Re: Fwd: Re: [boost] Re: Digest Number 754

From: Hillel Y. Sims (hys420_at_[hidden])
Date: 2001-08-06 13:15:07

--- In boost_at_y..., "Greg Colvin" <greg_at_c...> wrote:
> 
> Hillel -- does the list include close() and free() and such?
> 

Alexander Terekhov posted the full pthreads list a few days ago, it does
not include free(), but actually does include close()/fclose() and several
other cleanup-type functions. I can see this could be a real problem when
using pthreads. Well, under VMS there's nothing I can do about it except
live with it (if our app crashes at the moment of thread cancellation,
it's not really the end of the world, and still better than Windows/Java
at least).

Aside from the issue of catch (...) requiring the exception always being
reraised (which I still hold it's really invalid for any mid-layer code to
be eating any exceptions from catch (...) anyhow so that shouldn't be that
big a concern), what kind of fully safe mechanism would have to created in
the language (and would probably have to be completely independent of
pthreads?) to safely cancel threads that does not have the problem of
potentially throwing from a destructor or "throw()" code by calling
close()?

hys

__________________________________________________
Do You Yahoo!?
Make international calls for as low as $.04/minute with Yahoo! Messenger
http://phonecard.yahoo.com/