[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: no constants please



Richard Kelsey <kelsey@xxxxxxx> writes:

>    It's like you blow up a building and then complain that there's too
>    much dust to accurately say anything has been damaged at all.  "Show
>    me the damaged part of the building."  I can't, the building is gone.
> 
> You said 'easy' and 'a minor change'.  Now you say that you
> can't tell me what that change is?

"Show me the damaged part of the building."  

>    SCHEME_FALSE should not be assumed to be a C constant;
> 
> The SRFI doesn't do this.  All it says is that evaluating
> SCHEME_FALSE won't cause a GC.  

This is not acceptible.  Sorry, but no.  

> Even if it could cause a GC, it is easy for the user to arrange
> things so that it won't.  If it's easy for the user, why shouldn't
> the implementation do it and save the user the trouble?

Because the user might not care!  And because it might be expensive!
Seems entirely reasonable for me to require the user to declare when
he wants to block all GC, rather than to just gratuitously do it
sometimes rather than others.  Note that this is a *pervasive* problem
in the SRFI, and it's easily fixed!

Thomas