From: James Irving (
Date: Mon Jun 30 2008 - 18:10:13 CDT

Hi Axel,
That seems to be a successful work-around, after putting
in my login script VMD no longer crashes when I'm loading
Thanks very much!

2008/7/1 Axel Kohlmeyer <>:
> On Mon, 30 Jun 2008, James Irving wrote:
> JI> Hi,
> james,
> JI> I was wondering whether anyone else has had this problem.
> i don't have a mac and thus never seen this, but it reminds me
> of a discussion on this list recently. it most likely is some
> underlying issue of the operating system itself, but if it is
> limited to the file chooser dialog, you could try to work around
> it by setting the environment variable VMDFILECHOOSER to FLTK
> before starting VMD, or in your .vmdrc file as
> JI> I've been using VMD 1.8.6 for a couple of months on my OSX leopard
> JI> macbook pro, and recently VMD 1.8.7(CUDA). Both have worked fine.
> JI>
> JI> Two obvious changes to my system: an auto-upgrade from OSX10.5.2 to
> JI> OSX10.5.3, and installation of ActiveTcl 8.4.16 followed by 8.5.2
> JI>
> JI> VMD opens OK from the Applications folder
> JI> When I bring up the 'load molecule' dialog at some point VMD crashes -
> JI> either straight away, or after I've loaded a coordinate set, or after
> JI> loading a trajectory, but always by loading a second trajectory.
> JI> The error on crash is either 'Bus error' or 'Segmentation fault'
> JI> This happens on both versions of VMD
> JI> ActiveTcl version made no difference
> JI>
> JI> When I run VMD from within gdb, the error messages produced are typically:
> JI>
> JI> Program received signal EXC_BAD_ACCESS, Could not access memory.
> JI> Reason: KERN_INVALID_ADDRESS at address: 0xa099b00c
> JI> 0xffff1281 in __longcopy ()
> JI>
> JI> or
> JI>
> JI> Program received signal EXC_BAD_ACCESS, Could not access memory.
> JI> Reason: KERN_INVALID_ADDRESS at address: 0x440b0004
> JI> 0x92d8daba in _CFDoExceptionOperation ()
> no clue where this is coming from, but i would suspect
> the graphics driver. this looks like it is an error about
> copying data from the kernel to userspace or vice versa
> and graphics is a prime candidate for that...
> cheers,
> axel.
> JI>
> JI> Thanks for any assistance,
> JI> James
> JI>
> --
> =======================================================================
> Axel Kohlmeyer
> Center for Molecular Modeling -- University of Pennsylvania
> Department of Chemistry, 231 S.34th Street, Philadelphia, PA 19104-6323
> tel: 1-215-898-1582, fax: 1-215-573-6233, office-tel: 1-215-898-5425
> =======================================================================
> If you make something idiot-proof, the universe creates a better idiot.

Dr. James Irving
NH&MRC C.J. Martin Fellow
School of Biomedical Sciences
Building 13D
Monash University
Wellington Road
Melbourne 3800