🚀 go-pugleaf

Somewhere along the line, a patch ( no way to tell which one ) added
three lines too many.

Since I've run into the problem, I suspect others may also have
inherited this little problem.  What follows are two version of the
histdups awk script.  The faulty one first then the working one.  If
any one recognizes the patch that produces this result, I'd sure like
to know as it may have introduced other problems.

Here's the faulty histdups ( note the extra print line at the end ):

# Awk program to merge history lines for the same article in a sorted history
# file (such as is generated during the mkhistory processing).
BEGIN { FS = "\t" ; OFS = "\t" ; mesgid = "" }
{
	if ($1 != mesgid) {
		if (mesgid != "")
			print mesgid, dates, names
		mesgid = $1
		dates = $2
		names = $3
	} else
		names = names " " $3
}
END {
	if (mesgid != "")
		print mesgid, dates, names
}
	print mesgid, dates, names
}


Now my known working version (pre-patch):

# Awk program to merge history lines for the same article in a sorted history
# file (such as is generated during the mkhistory processing).
BEGIN { FS = "\t" ; OFS = "\t" ; mesgid = "" }
{
	if ($1 != mesgid) {
		if (mesgid != "")
			print mesgid, dates, names
		mesgid = $1
		dates = $2
		names = $3
	} else
		names = names " " $3
}
END { if (mesgid != "")
	print mesgid, dates, names
}


If memory serves me right,  there was a patch that addressed duplicates
some time ago.  It was not official (i.e. from geoff or henry).  I can't
find any such patch in my cnews source tree but I wonder.

ger
-- 
G. Roderick Singleton, System and Network Manager, JTS Computers 
{yunexus | uunet | geac | torsqnt}!gerry@jtsv16.jts.com