Re: TclX loop updated

Evil Son wrote:
 I've also updated the test with 8.6a2 (which I presume is unoptimised)

We know there are some significant issues with 8.6a2 (my testing
indicates that 8.5 goes around 50% faster, at least for some kinds of
tight loop) as it is using a substantially different execution model
internally. On the other hand, the new model brings quite a lot of
benefits, notably including coroutines. Guess we need to do a lot more
optimization. :-\

Donal (no, the location of the problem isn't very obvious when I
profile things).
.