Re: [MIT-Scheme-devel] multi-threading problem: Unassigned variable: root-continuation-default

[Matt Birkholz]

> From: Matt Birkholz <address@hidden>
> Date: Fri, 4 May 2012 12:14:34 -0700
> 
> [...] but I don't doubt you can get yourself in another fix if you
> keep on like THAT!

I don't mean to discourage you from hacking -- just the opposite!  You
might have found yourself at the bottom of a dark pit, but it was not
all chaos.  The continuations, their control points and dynamic
states, look to be working correctly, though perhaps surprisingly.
And an interrupt reset the console-thread.  A surprise pitfall or two
is more than half the fun, ya?

The patch boarding up this pit is a kludge that enforces the
expectation that block-on-io-descriptor is not unwound except by the
thread holding the io registration.  There might be situations where
such intimate cooperation among threads is intended, where I have
frustrated the intuitions of a different user, so I went no further
than block-on-io-descriptor.

Caveat within-continuation emptor.