How is the following?
<p>
<b>Post-finalization note 1</b>: In <code>regexp-replace</code> and
<code>regexp-replace-all</code>, the <code>subst</code> is defined to
be a string, integer or symbol, however the reference implementation
as well as equivalent procedures in IrRegex and SCSH also allow a list
of any of these, with the results concatenated. In addition, the
reference implementation accepts a procedure of one argument, which is
applied to corresponding <code>regexp-match</code> object to obtain
the substituted string (as in Emacs <code>regexp-replace-in-string</code>).
This can be used to avoid the ambiguity whereby <code>'pre</code> and
<code>'post</code> have predefined meanings and can't refer to submatch
names. Implementators are encouraged to support these two extensions.
</p>
Indeed, the existing post-finalization notes have been incorporated by the vote, though I mention the fact in each case for the avoidance of doubt. Currently the R7RS-large set/bag, immutable list, flonum, comparator, and vector libraries all have post-finalization notes incorporated. SRFIs 48 (format strings) and 149 (syntax-rules template extensions) will be balloted. SRFI 97's note is purely editorial in nature.
On 2020-04-09 8:55 a.m., John Cowan wrote:
> I think this is a case for a post-finalization note, which is a
> recommendation to implementers placed in the Status section of the
> SRFI. You can look at SRFI 113 for what they look like if you want to
> draft one. On this set of facts, I'm reasonably sure Alex and Arthur
> would accept it.
My thoughts on post-finalization notes are that when people vote on a
SRFI for inclusion in R7RS-Large, they are explicitly voting for or
against the SRFI as amended, unless the ballot question explicitly rules
out a note.
So while a post-finalization note is a recommendation in the SRFI, it
becomes normative if adopted into R7RS-Large.
-- v