Does this patch help? It seems like only the parent should
attempt the lock. I will try it out for a while. --mark--
--- /usr/bin/hp-systray.orig 2011-03-18 11:01:58.000000000 -0700
+++ /usr/bin/hp-systray 2011-03-18 11:04:40.000000000 -0700
@@ -82,7 +82,10 @@
log.warn("No hp: or hpfax: devices found in any installed CUPS
queue. Exiting.")
sys.exit(1)
- mod.lockInstance()
+ # only attempt the lock if this pid is the session leader,
+ # otherwise it creates a race condition where children get there first
+ if os.getpgid(os.getpid()) == os.getpid():
+ mod.lockInstance()
r1, w1 = os.pipe()
log.debug("Creating pipe: hpssd (%d) ==> systemtray (%d)" % (w1, r1))
--- /usr/bin/hp-systray.orig 2011-03-18 11:01:58.000000000 -0700
+++ /usr/bin/hp-systray 2011-03-18 11:04:40.000000000 -0700
@@ -82,7 +82,10 @@
log.warn("No hp: or hpfax: devices found in any installed CUPS queue. Exiting.")
sys.exit(1)
- mod.lockInstance()
+ # only attempt the lock if this pid is the session leader,
+ # otherwise it creates a race condition where children get there first
+ if os.getpgid(os.getpid()) == os.getpid():
+ mod.lockInstance()
r1, w1 = os.pipe()
log.debug("Creating pipe: hpssd (%d) ==> systemtray (%d)" % (w1, r1))
