rsync - Jan 2004 - How match.c hash_search works with multiple checksums that have identical tags

I am trying to understand how match.c works. 

I am reading the code and something doesnt look quite right. This is usually a
sign that I am missing something obvious.

Here is what I see.

build_hash_table uses qsort to order targets in ascending order of //tag,index//
into the array of checksums.
It then accesses the targets in ascending order and writes the index at the
tag's location in the tag_table. For any
set of targets with identical tags, the highest valued index is the resulting
value in that location in the table.

hash_search initializes "j" from the tag_table resulting in
"j" being the largest index in a group of targets with the same
tag value. The for loop that uses "j" as its primary iteration control
iterates positively (ie. j++). The termination condition
for the loop is any tag value that isn't equal to the initialized tag value.

It appears to me that only the last target of a group that has an identical tag
value will be processed.

What am I missing?? 

wally

On Mon, Jan 26, 2004 at 11:28:44AM -0500, Wallace Matthews
wrote:
> For any set of targets with identical tags, the highest valued index
> is the resulting value in that location in the table.
No, it's the lowest index because the loop is run in reverse:

	for (i = s->count; i-- > 0; )
                tag_table[targets[i].t] = i;

Thus, the loop that scans the targets with the j++ increment is handling
things correctly.

..wayne..