problems with 'picking' on various models of the 4D

[Steve Dempsey]

Has anybody noticed any problems with 'picking' on the 4D workstations?

Recently, several people running my molecular modeling software (MMS) have
called to report that it has stopped working.  I have traced the problem to
erroneous results being returned by 'endpick'.  Just to make life complicated,
different models of the 4D report different results, and on the one 4D that
I have routine access to (a 240GTX running 3.1G) the problem does not occur.

Most bizarre has been the behavior of Personal Irises.  These machines are
running 3.1D and 3.1G.  In three cases people reported identical situations:

    1. They had been using my software for weeks or months without problems.
    2. Their machines crashed because of a local power failure.
    3. My program would not operate properly after power was restored.

No hardware or software changes had taken place between steps 1 and 3.
In one case the PI was turned back off for the weekend, and when it was
restarted on Monday morning the problem was gone!

What I am expecting from 'endpick' (and what I see on the 240GTX) is a function
value of 1 (indicating 1 hit) and a pick buffer containing two values; a 1
(length of name stack) and the value placed on the namestack with 'loadname'.

The Personal Irises are returning 0 from 'endpick', indicating that no hits
occured.

A 70G (3.141592638...) and and a 70GT (3.0) report multiple 'null' hits in
addition to the real hit.  For example, if 'endpick' returns a value of 3
the pick buffer contains:

	pickbuf[0] =	0		0 length namestack
	pickbuf[1] =	0		0 length namestack
	pickbuf[2] =	1		namestack has one value
	pickbuf[3] =	expected value

--------------------------------------------------------------------------------
Steve Dempsey						          (619) 534-0208
Dept. of Chemistry Computer Facility, B-014	   INTERNET:   sdempsey at ucsd.edu
University of Calif. at San Diego		     BITNET:   sdempsey at ucsd
La Jolla, CA 92093				       UUCP:   ucsd!sdempsey