[pcwalton]

To be fair to Nim, I don't see any reason why it couldn't be made memory safe by using the Boehm GC (though I'm not an expert in Nim by any means). Of course, using the Boehm GC negates the advantages of the thread-local heaps, but I don't think that Nim's implementation of them scales up to large-scale software in any case for the reasons I detailed in my other comments. IMHO, if you have a garbage-collected, multithreaded language that must compile to C (and doesn't need interoperability with a reference-counted object system like e.g. Swift does), the Boehm GC is the best choice.

[parley]

Thanks for correcting me, Patrick. I certainly didn't mean to be unfair (especially as I replied to a comment I felt wasn't being completely fair itself, intentionally or not), and I should have been more precise about the use cases.

I agree about the GC considerations. I meant my points to mainly apply to the use cases where safety such as that offered by Boehm is eschewed in order to achieve other powers at its expense, which I feel is brought up a lot by Nim proponents as strengths during these discussions.