Change-log for cht
2.6 (mk)
- Reordered p2c portability functions to make gcc411/glibc255 happy

2.5 (mk)
- Code changes to make splint happy.
- Added Ag,Au,Hf,Hg,Ir,Lu,Mo,Nb,Os,Pd,Pt,Rh,Re,Ru,Ta,W,Y,Zr

2.4 (mk)
- Modified Makefile.in to actually use the CC and CFLAGS information propagated 
  by configure.
- Removed a few unused functions (from the builtin copy of p2c.h) that might 
  cause portability problems. 

2.3 (mk)
- supplied exact masses for the elements added in 2.2,
  added Al,As,Ba,Be,Bi,Li,Cs,In,Ga,Ge,Pb,Rb,Sb,Sc,Se,Sn,Sr,Te,Ti,Tl,V,Zn
  added greek Phi (@F) as alternative symbol for Ph
  Changed Makefile.in to build version 2.3 

2.2 (mk)
- added Ca,Cd,Co,Cr,Cu,Fe,Mn,Ni

2.1 (mk)

- regular arrows are no longer counted as C2 fragments
 
2.0 (mk)

- added support for 1.5a25 'decoration' flag to ignore 'electron pairs',
  boxes and brackets
- the 'half arrow' bond type is now ignored
  * TODO : _isolated_ regular arrows should be ignored as well 

1.9 (mk)

- added support for latex-like brace syntax in subscripts (e.g. C_{10}H_{22})
- added support for two-digit subscripts (C_1_0 was interpreted as C_1)

1.8 (mk)

- interactive mode added, one can now call cht with '-i' followed by
  a sum formula in chemtool notation, e.g. 'cht -i C_6H_5_NH_2' or
  'cht -i C_1_2H_2_5OH'

1.7

- Mg added
- small changes in output formatting
 
1.6

- a nasty bug in the label parsing routine was removed (that which was 
  fetched in during the code cleanup in version 1.4)
   
1.5

- wrong mass for B. All values checked with tabular ones.
- composition is calculated (format: 00.00%C;00.00%H;00.00%N; ... )

1.4

- added test for bonds overlap. If some bonds are overlapped (within the
  interval gave by "Sensi" parameter; 5 points by default) then the
  exclamation mark precede the summary formula (one "!" for each such overlap)
- only the "top label" of overlapped labels is taken into calculation

1.1 to 1.3:

- added support for bond type 11 (aromatic ring), Martin's patch.
- Ac is CH_3CO, not CH_3 (uff!)
- removed bug in the label parser (labels XxxYyy were interpreted as XxxY, e.g.
  --C(Me_2BuSiO)_3 were interpreted as --C_20H_48O_3S_3 - note S instead of Si)
- removed bug in the label parser (labels XXXXX??? bring cht into infinite loop,
  e.g. "Guo" instead of "Gua", NE instead of NEt)
- Readln(file,X1,X2,Y1,Y2 ...) used, instead of Readln(file,string); no 
  Val(Copy(Pos........))) construction needed.
- Now the mass of the label(s) is calculated, even if no bond has been drawn 
  (however when some bonds have been drawn, then the labels which are not con-
  nected to the structure are ignored - it is good for the caption(s) not to be
  involved into calculation)