14:48:20nyef``beach: Looked over clordane.pdf, your approach looks workable, but is still somewhat heavyweight for non-debugged threads, which could affect which levels of debug optimization you use it for.
14:50:16beachHello nyef``. Thanks! You mean you looked at the appendix?
14:57:03beachAre you saying that it could be doable?
14:57:36nyef``It'd be a bit of a lashup, and there's a chance of corrupting the running SBCL image, but it could be done.
14:58:40nyef``And that's without using ptrace(). If you're willing to work with a process boundary, things become a bit easier in some ways, but possibly harder in others.
14:59:28beachI have no intention of going outside the process boundary. :)
15:02:18nyef``Somewhere or other you have something on putting "unsafe" things in separate address spaces, so that they can't corrupt the host environment. This is _exactly_ that use-case.
15:03:53beachHmm, OK. That's in the LispOS document.