Just one editorial correction: "coincides" always needs to be followed with "with". Otherwise LGTM.

However, I spotted a different problem, but I'm too tired to fix it now. I'll try to do so tomorrow, or at worst Friday morning UTC-4 time. Basically, this SRFI still says that binary positions are always exact integers, whereas we changed SRFI 192 to say that they don't have to be.

On Tue, Aug 25, 2020 at 11:19 PM Shiro Kawai <xxxxxx@gmail.com> wrote:

I PR'ed a few minor changes.

https://github.com/scheme-requests-for-implementation/srfi-181/pull/8

- The arity of close and flush procedures weren't specified. Made it clear that they take no arguments.
- Clarify that the "position" returned by get-position conincides the position next read! and/or write! performs the action.
- get-position of make-custom-textual-output-port said an exact integer, which should be an implementation-defined object.

On Sun, Aug 23, 2020 at 6:25 PM Arthur A. Gleckler <xxxxxx@speechcode.com> wrote:
On Sun, Aug 23, 2020 at 8:54 PM Arthur A. Gleckler <xxxxxx@speechcode.com> wrote:

Unless someone requests otherwise, I will publish a new draft with this language in it tomorrow (by pulling from your repo), then wait until Thursday (PT) before finalizing.

I published the new draft without waiting, but I'll wait until Thursday before finalizing.