WHAT'S NEW:

1) Indexing: the pretty-printers now fix up some of the deficiencies of noweb's
indexing mechanism.

"finduses" generates spurious references to identifiers when it sees a string
matching an identifier inside a comment or string literal.  For example, try

	noweave -autodefs c -index t.nw  > t.tex
	latex t
	latex t

on the file

<<Section A>>=
int num;
<<Section B>>=
/* num is the number ... */
<<Section C>>=
printf("num is the number of things\n");
@
\nowebindex

noweave (finduses) will generate a reference to num in each of sections B and C.

Now try 

	noweave -autodefs c -filter C.filter -index t.nw  > t.tex
	latex t
	latex t

(NOTE: -filter first, -index next!)
You get a pretty-printed version without these references.  All pretty-printer
filters suppress these spurious references.


2) A bug in pretty-printing operators has been fixed.  For example, if ``=='' is
meant to be typeset as ``\equiv'', say, but ``==='' is supposed to be left alone, the
old version would typeset ``==='' as ``\equiv=''.


3) For those who know about OOT: the autodefs.oot now does not index identifiers
declared in subprograms.  This reduces clutter.  However, it has the capability to
index such identifiers selectively if the user supplies an argument to the filter,
the name of a file listing subprogram names one per line.  (See ootdefs.nw for
details.)