[Ifeffit] Problem opening up project files

Bruce Ravel bravel at bnl.gov
Wed Jan 20 14:47:52 CST 2010


On Wednesday 20 January 2010 03:23:39 pm you wrote:
> Was the problem an ifeffit problem --  truncating strings and then not
> recognizing them??  If so, what happened?

Yes, it was.  Here's a "screenshot" of the simplest demonstration:

  bruce at clotho [horae] > ifeffit
   Ifeffit  1.2.11d Copyright (c) 2008 Matt Newville, Univ of Chicago
                  command-line shell version 1.1 with GNU Readline
  Ifeffit> set $text = 'bad string ( bad'
    ...  >

Since a lot of text gets sent to Ifeffit all at once, the only way for
Athena to avoid this problem is to make sure before sending text that
all parens get closed.

The reason that Athena sends large blocks of text rather than sending
text line by line and testing the state of Ifeffit is for performance
reasons.  The Ifeffit::ifeffit subroutine is almost always the slowest
part of anything my codes do, with the possible exception of the
Ifeffit::get_echo subroutine.  Sending lareg blocks of text is an
enormous perfomance enhancement.

> > P.S. While I am at it, I'll look into the issue of non-ascii
> > characters in filenames that Mingliang asked about earlier today.
> 
> I think this is an Ifeffit problem.  Well, you could change
> directories before reading the file, or avoid read_data() all together
> (which would help avoid problems of non-ASCII characters in the file
> name itself or in the file contents).  But I think that if a Fortran
> open is used, the problem is not easily solvable.

Hmm... good point.  Changing directories, reading the file, changing
back is a good idea, though.  However, it will not solve the issue of
data files themselves which contain non-ascii characters.  In either
case, Athena could and should do an analysis of the filename to
determine if this is a problem and then post a /useful/ error message.

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/



More information about the Ifeffit mailing list