[Ifeffit] problem with the current Ifeffit installer for windows

[Bruce Ravel]

Hi all,

Last week, I discovered a problem that might affect some of you.  The
current Windows installer package, ifeffit-1.2.11.exe, has a problem
to do with how the ifeffit library was compiled.

Fortunately, this problem was subsequently fixed.  If you have
installed the updates that have come out since ifeffit-1.2.11.exe was
released, then you are fine.  If, however, you have installed the
installer *without* installing the updates, then your version of
Artemis will be performing its fits slightly incorrectly.

To find out if your installation is affected, start Artemis and select
"About Ifeffit" from the Help menu. 

  * If the version number for Ifeffit displayed in the echo area at
    the bottom of the Artemis window says "1.2.11" then you need to
    update.

  * If the version number says "1.2.11c" then you are fully updated
    and will not suffer from this problem.

If you need to update but are, for some reason, unable to do so by the
normal channel, there are instructions for updating by hand at this
wiki page:
    http://cars9.uchicago.edu/ifeffit/IfeffitOnWindows/WinUpdater

If you are using the incorrect version of the Ifeffit library, your
fits will still be reasonable, but the best fit values of your
parameters and all statistical parameters will be calculated
differently from the normal version of Ifeffit.  In my limited
experience, best fit values tend not to different outside of their
error bars, but they are different.

The reason I noticed this problem is because here at Brookhaven, our
internet access goes through a proxy server.  If you are not
configured to have the runner go through the proxy server, then it
will be unable to find and download the updates.  The solution I found
for my colleague's computer was to set the http_proxy variable in 
C:\Program Files\Ifeffit\config\updater.ini to the address and port of
the local proxy.

I will work on making a new installer next week when I return from an
experiment at the APS.

For those affected by this long-standing but recently-discovered
problem, I apologize.

B

-- 

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

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

 My homepage:    http://xafs.org/BruceRavel
 EXAFS software: http://cars9.uchicago.edu/~ravel/software/exafs/