Re: ringPointerTest

[Andrew Johnson <anj@aps.anl.gov>]

On 11/18/2014 05:02 PM, Andrew Johnson wrote:
> Actually I even get that when I run the test in a priority=100 thread
> (VxWorks task priorities go from 1=highest to 255=lowest).

... but we only use a subset of those priorities.

Fixed by doing as I originally suggested and setting the thread priority
to something in the middle before running the testPair() subroutines. I
committed that change.

Now I'm getting a failure in scanIoTest on VxWorks 5.5.2, 68K:

***** scanIoTest *****
1..10
Starting iocInit
############################################################################
## EPICS R3.15.0.2-DEV $$Date: Tue 2014-10-07 01:50:24 -0500 $$
## EPICS Base built Nov 18 2014
############################################################################
iocRun: All initialization complete
# Testing single callback thread
#     using 11 ioscan sources, 5 records for each, and 101 loops
ok  1 - No out-of-order processing
not ok  2 - All per-record process counters match number of loops


This was apparently caused by the test not waiting long enough for the
callback threads to finish because changing the epicsThreadSleep() delay
from 0.1 to 1.0 was enough to resolve that.

Committed.

Now all tests pass on vxWorks-68040. More tomorrow.

- Andrew
-- 
People everywhere confuse what they read in newspapers with news.
-- A. J. Liebling