[Ifeffit] Athena crash report

Bruce Ravel bravel at bnl.gov
Mon Oct 28 09:47:07 CDT 2013


Hi George,

I now know what is causing the problem.  For one of the main libraries
that Demeter depends upon that, your computer has a version that is
actually ahead of any of my computers.  One of the things that library
does is manage parameters, including type constraints.  That means
that it does early checking that, for example, a numerica parameter is
being set to a numeric value.

A recent version of that library changed its definition of what
constitutes a number.

Apparently, your initialization file (~/.horae/demeter.ini) has a
value for the default of the lower bound of the normalization range of
"15.".  That value, with the trailing dot, is causing the problem.
That is, I can replicate your problem by setting that parameter to
that value.

I know how to fix Demeter to do the right thing given the current
state of that library and such a perfectly reasonable numeric value as
"15."  That fix will make it into the next release.

The immediate work-around is to edit the file ~/.horae/demeter.ini and
search for this section:

     [bkg]
     e0_fraction = 0.5
     spl2 = 0
     nor2 = -100
     clamp2 = 24
     e0 = derivative
     kwindow = hanning
     pre1 = -150
     dk = 0
     fnorm = true
     nor1 = 150
     nnorm = 3
     kw = 2
     nclamp = 5
     spl1 = 0.0
     clamp1 = 0
     rbkg = 1.0
     flatten = true
     pre2 = -30

You will have a line like

     nor1 = 15.

Edit that to read

     nor1 = 15

and try restarting Athena.  If you get a similar error message, search
through that section or elsewhere in the .ini file for numeric values
that have a trailing dot.

Thanks for finding and reporting this!
B


-- 
  Bruce Ravel  ------------------------------------ bravel at bnl.gov

  National Institute of Standards and Technology
  Synchrotron Science Group at NSLS --- Beamlines U7A, X24A, X23A2
  Building 535A
  Upton NY, 11973

  Homepage:    http://xafs.org/BruceRavel
  Software:    https://github.com/bruceravel



More information about the Ifeffit mailing list