Re: Fix to long-standing crashes in GC

[Stefan Monnier <monnier@iro.umontreal.ca>]

>> The sweep step simply has to be atomic, just like most of the GC itself
>> (afterall, that's why the mark&sweep algorithm is in the category called
>> "stop-the-world").

> True -- but if you try to debug GC errors, I prefer the code to be as
> safe as possible.

If you want it for debugging purposes, feel free to switch the sweeping
of strings w.r.t the sweeping of cons cells.  I've never had to debug in
there, so I wouldn't have any preference either way.  As I said, it
shouldn't make any difference w.r.t to correctness.
The two "reasons" why I undid this part of your change were:
- you also moved the sweeping of cons cells to before the sweeping of weak
  hash tables, which sounds seriously wrong.
- I started from a full undo of your patch.


        Stefan