On Mon Apr 25 06:10:50 2016, sml...@gmail.com wrote:
> The following works fine:
>
> for (^100).race(batch=>1) { {1 + $^a}(10) }
>
> But the following pretty reliably crashes:
>
> for (^100).race(batch=>1) { {1 + $^a}.assuming(10)() }
>
> If even crashes when EVAL is used to make sure each thread has a
> completely different closure:
>
> for (^100).race(batch=>1) { EVAL(q[ ({1 + $^a}) ]).assuming(10)() }
>
I believe `assuming` uses `EVAL` under the hood, and that it's `EVAL` having
reentrancy issues that is at the heart of it. So adding an extra `EVAL` is only
making it worse. (And yes, this is fairly high on my list of stuff to hunt down
and fix.)
/jnthn
