On Mon, 2009-06-01 at 13:16 +0300, Abdulaziz Ghuloum wrote: 
> [Let's call import-ikarus--print-greeting.sps main.sps.]
> 
> What about "ikarus foo bar baz" where foo, bar, and baz are nonoptions?
> 
> Should that be equivalent to
>    ikarus --repl --load foo bar baz   (current behavior)

No.  Eliminating the current behavior, as you suggested, is a good idea.

> or
>    ikarus --r6rs-script foo bar baz   (current scheme-script behavior)

I don't like this because it assumes --r6rs-script is the desired mode.
If this assumption becomes established (e.g. by shell script writers), I
foresee problems when --future-whatever supersedes --r6rs-script.
If/when that superseding happens, I don't think --r6rs-script should
remain the implicit mode, but there may be things which rely on the
assumption that it is.  Seems better to avoid this altogether by
requiring the mode to be explicit.

> or
>    ikarus --repl --r6rs-script main.sps foo bar baz
>                                       (consistent with proposed  
> behavior)

I suppose this is the right one because --r6rs-script is not actually
assumed, REPL mode is.  When --future-whatever supersedes --r6rs-script,
this will still be the same.

And I suppose this should be an error:
  ikarus --not-for-ikarus

And
  ikarus -- <anything>*
should be equivalent to
  ikarus --repl --r6rs-script main.sps <anything>*
so that option-appearing arguments can be passed through.

> or
>    ikarus --repl --r6rs-script foo bar baz
>                                       (more useful, foo replaces  
> main.sps)

Same issue about assuming --r6rs-script and risking establishing
assumptions which may become invalid.

-- 
: Derick
----------------------------------------------------------------

Reply via email to