[c++-pthreads] Re: thread-safety definition

[Matt Austern]

On Jan 12, 2004, at 11:35 AM, Ted Baker wrote:

>>> This problem could be eliminated by specifying that cancellation
>>> is a special case, that cannot be caught by catch(...), i.e., that
>>> it can only be caught by a handler that names it explicitly, or that
>>> it cannot be caught at all.
>>
>> That was originally proposed, and it's a bad idea.  There's too much
>> code of the form
>>   catch(...) {
>>     do_some_partial_cleanup();
>>     throw;
>>   }
>> This is important, and it's recommended style.  Uncatchable
>> exceptions would be a major change in the C++ exception model.
>
> This existing code does not rely on being able to catch thread
> cancellation, since it was written with only normal exceptions in
> mind.  It is precisely because existing code does no know about
> cancellation that we are having this dicussion, I thought.

That's a fair point.  I suspect you're right that we'll be wasting our
time if we hope we can do something that will make C++ code that
doesn't  know about cancellations work correctly even in the
presence of multiple threads and cancellation.

The real question we should be asking ourselves is how invasive
the necessary modifications would be under whatever proposal
we're looking at. (Modifications both to the C++ standard and to user
code.)

I still think that uncatchable exceptions aren't the right answer.  
Think
about the scope of the modifications that would be required:
  1. To a good first approximation, all code that catches and rethrows
       exceptions, and that might be used in a cancellable thread, would
       have to be modified.   (The basis of my claim: if you have to do
       some cleanup whenever an exception passes through your code,
       you'll still have to do it for cancellations too.)
   1a. This isn't too different from saying that all code that catches
          exceptions will have to be modified.  Most code that catches
          an exception ends up rethrowing it.
    2. Now think about what kind of modification we're talking about.
        Again, to a good first approximation, cleanup is cleanup.  So
        the example I've give above would be rewritten:
        catch(...) {
          do_some_partial_cleanup();
          throw;
        }
        catch (CancellationException) {
          do_some_partial_cleanup();
          throw;
        }
       We don't really want that to become recommended C++ style,
       do we?

			--Matt