>>However, the possibility is to start a thread in parallel, and if the >>initial call to the VM has not returned after a given time, you can kill >>the VM thread. There is currently no simple way to abort the VM >>execution from another thread, but this is something I'll think about >>adding. > > > And will the GC be alright after such a thing (i.e. not memory leaks)? > IIRC this was the problem with forcing Python threads to stop (currently > there is no clean way to kill a Python thread).
Yes, mostly ok. The only thing that will leak is the field id cache which is a per-thread object. It only store (id/name) objects fields relations so might be a few K at worse. I'll think about a way to perform a proper cleanup. Nicolas -- Neko : One VM to run them all (http://nekovm.org)
