Subject: Re: [nmr_sparky] Runaway Strips
From: Mandar T. Naik
Date: Jun 6, 2006
Previous: 160 Next: 162
Hi Thomas,
Thanks for the technical insights. In few of my attempts to reproduce that
behavior I found running orthogonal view ov on any strip, then closing
those new views and further deleting the original strip causes this problem.
But this behavior seems to be explainable after reading your email and I
generally dont do this in routine operation. There must be something else,
which I will try to keep track of.
To answer your second question, yes, the number appended are always high,
sometimes even in thousands.
Do you think minimizing all spectral views while running strips will have any
implication? I generally dont like my desktop cluttered and keep most or all
views minimized and sk from the peak list.
Best regards
-mandar
On Wednesday 07 June 2006 01:08, Thomas Goddard wrote:
Hi Mandar,
Ive never seen strip plot strips added to the Sparky Windows menu.
That is a bug. I tried to make that happen but was not able to. I looked
at the Sparky code and do not see any way it could happen. The Sparky
code only adds a spectrum view to the Windows menu if it is a top level
window. Do you have any ideas about how to consistently produce the
problem? I guess the entry in the Sparky Windows menu typically has
a pretty high number appended (eg hnca/87)? I can see in the code that
if a strip did get added to the Windows menu then it would never be
removed (since it is not a top level spectrum window), and then after
the spectrum is deleted, clicking that entry would cause a crash. My
only theory about how it could happen is that it is being added to the
menu before the spectrum window is displayed and Tk (the windowing toolkit
used by Sparky) is incorrectly reporting that it is a top level window.
I also see that Sparky takes 10 seconds when I delete the first peak
of a project with a few thousand peaks. That is unfortunate. Im guessing
it is creating peak lists that are not necessary. It would require more
investigation than I have time for now, though I suspect it could be fixed.
Thanks for reporting these problems.
Tom