comp.lang.ada
 help / color / mirror / Atom feed
From: Jim Podmore <jim@jim-podmore.demon.co.uk>
Subject: Signals and delays
Date: Wed, 28 Aug 2002 08:06:36 +0100
Date: 2002-08-28T08:06:36+01:00	[thread overview]
Message-ID: <3D6C767C.6020305@jim-podmore.demon.co.uk> (raw)

I have an application in GNAT running under Irix 6.5 which uses a pragma 
attach_handler to catch a signal (sigusr2) from another process.

All seems to work fine, except once in a while the program stops, with 
all of the cyclic tasks stuck on delay statements that should have 
expired long ago.  The problem seems to occur at about the same point in 
the regular cycle of the task execution each time, and one thing that 
struck me is that one of the delay statements is carefully timed to 
occur at the exepcted time of arrival of the next sigusr2 signal.  When 
I changed the timing slightly to avoid any conflict, the problem 
(fingers crossed) seemed to go away.

So is there a known problem with conflicts between signal handling and 
delays / real time clock ?




                 reply	other threads:[~2002-08-28  7:06 UTC|newest]

Thread overview: [no followups] expand[flat|nested]  mbox.gz  Atom feed
replies disabled

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox