[PEAK] Re: trellis.Set.discard

[Phillip J. Eby]

At 10:06 PM 10/11/2008 +0300, Sergey Schetinin wrote:
>I was investigating exactly what is happening and stumbled over
>something odd. When I add some debug prints, the version with
>@atomically starts to print "False" as well. The minimal change to
>trigger this is to add str(self.queues) to Controller.schedule in
>stm.py. I imagine there's something in Cell.__repr__ that changes the
>behavior, but don't yet understand what is it. It doesn't seem to
>force the rule to run or anything like that.

It's probably actually the __repr__ of a Set or List or some such, 
called recursivel from a cell's __repr__; those would access data, 
causing a read to occur at an inopportune moment.  Whether that has 
anything to do with anything else, I couldn't say.  ;-)