handle larger data sets - I think

README.md

@@ -8,10 +8,6 @@ A Go implementation of Hochbaum's PseudoFlow algorithm as [implemented here in C
 
 An example command-line program is cmd/pseudo/main.go.
 
-<h2>Status</h2>
-
-Release 2.2.  (Note: large data sets like "_data/BVZ-tsukuba0.max" are VERY MEMORY INTENSIVE and cannot be run on 8 GB machines; for that data set you will need over 128 GB of memory.)
-
 <h2>Prior Versions</h2>
 
 The original port of the C source code is in the subdirectory [v1.2][v1.2]. It is used in an example command-line program cmd/pseudo/main_v1.2.go.  It is fine for command-line or utility programs, but it is not safe for concurrent use in a server.

pseudo.go

@@ -250,7 +250,7 @@ type node struct {
 func (s *Session) newNode(number uint) *node {
 	return &node{
 		number:    number,
-		outOfTree: make([]*arc, int(s.numArcs)),
+		// outOfTree: make([]*arc, int(s.numArcs)),
 	}
 }