CVS difference for ai05s/ai05-0212-1.txt

--- ai05s/ai05-0212-1.txt	2011/03/17 02:58:31	1.8
+++ ai05s/ai05-0212-1.txt	2011/03/17 03:01:59	1.9
@@ -1,4 +1,4 @@
-!standard A.18.2(9/2)                                  11-03-08  AI05-0212-1/06
+!standard A.18.2(9/2)                                  11-03-15  AI05-0212-1/07
 !standard A.18.33(0)
 !class Amendment 10-04-26
 !status work item 10-04-26
@@ -1499,5 +1499,41 @@
 the algorithm itself!
 
 [Editor's note: This is version /07 of the AI.]
+
+****************************************************************
+
+From: Tucker Taft
+Sent: Tuesday, March 15, 2011  2:45 PM
+
+The line below could be simplified to "for N in Node loop"
+for what it's worth.  Actually seems a little clearer to me to leave off the
+'First/'Last, since you use simply "Node" as the index subtype when declaring
+the arrays.
+
+>          for N in Node'First .. Node'Last loop
+
+****************************************************************
+
+From: Robert Dewar
+Sent: Tuesday, March 15, 2011  2:53 PM
+
+> Here is a revised version of Dijkstra's shortest path algorithm.  It
+> is simpler than the previous version, (it even uses anonymous array
+> types where appropriate) and only makes use once of the new iterator
+> form. It compiles properly with the partial implementation of
+> iterators in GNAT, and the code incorporates the suggestions made at
+> last meeting, including the precondition that the source node is
+> connected to the given graph.  a proper postcondition seems out of
+> reach: describing the set of possible paths would be longer than the
+> algorithm itself!
+
+Where are power sets when you need them? :-) :-)
+
+****************************************************************
+
+From: Ed Schonberg
+Sent: Tuesday, March 15, 2011  2:59 PM
+
+Of course, this was left-over from an earlier version. Will update
 
 ****************************************************************