Re: [Ifeffit] Has anyone got this technical problem during the fit?

Hi Nirawat, I see the same behavior your describe -- fit 61 is empty and the big green button is disabled. It is hard for me to know exactly what happened -- I would need to be able to reproduce the events leading up to the problem in order to track it down. That said, I can make a guess. I am guessing that you were in the middle of a very long session with Artemis, having run many, many fits during that session. I am guessing that you ran into Ifeffit's built-in memory limits. This is an unfortunate side-effect of the language Ifeffit is written in. The best work-around for this problem is to save your work often and to save your work to different filenames. In this way, if you run into a problem of this sort, you can at least "go back in time" in the sense of using a recently saved project file. That is, if you are working on "myfit.apj", after a while, save it to "myfit_01.apj" and so on. The best thing I was able to do in the limited time I have to spend on this was to remove all the fits after fit #61 from your project file. That is attached. (I also removed the first 50 fits, just to make it a bit smaller.) This loses the 25 or so fits you somehow did after running into your problem, but it at least enables you to load and continue using this fitting project. Hope that helps. Save early and often. B On Friday, August 05, 2011 07:53:52 pm NIRAWAT THAMMAJAK wrote:

On 4 August 2011 18:39, Matt Newville wrote:

...

Hi Nirawat, I think it would be really helpful to see the trapped errors. I can't promise that I could figure it out, but if you send the project file Bruce or I might have a chance.

-Matt

On Aug 4, 2011 9:18 AM, "NIRAWAT THAMMAJAK" wrote:

...

Dear All,

I was just wondering if some of you may have experienced a problem like

this

...

in the fit;

I was trying to improve the precision of some fit parameters by doing

guess,

...

def, and set. In the 61st fit, I tried to split parameters from

constraints

...

that i have made. Then, the program didn't show me the results of the

fit. I

...

knew that something might have gone wrong, the program forced me to

close.

...

Fortunately, there was a window asking me to save it before closing the project.

I opened again the project file after it was safely saved. This time, I could not perform any fitting - the green 'Fit' button was frozen and

it's

...

not possible to click it. The red message below showed "Artemis trapped

one

...

or more errors! Error message dumped to the screen", but I couldn't see

such

...

error message on the screen.

I closed the project and opened the file again, the 'Fit' button seemed

to

...

work this time. Anyway, I still had that red message below said "Artemis trapped one or more errors!...". After clicking at the data file, the red message was gone and replaced by the number of data points, "22

independent

...

points data points (Nyquist): ...". The numbers of the fit and FT

parameters

...

were then shown in the boxes on the left as usual, but the title of data

was

...

replaced by "o"!

I thought it was okay, then I performed other fits. It worked fine. Until the recent fit that I have it blown up, so I saved and closed the

project.

...

Now when I try to open the file, I have got the earlier problem, frozen 'Fit' button. Clicking at data file doesn't help as it used to be. I have tried another saved file in this series of .apj files (same data), they

have

...

got the same problem. I have tried open another data file like eto.apj in the demo, it's also got that problem with the red message "Artemis

trapped

...

one or more errors!.." too.

I don't understand what should have gone wrong with the program. Your

help

...

will be very much appreciated.

Thank you, Nirawat.

-- Nirawat Thammajak DPhil Candidate in Inorganic Chemistry Balliol College Oxford OX1 3BJ United Kingdom

_______________________________________________ Ifeffit mailing list Ifeffit@millenia.cars.aps.anl.gov http://millenia.cars.aps.anl.gov/mailman/listinfo/ifeffit

-- Bruce Ravel ------------------------------------ bravel@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/ifeffit/Demeter