From: Paul Brannan <pbrannan@...>
Date: 2004-08-30T22:43:22+09:00
Subject: Re: Request For Comments: exception safe ConditionVariable#wait

On Sat, Aug 28, 2004 at 05:38:53PM +0900, Bill Kelly wrote:
> Wow, nice catch! Thanks - !
> 
> Hmm... Would that mean it's possible for 
> 
>      ensure
> # -------> %exception here%
>        mutex.lock if unlocked
>      end
> 
> potentially an exception firing once we're in the
> ensure block as well?  Or is ensure implemented to
> have the properties of your no_raise style block,
> I wonder?

Actually the exception has the potential to occur when calling the
Mutex#lock method, if Ruby decides to switch threads just at the point
of the call (so it's not really between the ensure and the next
statement).

> Wow your observation has really got me curious now:
> Can the exception raised by timeout occur anywhere
> in:
>   - a rescue block?
>   - an ensure block?

I'm not an expert on the Ruby interpreter, but if I understand eval.c
correctly, it can occur any time Ruby can switch threads, which
includes but is not limited to:
  - each time a node is evaluated
  - each time (just before) a method is called
  - each time an extension waits for IO using rb_thread_select

> That would seem to make it really hard to guard
> against...

I agree.  I've never really liked the thread-based timeout for exactly
that reason (that and my C++ extensions don't always play nicely with
Ruby threads, since they keep data on the stack).  Instead I use an
event loop, and let the event loop handle my timeouts; this guarantees
that timeout exceptions are only raised from well-defined points in the
code.

Paul