• Modified mopcrt.c for mopac to do optimizations for excited states. A typical command is like: $ACHOME/exe/antechamber -fi mol2 -fo mol2 -i test.mol2 -o test_bcc.mol2 -c bcc -nc NET_CHAEGE -m MULTIPLICITY -j 4 -df 0
• Fixed two bugs about memory allocation in bondtype.C
• Improved database.c and it is now more robust

• Modified the charge.c function and the parameters of calculating the Gasteiger charges. The parameters of a,b,c,d are from those of Gasteiger et al. (J. Gasteiger, M. Marsili, Tetrahedron, 36, 3219-3228 (1980)) except those for phosphate, which are borrowed from sybyl7.2. The charges calculated by the antechamber program may slightly different from those by sybyl due to a more stringent converge criterion is applied. Unlike sybyl, which needs the users to assign atomic formal charges before the charge calculation, the antechamber.c program could assign formal charges automatically.
• Added a new function of spaceline() to judge if a line is a space line (only has spaces or control characters) or not. We found that strlen(line) == 1 fails for files generated by Windows (there is cntr-M at the end of each line either visable or invisable)
• Revised prep.c to read in the atom connectivity information from a prepi file and no atomic connectivity is judged for the prepi input. Now antechamber and parmchk wrok properly even when the input structures are not good enough.
• Created an atom type description file for Tripos atom types (ATOMTYPE_SYBYL.DEF) and revised the atomtype.c program to recognize sybyl atom types. The wmol2() funcition in mol2.c was also modified accordingly.
• Improved database.c greatly. Now database.c can handle almost any multiple-record database file only if it is well organized (some symbols to separate each record, such as "$$$$" of sdf, "@MOLECULE" of multiple-mol2 files). It can even properly handle a file like all_amino94.in in $ACHOME/dat/leap/prep. A sample input file (sample.def) is provided in the $ACHOME/test/database.
• Added a new atom type, cz for the sp2 carbon in a guadinium group. It was found by Cristophe Guilbert at UCSF that the original gaff.dat parameter set breaks the planar conformation of 2-amino-1H-benzo[d]imidazol-3-ium. We studied a set of similar molecules and found that only 2-amino-1H-benzo[d]imidazol-3-ium (the first molecule) caused trouble. We solved this problem by simple applying a larger force constant of improper dihedral angles involved by cz (X -X -cz-X 10.5 180. 2.). All the other parameters involved by cz are borrowed from those of c2. We think it may not a good idea to introduce atom types for five-memberred rings because (1) many atom types (sp2 C, sp3 C, sp2 N etc) are added; (2) the current parameter set works fine for most of molecules that have five-memberred rings.

	68,69c68
	<       if(ringnum == 0)
	<               return;
	---
	>

     64         RING *ringbak;
     65         int ringbaknum = 0;
     66         int tmpint;
     67         int index;
     68         if(ringnum == 0)
     69                 return;
     70         ringbak = (RING *) malloc(sizeof(RING) * ringnum );
     71         if (ringbak == NULL) {
     72                 fprintf(stderr, "memory allocation error 
                                   for *ringbak\n"        );
     73                 exit(0);