Re: SCM_NEWCELL{2}?

[Maciej Stachowiak <mjs at eazel dot com>]

Mikael Djurfeldt <mdj@mdj.nada.kth.se> writes:

> Dirk Herrmann <dirk@ida.ing.tu-bs.de> writes:
> 
> I buy your arguments for coding "pedantically right", and think we
> could go through with such a change.  This is a big decision, though,
> and we should hear what Maciej and the others have to say first.

I think in general limiting use of SCM_CAR, SCM_SETCAR and such to
actual pairs rather than cells in general is a good idea. It shouldn't
even be that hard a change, because any type-specific macros of that
kind can just be #defines of SCM_CELL_FIRST and SCM_CELL_SECOND (or
whatever those end up being called).

> I also think we should change SCM_SETCAR to SCM_SET_CAR.  But while
> doing all these changes, it's important that we keep track of backward
> compatibility so that we don't break old applications more than
> necessary.  We should, for example, preserve old macros and try to
> make sure that only the "debugging modes" will break compilation on
> violations, at least in the beginning.

Agreed generally; although I am not sure if SCM_SETCAR is something we
want to enourage app authors to use, being copatible should not be
hard in this instance.

 - Maciej