August 22, 1997

	D.Jeff Dionne	Started moving things around to split things into
			libspice and driver code.

August 13, 1997

	D.Jeff Dionne	Cleaned up the memory management even more.  Now
			uses the OS' malloc() free() and realloc().  This
			change makes spice work perfectly on Linux and
			Win32 with cygwin32.

			Cleaned up readin.f to properly allocate memory for
			new elements.  It used to make assumptions about
			the physical layout of memory blocks!  Removed an
			array of pointers to circuit elements that was never
			used and caused problems when blocks move around.

March 16, 1996

	D.Jeff Dionne	Completely re-wrote the memory manager from scratch
			in C.  _MUCH_ better this way (understandable too!).
			Spice will now process the entire benchmark.in
			without any trouble, with or without output to rawfile.

		ToDo:	Split the source tree up into subdirectories.
			Clean up the new memory stuff a little (there are
				still a few hacks in there).
			Figure out what is wrong with my use of sbrk().  Just
				using malloc() for now.
			It appears there may be something STILL funny with
				rawfiles and time on dctran :-(

Jan 27, 1996

	D.Jeff Dionne	Added a new field to readin.f so we can tell what
			type of analysis this is.  See readin.f and dctran.f
			for comments.  Updated spiceparse to properly read
			the new format in.  Did not bump the X in rawfileX 
			key though, since this stuff has not yet been
			released officially.

Jan 25, 1996

	D.Jeff Dionne	Fixed readin.f to allow -ve values for capacitors
			and inductors.  This allows some op-amp models from
			ADI to work.

		ToDo:	Test all possible analysis to make sure that -ve
			values for these does not cause any problems.

Jan 23, 1996

	D.Jeff Dionne	Created (and put stuff in) doc and models dirs.

Jan 21, 1996

	D.Jeff Dionne	Removed really stupid and improperly implemented
			check for free memory from dctran.f.  As if checking
			was going to prevent an error or as if not checking
			would let one go un-noticed! (even if it had been
			implemented correctly to check real space insted of
			managed core).  Removed avlm4.f and redundant 
			friends, which fixed a random out of memory bug.

Jan 20, 1996

	D.Jeff Dionne	Cleaned up more of the rawfile code.  Removed
			what looks like arbitrary closing of the rawfile
			at odd times, and added a magic string at the
			begining of a new analysis so that we can tell
			where one ends and a new one begins.

		ToDo:	Need to add a field to the header in the rawfile
			so we can tell what type of analysis this is!

Jan 19, 1996

	D.Jeff Dionne	Arrgh!  Fixed a serious bug left over from before me,
			with rawfile code, in *pheadr*.  After finding some
			more pristine 2g6 code than the Linux port of the
			BSD port, found interger*2 was incorrectly changed to 
			interger, causing memory corruption with the
			original memory array in the BSD port, and wild
			indexes causing segfaults with the new dynamic 
			allocation code.  Had to put it back the way 
			Berkeley intended :-)  Unfortunately, this breaks 
			g77 again...

		ToDo:	Should report this bug, or at least get the bug
			fixed in the 'standard' distributions until we're
			ready for prime time.

Jan 18, 1996

	D.Jeff Dionne.	Added dynamic memory allocation.  Cleaned up
			various bits all through the code, some dead
			code removed.

			Simple memory debugging added unix.c

		Note:	There is something funny with *readin* that prevents
			the first allocated block from being moved.

Jan 6, 1996

	D.Jeff Dionne.	Re organized the directory tree.

			Included spiceparse, however it has just begun.
			SpiceParse is to become a GUI shell and post 
			processor.  In the near term, it will output
			to octave (matlab) 'm' file.

Dec 29, 1995

	D.Jeff Dionne.  Re-wrote ushift.f in C and put it in unix.c
			Also removed none portable VAX asm code.

Dec 23, 1995

	D.Jeff Dionne.  'Bugs' fixed that break compile on g77.

From standard Linux distribution of the BSD port.
