Remove "primes-under-million" list

In the past, this was a static list used to optimize prime factors
computation. Now that the dependency has been reversed, there is
no point in keeping this list which can be obtained by
"1000000 primes-upto" as easily.

extra/benchmark/binary-search/binary-search.factor

@@ -1,13 +1,17 @@
 ! Copyright (C) 2008 Slava Pestov.
 ! See http://factorcode.org/license.txt for BSD license.
-USING: binary-search kernel math.primes.list math.ranges sequences
-prettyprint ;
+USING: binary-search compiler.units kernel math.primes math.ranges
+memoize prettyprint sequences ;
 IN: benchmark.binary-search
 
-: binary-search-benchmark ( -- )
-    1 1000000 [a,b] [ primes-under-million sorted-member? ] map length . ;
+[
+    MEMO: primes-under-million ( -- seq ) 1000000 primes-upto ;
+] with-compilation-unit
 
 ! Force computation of the primes list before benchmarking the binary search
 primes-under-million drop
 
+: binary-search-benchmark ( -- )
+    1 1000000 [a,b] [ primes-under-million sorted-member? ] map length . ;
+
 MAIN: binary-search-benchmark

extra/math/primes/list/authors.txt

@@ -1 +0,0 @@
-Doug Coleman

extra/math/primes/list/list.factor

@@ -1,4 +0,0 @@
-USING: math.primes memoize ;
-IN: math.primes.list
-
-MEMO: primes-under-million ( -- seq ) 1000000 primes-upto ;