Re: [PATCH] %nil-handling optimization and fixes v1

[Neil Jerram]

Mark H Weaver <address@hidden> writes:

> I agree that the names are uncomfortably long.  We could shorten them
> without much loss of clarity by replacing "lisp_nil" with "nil" and
> "and_not" with "not", yielding:
>
>   scm_is_false_assume_not_nil  scm_is_true_assume_not_nil
>   scm_is_false_not_nil         scm_is_true_or_nil
>   scm_is_false_or_nil          scm_is_true_not_nil
>
>   scm_is_lisp_false            scm_is_lisp_true
>
>   scm_is_null_assume_not_nil
>   scm_is_null_not_nil
>   scm_is_null_or_nil
>
>   scm_is_bool_not_nil
>   scm_is_bool_or_nil
>
> I can still do this if y'all would prefer the shorter names.

FWIW, dropping "lisp_" looks OK, but I'm not sure about dropping
"and_".  "scm_is_false_not_nil" feels notably harder to understand
than "scm_is_false_and_not_nil".

> Are there any remaining objections to mapping scm_is_false/true/null
> as follows?
>
>   scm_is_null   -->  scm_is_null_or_lisp_nil
>   scm_is_false  -->  scm_is_false_or_lisp_nil
>   scm_is_true   -->  scm_is_true_and_not_lisp_nil

(Not from me - but I assume you're checking with others...)

    Neil