Yup. The spec says "Continuation frames can be annotated with continuation marks." so I naively made each continuation frame have a slot for its marks. Then I realized that adding a mark mutates the continuation frame and all the complexities come out.
Could there be another explanation? I think srfi should be implementable solely from the spec and allowing different models as far as it follows the spec. On Sat, Oct 15, 2022 at 12:28 PM Marc Nieper-Wißkirchen < marc.nie...@gmail.com> wrote: > Am Sa., 15. Okt. 2022 um 14:21 Uhr schrieb Shiro Kawai < > shiro.ka...@gmail.com>: > > > > Is it possible that more than one thread attach continuation marks to > the same continuation simultaneously? > > I couldn't come up a code that does so, but I'm not susre if it's > impossible. For example, can an implementation-level error be triggered > right after a continuation passed from a different thread, and the handler > attaches a new mark? > > Attaching continuation marks does not mutate but, abstractly, conses > to a continuation frame. In principle, it is no different from > attaching frames by non-tail-calling procedures. > > Does this help and answer your question? > > > >
