MIR4

Notes on what was done to miriad for the MIR4 transition:

MIR42 notes: (June 2011)
========================

- large fits files (8+ GB) 
- new memory managment (mm)
- new ptrdif (close to an earlier WSRT version using prtdiff_t) for mm
- mosaicing changes
- WCSLIB integration
- (more to come)


MIR41 notes:
============

intf2c.c	longinteger in .f2c files
		int8 is used, and become 'int' or 'long long int'
		depending on -DMIR4 (via miriad.h)

scrio.c		scrread/scrwrite are now using int8, which again
		via -DMIR4 will become int vs. long long int


mapper.for	use integer*8  offset8, length8
model.for	use integer*8  offset8, length8

		It is quite possible, since other integer variables are used to compute
		the offset8 and length8, that UV files cannot be arbitrary large.

		Probably important to note that we decided to keep ncorr and nwcorr 
		as integer*4

		this will be quite hairy to figure out


problem cases:


gpcal.for	arrays are not addressed in main / compareg, complicating
		figuring out sizing w.r.t. MAXANT
		16	26870912
		32	27141248
		64	28801152
		128	34598528
		256	57661568