[Ifeffit] Bug report

[Scott Calvin]

Hi Bruce,

Two issues I've found with the latest Artemis release; one is a bug, and
one is undesirable behavior (at least to me!).

My specs:

Running Artemis 0.8.000 under Windows XP Professional.

The bug:

To see this bug, open examples/Artemis/EuTiO3/eto.apj.

Click Fit and let it use the default options for naming. Now un-include the
Ti edge from the fit. Change amp_ti, e0_ti, and sso_ti to "set." Run a new
fit (so that it saves the old one). Change the k-range on Eu and run a new
fit again (#3).

Now highlight fits 1 and 2 under the Eu edge and try to "discard selected
fits." 

For me, it deletes fit 1, but not fit 2. It also dumps an error to the
screen, which I'll send to you separately as an rtf file. (Incidentally,
Artemis used to put these error messages in a file somewhere--does it still
do that? It seems silly to have to do a screen capture to send it to you.)

This behavior seems to be associated with multi-edge fits where some of the
edges are turned off for some of the fits, but I wouldn't swear to it.

The undesirable feature:

This one is easy to see; even the plain vanilla copper project will do it.
Basically, restoring a fit model tends to take subsequent journal entries
with it. That is not how a journal should work--if I try five fits and take
notes along the way, and then decide to return to the second fit, the notes
I've written since the second fit (including, say, the formula for the
potion which will allow cats to recite Shakespeare) shouldn't vanish.

--Scott Calvin
Sarah Lawrence College