On Feb 7, 2011, at 13:02 , Simon Perreault wrote:
Furthermore, you're ignoring alignment issues.
What potential alignment problem can occur with the union that Dave
proposes? My understanding is that C compilers will ensure that the
union is aligned to the largest member, so my understandin
o: it triggers a specific callback in an event_loop, which can be
across threads. I use this in one program to implement a thread pool:
I have a "main" event_loop which does most network I/O processing,
then I use event_active to pass some specific tasks to other t threads.
Evan
--
, and
using event_active() to "trigger" an event to be run in another thread.
This doesn't really address the multi-threading issue though, since it
is a fair amount of manual work.
Evan
--
Evan Jones
http
timeout.
Hope this helps,
Evan Jones
--
Evan Jones
http://evanjones.ca/
***
To unsubscribe, send an e-mail to majord...@freehaven.net with
unsubscribe libevent-usersin the body.
This seems like a good interpretation, and makes sense now that I
think about it.
ONCE|NONBLOCK doesn't seem reasonable to me atm.
Adding an event_err() for that combination might make sense then?
Evan
--
Evan Jones
http://evanjones.ca/
*
VLOOP_NONBLOCK; poll the loop once, run highest priority, then exit
11 = EVLOOP_BLOCK_FOR_EXPLICIT_EXIT; run until there is an explicit exit.
Thanks,
Evan Jones
--
Evan Jones
http://evanjones.ca/
***
To unsubscribe, send