>Interesting, not sure what this means. I guess starting up with an XML
>file is possibly in a different thread than reading it from inside
>Analysis, so that might be why there is a difference. Doing a google
>search on that error comes up with loads of hits, but nothing that
>illuminating that I could see. We haven't seen that specific error on
>Linux that I know (at least not very often, I suspect we must have seen it
>at some point). One of the things we hope to do in the near future is to
>get a Mac so that we can better support that OS. (We actually managed to
>get it to compile ok using the Aqua Tk on a friend's computer a couple of
>weeks ago, but frankly it wasn't an improvement on the ordinary Tk, and it
>also crashed horridly when actually trying to use the Tk.)
Actually, just before the abort occurs, analysis has opened the xml file
specified on the command line. A window briefly flashes up about reading
the peptide structure, and window 1 with the HSQC spectrum is displayed.
The crash happens around the time that window 2 to which the 3D HSQC NOESY
and HSQC TOCSY belong is to be displayed as a window with 2 strips.
For me, analysis 1.02 worked just about flawlessly. Therefore, whatever
causes this is probably related to some change made subsequent to the
release of 1.02 and a diff of files may reveal the change. The error
message states that the problem is NULL tstate in PyEval_RestoreThread.
I take this to mean that a thread state is not assigned a value. This
suggests to me that the problem may be a failure to assign a value, possibly
a pointer value, and may involve a failure in memory allocation.
Sincerely,
--
Bruce D. Ray, Ph.D.
Associate Scientist, and Operations Director
NMR Center
IUPUI
Physics Dept.
402 N. Blackford St.
Indianapolis, IN 46202-3273
|