Works for me.  In R7RS-large, only SRFI 128 is used, so all R7RS SRFIs should be understood as using SRFI 128.  In addition, Schemes that do support both 114 and 128 should support them compatibly.

On Fri, Mar 27, 2020 at 6:33 PM Shiro Kawai <xxxxxx@gmail.com> wrote:
I'll make a PR.

But as a part of R7RS-large, I'd like to have some agreement, down the road, that a portable code can
rely on those comparator to be treated as srfi-128, because srfi-128 and srfi-114 comparators are not 
necessarily compatible to each other and writing code to consider both possibilities will be cumbersome.

Gauche supports both srfi with integrated comparator so it's no issue.  If an implementation supports
only one of them, it's also no issue.  If an implementation supports both but as disjoint types, it'll be an issue.

So, could we have either mini-srfi or some other means, to agree that a portable R7RS-large code can
assume either (1) the implementation provides only srfi-128, or (2) if the implementation provides both128 and 114, the comparators are compatible to each other?






On Fri, Mar 27, 2020 at 11:51 AM Arthur A. Gleckler <xxxxxx@speechcode.com> wrote:
On Thu, Mar 26, 2020 at 9:05 PM John Cowan <xxxxxx@ccil.org> wrote:
I think a post-finalization note is appropriate:  "The author of this SRFI recommands that implementers used SRFI 128 comparators rather than SRFI 114 comparators."'

Okay, I've added that as a post-finalization note.
 
Adding to the existing implementation would be great.  I'm not sure why I left them out.

Shiro, if you or John or someone else would like to add them to the sample implementation, please do.

Thank you both.