That's how I spiked with to see if Java runs a 'finally' block after it receives a 'SIGINT'
On Tue, Mar 2, 2010 at 12:29 PM, kristian <m.krist...@web.de> wrote: > what about > > http://gist.github.com/319889 > > ??? > > regards > Kristian > > On Wed, Mar 3, 2010 at 1:48 AM, Abhijit Hiremagalur > <a...@pivotallabs.com> wrote: >>>> For some additional context, our JRuby application is consuming JMS >>>> messages from a Websphere MQ instance. The 'ensure' blocks that we'd >>>> like to run close connections to this queue gracefully. Not doing this >>>> upsets the folks who manage these queues :) >>> >>> There's another option: don't rely on SIGHUP :) No other JVM-based >>> program would. >>> >> >> Interesting, a little spiking showed Java 'finally' blocks aren't run >> on a SIGHUP or SIGINT either, so JRuby's behaviour appears consistent >> with Java's. >> >> https://gist.github.com/881542ca3cca1281e959 >> >> So if SIGHUP isn't an option, what would you say is a good way to tell >> a long running JRuby process to stop gracefully? >> >> I've googled around a little to see how Java folks approach this and >> haven't found a clear/consistent answer. --------------------------------------------------------------------- To unsubscribe from this list, please visit: http://xircles.codehaus.org/manage_email
