Remaining memory leaks

[Alex Kiernan]

This is the last one I had left over from the last debugging session:

New memory leaked: 331008 bytes (2.91%); potentially leaked: 0 bytes (0%)

MLK: 331008 bytes leaked at 0x1890048
  * This memory was allocated from:
	malloc         [rtlib.o]
	realloc        [rtlib.o]
	x_realloc      [xmalloc.c:108]
	BUFFappend     [chan.c:74]
	NClist         [nc.c:612]
	NCproc         [nc.c:844]

I'm going to try removing the tests against "cp->{In,Out}.size >
BIG_BUFFER" at the tail end of CHANclose to see if thats how they're
getting leaked (since there's places other than the main channel table
which do "CHANclose(p); p = NULL;" kind of stuff).

Anyone any thoughts?

-- 
Alex Kiernan, Principal Engineer, Development, Thus PLC