From vbarr@cme.nist.gov Wed Jun  6 17:56:51 2001
Received: from dribble.cme.nist.gov (dribble.cme.nist.gov [129.6.32.31])
	by swi.psy.uva.nl (8.11.2/8.11.2) with ESMTP id f56Fuo302971
	for <prolog@swi.psy.uva.nl>; Wed, 6 Jun 2001 17:56:51 +0200 (MET DST)
Received: from blur.msid.cme.nist.gov (blur.msid.cme.nist.gov [129.6.77.135])
	by dribble.cme.nist.gov (8.9.3/8.9.3) with SMTP id LAA00352
	for <prolog@swi.psy.uva.nl>; Wed, 6 Jun 2001 11:56:48 -0400 (EDT)
Message-Id: <200106061556.LAA00352@dribble.cme.nist.gov>
Date: Wed, 6 Jun 2001 11:56:47 -0400 (EDT)
From: Vincent Manuel Barrilliot <vbarr@cme.nist.gov>
Reply-To: Vincent Manuel Barrilliot <vbarr@cme.nist.gov>
Subject: Re: [SWIPL] SWI-Prolog 4.0.6 problem on Solaris 2.6
To: prolog@swi.psy.uva.nl
MIME-Version: 1.0
Content-Type: TEXT/plain; charset=us-ascii
Content-MD5: mSwCHvmt1vvXqCF2gzk9ew==
X-Mailer: dtmail 1.2.1 CDE Version 1.2.1 SunOS 5.6 sun4u sparc 

Jan,

Thank you very much for your fast answer !

Look: I think that's JPL trying to load its stuff. But WHY is it trying to load it from my work directory and not
from its installation directory ?

tat64("/proj/ppe/PCR/Vincent/Prototype/ProE/Java/jpl/fli/predicate_t.class", 0xE43800D0) Err#2 ENOENT
stat64("/msid/depot/pub/jade-2.1/jpl/fli/predicate_t.class", 0xE43800D0) Err#2 ENOENT
poll(0xEF005AC8, 0, 10)                         = 0
stat64("/proj/ppe/PCR/Vincent/Prototype/ProE/Java/jpl/fli/PointerHolder.class", 0xE437ECA8) Err#2 ENOENT
stat64("/msid/depot/pub/jade-2.1/jpl/fli/PointerHolder.class", 0xE437ECA8) Err#2 ENOENT
poll(0xEF005AC8, 0, 10)                         = 0
stat64("/proj/ppe/PCR/Vincent/Prototype/ProE/Java/jpl/fli/atom_t.class", 0xE437FFB0) Err#2 ENOENT
stat64("/msid/depot/pub/jade-2.1/jpl/fli/atom_t.class", 0xE437FFB0) Err#2 ENOENT
poll(0xEF005AC8, 0, 10)                         = 0
stat64("/proj/ppe/PCR/Vincent/Prototype/ProE/Java/jpl/fli/qid_t.class", 0xE4380150) Err#2 ENOENT
stat64("/msid/depot/pub/jade-2.1/jpl/fli/qid_t.class", 0xE4380150) Err#2 ENOENT
sysconfig(_CONFIG_CLK_TCK)                      = 100
poll(0xEF005AC8, 0, 10)                         = 0
times(0xE4380FE8)                               = 231422783
access("/proj/ppe/PCR/Vincent/Prototype/ProE/PART_SLOT.pl", 4) Err#151 ESTALE

BING !!
PART_SLOT.pl is on NFS, and it looks like the handle is stale. But I don't understand why. I generate this file 
from C code, then fclose() it (successfully, I've checked that), then send a notification (via socket) to the Java 
agent so that it starts a Prolog engine and consult() it.

What I didn't tell you is that the C code generating the PART_SLOT.pl is running on a remote computer (with a lot 
of memory to run Pro/ENGINEER) while the Java agent is running on my computer.
I had a look on the net and figured out that maybe when a computer changes a file on a NFS it creates a new handle 
for it, making invalid those used by other computers and which may not be up to date. (they should be updated 
automatically, don't they ??).
I have tried to run both of the wrapping agent and Pro/E on my machine so that there is no problem of NFS handle, 
and it works !!!

> system level and what errors are returned.  Finally, is it just this
> file, any file on this file-system, in this directory or just any file?

Your guess put me on the track, and thank you for telling me about truss, I think this thing is gonna be my life 
saver in the future !
Thank you very much Jan ! :-)

Vincent


