Re: Language translation proposal 1.4

[Mikael Djurfeldt <mdj at mdj dot nada dot kth dot se>]

Michael Livshin <mlivshin@bigfoot.com> writes:

> Ian Bicking <ianb@colorstudy.com> writes:
> 
> > Should load also potentially take a :language argument?  Maybe this is
> > getting to be too many ways...
> 
> seems like a good idea to me, FWIW.

Well, the arguments to `load' are already specified by R5RS.  This
doesn't necessarily mean that we can't provide a Guile version of
load.  (The r5rs-module still exports standard `load' of course.)

This has been done with `set!' in the core module, and `map',
`map-in-order', `for-each', `member' and `assoc' in SRFI-1.

But, I think we should keep this kind of shadowing to a minimum.  The
primitives above have strong reasons for being non-standard.  `set!'
is needed for the object system.  The SRFI-1 primitives are very good
suggestions which probably are intended as candidates for next version
of the Scheme standard.

A better alternative could be to introduce a new primitive:

  load-using-language FILENAME LANGNAME

> > In Python, at least, it is important that whitespace at the beginning
> > of the line is preserved.  So it would be fine to check if the first
> > non-whitespace character is a comma, but the whitespace needs to be
> > preserved.
> 
> perhaps we could mandate that there be _no_ whitespace before the
> comma for it to be special?

That seems OK to me.  A possible counterargument could be that it a
fragile syntax, but if no-one protests, I'll put this into the
proposal.