2005-4-16

The illegal jump to for-loop in emx.f has been fixed (masked 
with -w Makevars but gfortran -w -fPIC -O2 -g -c picks it up).

2005-3-23

To apply -Wall -pedantic to emx.f, nucfama.f, pointr.f.


2005-3-18

add control.pointer function and reflect in INDEX

emx.f:
. pass ascertainmentr selection as strings

nucfama.f:
. add err= in backsapce

pointer.R:
. unlink junkfile as wild card fails under Windows


Note that when added a blank line to example file poijob,
.Fortran("nucfama") would work under Windows, hopefully
this problem can now be dealt with by backspace(err=).


2005-3-16

Redo with successful testing of the docuementation example

emx.f:

. warnings in emx.f to be fixed

nucfama.f:

. add 'n' (to differentiate from pointr.f):
  nspace, nsp1, odds, par, xno, xno1, xno2
. odds, reorder
. /add/ block and reorder x(60)
. /pets/ block size
. program --> subroutine, comment getarg, stop --> return

pointr.f:

. reorder /para/, /auxp1/, /odds/, /intg/, /mini/, /data/, /data2/
. debug --> idebug
. add three commas (,) to =e18.11, and two 1h;4x

The changes required for emx.f, nucfama.f and pointr.f
for stand-alone executables under Windows

. comment subroutine statement
. uncomment getarg
. return --> stop

However, it remains to work out under Windows. A way out is
to use the Cygwin program obtained in 2001, or include this R
package in a bundle with R/gap.

Other things to implement relate to split=T, e.g.,

 MATING TYPE SELECTION(00,01,02,10,11,12,20,21,22)?
 POINTER SELECTION(0,4,5,6)?
 POINTER DEGREE SELECTION(1,2,3,4,5)?
 ASCERTAINMENT COMPLETE(C) OR MULTIPLE INCOMPLETE(M)?

along with some warnings in the compiling, but thse may not 
be a priority. 


2005-3-15
first pass.

