type parser, which might not be so clean.
>
> There are operations: difference, union, intersection, and cut. These can recur, and there are predefined list of colors that can be the targets of these results, which are calculated from sources that are features. In addition, the specification of colorings can be multiple, as in --color ... --color ...
>
> Can anybody see where they would overcome some of these difficulties?

-- 
Caml-list mailing list.  Subscription management and archives:
https://sympa.inria.fr/sympa/arc/caml-list
Beginner's list: http://groups.yahoo.com/group/ocaml_beginners
Bug reports: http://caml.inria.fr/bin/caml-bugs