[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: ./check failures



Graham Phillips wrote:
> Patrick, 
> 
> > Please send them along if its not obvious what went wrong.

Hmm... looks like a race condition of some sort.  The DataProtocol is
registered at startup by the node bootstrapping code, so the MID
should always be found.

If you're interested in debugging this problem, there is a better
shell script for debugging Ping in the java/antrs/apps/ping/ directory 
(ping.start).  It fires up multiple xterms.  We only have glorified
printf debugging available... unless you've got some good tools for
you JVM...

You'll want to turn on the tracing (just edit ping.start).  Probably
setting janos.debugFlags to ANTSR_STARTUP and ANTSR_DYNLOAD is sufficient.

If you're not so interested in debugging the problem, could you 
edit test/java/antsr/ping-regression.start, change the line 
	JVMOPTS=-Djanos.debugFlags=NONE
to
	JVMOPTS=-Djanos.debugFlags=ANTSR_STARTUP,ANTSR_DYNLOAD

and then re-run the test (manually) via:
	env JANOSVM=/asterix/janos-20000214/tools/bin/janosvm sh ping-regression.start

Send me the output if that fails again (it may succeed occassionally
if its really a race condition).

-Pat

----- ----- ---- ---  ---  --   -    -      -         -               -
Pat Tullmann                                       tullmann@cs.utah.edu
      That which does not kill you just didn't try hard enough.




[ Janos ] [ OSKit ] [ Network Testbed ] [ Flick ] [ Fluke ]
Flux Research Group / Department of Computer Science / University of Utah