Peter Prymmer <[EMAIL PROTECTED]> writes:
> On Tue, 28 Nov 2000 [EMAIL PROTECTED] wrote:
>> Well, for one thing TEST.COM should be doing:
>> Set Message /NoFacility/NoSeverity/NoIdentification/NoText
>> to try and keep the %RMS-E-FNF messages from being emitted in the first
>> place.
> Why did this requirement only come up just now? There are plenty of other
> tests that intentionally do nasty things to ensure that the proper error
> message is generated.
The SET MESSAGE has been in TEST.COM for a while, coincident with
piping and test changes earlier this year. Usually it does a good job
of supressing the vmsish error messages from the test suite, but
apparantly not in this case.
And yes, it should be inherited by subprocesses...I just did a test
and found that it was indeed inherited by subprocesses of
filter-util.t
>> A quick look at the filter-util.t test indicates that we may have
>> another one of the typical "need double quotes around -I" difficulties.
> This was the first thing that I tried. I tried double quoting the "-I"'s,
> I tried removing them altogether and I tried explicitly listing:
> "-I." "-I../lib" to no avail. Note that the later tests in filter-util.t
> also use the -I. $Inc idiom successfully on VMS.
Well damn.
Right now it's beginning to look like the error is coming from somewhere
deep in Perl. I'm building a VMS-debug version and see what gives.
--
Drexel University \V --Chuck Lane
======]---------->--------*------------<-------[===========
(215) 895-1545 _/ \ Particle Physics
FAX: (215) 895-5934 /\ /~~~~~~~~~~~ [EMAIL PROTECTED]
