Index: /branches/eam_branches/20091201/psphot/doc/footprints.txt
===================================================================
--- /branches/eam_branches/20091201/psphot/doc/footprints.txt	(revision 26745)
+++ /branches/eam_branches/20091201/psphot/doc/footprints.txt	(revision 26746)
@@ -1,2 +1,20 @@
+
+2010.01.18
+
+  * we are having some problems with footprints being freed to early in the second pass
+
+  * footprint memory history:
+
+    * entering psphotFindFootprints, on the first pass, we have an
+      array of detected peak in detections->peaks
+
+    * in pmFootprintsFind, footprints array owns allocated footprints
+      footprints each own their spans
+
+    * from pmFootprintAssignPeaks:
+      * the footprints have an array of peaks with valid memory references
+      * the peaks have just the pointer to the corresponding footprint (not owned)
+      * all peaks get assigned to a footprint
+      * at end of pmFootprintAssignPeaks, only a single copy of each peak is/should be available from the collection of footprints
 
 pmFootprintCullPeaks is very expensive (15 - 30 msec per object) and
