> At 11:17 AM 9/5/2000 -0700, Peter Prymmer wrote:
> >So now that the deadline pressure is off, would anyone object to making
> >vmspipe a little more optional rather than a required part of the build?
...
> It seems to me worth noting that the extra lines in the test output, though
> they certainly look bad, don't appear to cause the tests to fail. Could
> there not just be something quirky about a subprocess talking to another
> subprocess talking to a parent process whose output is a terminal device and
> whose messaging is turned off? I still think we shoudl fix it if we can
> nail down what's causing it, but I'm not convinced that you're better off
> without the piping mods just because the test output looks flaky.
To add some more data: on a system that has the extra LF's:
$ MMS TEST
will produce the LF's (same ones that Peter sees).
But:
$ @[.vms]test ".exe" 1
doesn't have extra LF's. Both were run as batch jobs so SYS$ERROR
output would go to the log file.
To compare results so far:
Who system compiler make result
-----------------------------------------------------------------------
Prymmer ? ? ? LF's
Lane VMS/AXP 6.2-1H3 DECC V5.5-002 MMK 3.8-1 ok
Lane VMS/AXP 7.1 DECC V5.7-004 MMS 3.0-05 LF's
Berry VMS/AXP 7.1 DECC V5.2 ? ok
Please feel free to fill in data for the ?'s.
Right now, it's looking like the extra LF's might be a MMK/MMS thing
(i.e., subprocess termination writes an EOF to its SYS$ERROR, MMS
picks gets the EOF in a MBX and writes as an empty record while MMK just
ignores the EOF?)
More tests underway...
--
Drexel University \V --Chuck Lane
======]---------->--------*------------<-------[===========
(215) 895-1545 _/ \ Particle Physics
FAX: (215) 895-5934 /\ /~~~~~~~~~~~ [EMAIL PROTECTED]
