********> bugfix.5

Correction Author: Mats Eriksson & David A. Pearlman 
Correction Date: 11/06/97

Programs: GIBBS

Severity: Modest

Problem: 

       When using namelist input, if one attempts to set the molecular 
       cap pointer NATCAP by setting IVCAP = 1 and NATCAP > 0, NATCAP
       will be reset to 0.

       This problem only manifests itself when namelist format input
       is used (not when number-in-columns format input is used). And
       this problem only occurs when IVCAP = 1.
       
Affects: 

       The cap atom pointer will be improperly set to 0.

Cause: 

       Internally, the program expected the variable to be used as
       NATCAP to be read in as MATCAP when IVCAP=1. This was not
       being done for the namelist input.

Fix:  Make the following changes to routine MDREAD

---To routine MDREAD in file mdread.f (GIBBS directory) -------

*** mdread.f	Fri Nov  7 13:19:32 1997
--- mdread.original	Fri Nov  7 13:16:00 1997
***************
*** 616,622 ****
  C ------------------------------------------------
  C
    125 READ(5,9048) NSTPE,NSTPA,DTE,DTA
!       READ(5,9048) IVCAP,NATCAP,FCAP
  C
  C ----------------------------------------------------------
  C If JFASTW=1, read in extra line defining the water residue
--- 616,622 ----
  C ------------------------------------------------
  C
    125 READ(5,9048) NSTPE,NSTPA,DTE,DTA
!       READ(5,9048) IVCAP,MATCAP,FCAP
  C
  C ----------------------------------------------------------
  C If JFASTW=1, read in extra line defining the water residue
***************
*** 640,651 ****
  C
        NRCP = NTR
        ISHAKE = 1
- C
- C NATCAP will be set by the call to RDPARM.  Subsecquently, we will overwrite
- C this value with the user-specified value stored in MATCAP.  if IVAP = =1
- C
-       MATCAP = NATCAP
- 
  C
  C IF INTR.GT.0, then the user is reading in internal restraints;
  c We cannot read these here, because if the user speifies
--- 640,645 ----

------------------------------------------------------------------------------
------------------------------------------------------------------------------

 
Temporary workarounds: 

     You can avoid this bug by using numbers-in-columns input.

Routines affected:

     GIBBS   Routine MDREAD in file .../amber5/src/gibbs/mdread.f


-- 
    David A. Pearlman                           email: dap@vpharm.com
Vertex Pharmaceuticals Inc.                    
130 Waverly St.                   "There are only 25 great people in the world &
Cambridge, MA 02139-4242         5 of them are hamburgers..." -- Cptn Beefheart