A_TE_HALT was being filtered incorrectly
- which means transitions could still be active during the join process
- which means resource actions could be overwritten (lost) during the join
- which means resources could be active more than once
- which is the bug I tried to *remove* yesterday
We can detect this with the R_IN_TRANSITION register flag
- Complain loudy if it's set during a join (so we're sure its fixed)
CVS patchset: 8789
CVS date: 2006/02/23 09:36:01
Mercurial revision: a1205f1a2dd76cd4a7f22be80e091e246f245c46