On Sun, Jul 11, 2010 at 5:36 PM, Mark A Saper wrote:

When running phenix.refine from the GUI, definitions for rigid body groups in phenix.refine that are syntactically wrong, causes the job to just hang with "running" status but no CPU is used. It would be nice if either the GUI checked for valid selections, or if phenix.refine returned an appropriate error code that would be recognized by the GUI.

Actually, the latter is exactly what should happen - the only reason why it wouldn't is if you've stumbled across a low-level C++ error that can't be propagated. I already modified the code to handle one such error a couple of months ago (it should be in 1.6.2) but there may be others I don't know about. The best way to check this is to run phenix.refine on the command line using the .eff file written by the GUI, which should end with a recognizable error message. (You can just send me the input files if you want.) However, I will try reproducing it myself with a test structure tomorrow. You can validate the selections in the GUI by clicking the "view/pick" button next to the input widget, which will load up the simple graphics window with your start model. It should be very obvious if the selection string is syntactically incorrect. Validating these automatically at runtime would be nice but it's more difficult (and time-consuming). -Nat

Hi Mark,

When running phenix.refine from the GUI, definitions for rigid body groups in phenix.refine that are syntactically wrong,

this is a strong statement. Sorry about the problem. At the moment I have no clue what you mean. So an example of a problem would be a starting point for discussion. See the page #49 here: http://www.phenix-online.org/presentations/2010_30APR_TX_PHENIX_WORKSHOP.pdf about bug reports. All the best! Pavel.