Bugs reported in PHOELIX
The PHOELIX distribution was updated on Janurary 4, 2000. If you got your copy
before that date you may have a number of bugs that we managed to find in the
last few weeks. That's the price of getting the new SUPRIM on board. You might
want to get a new copy if yours is an older distribution.
From now on I will list bugs that we find here for your consideration. Please
report any bugs you find to phoelix@scripps.edu
It should be noted that we find little bugs in the code all the time. In fact,
if we were writing code for Star Wars defense everyone would be charcoal by now
and cockroaches would rule the earth. There are a lot of lines of code here.
These bugs are generally of the nuisance category. By that I mean they are
annoying but don't affect the data. When we find one that is serious we will
post it here and, if you sent me mail telling me you were using PHOELIX, send
you some mail about it. We test the code as much as we can, but we are not
in the software business. We view this as a tool, just like you do. You should
always test this code for yourself. And if you use this software you have to
agree to let us know if you find any bugs.
Any bugs with fixes listed here are also fixed in the latest distribution.
6/21/96: A bug was introduced into ee, the equation evaluator
in the suprim package, during the change over to SUPRIM v5.1. The function of
torad and todeg (conversion between degrees and radians) got switched around,
so that if you said "todeg 3.14159", you actually got the radian conversion
from 3.14159 degrees. Similarly, if you said "torad 90", you actually got the
degree conversion from 90 radians! This makes a difference if you are using
the trig functions within ee, because they expect radians. The only part of
PHOELIX which uses this is the straightening algorithm, which rotates the
template to match the local orientation of the filament prior to calculating
a cross-correlation. As a result, the cross-correlation peaks may not be as
sharp as they might otherwise be. As a note, we didn't put in this rotation
of the template until recently, and the straightening worked ok without it;
it was just not optimal.
The fix: A new ee executable and a new img.h. Uncompress ee.Z and put the ee
binary in your suprim bin directory. Put img.h in the suprim include directory.
6/7/96: A fix for irll required for the new SGI's with Impact
graphics. When drawing, irll is not releasing the line, but rather playing
connect the dots. No big deal, but here's a fix.
The fix: a new irll executable and source. Uncompress irll.Z andput the
irll binary in your suprim bin directory. Put the irll.f source in
src/ir/irene.
1/30/96: A rather serious bug in str (SUPRIM straightening algorithm)
which resulted in scaling errors that increased as the curve to be straightened
increased in angle from the horizontal axis. The error is less than 10% at
angles less than 20 degrees, but is really serious at, say, 90 degrees. Reported
by Gina Sosinsky.
The fix: a new strlib.c. Download this file and move it to your
suprim5.1/src/str directory. Then cd to this directory and type "make". If you
don't want to do this, just grab a new copy of the entire distribution.
contents © 2000
awl