Performance boost: reduce instantiations of re.Scanner

- Lexer is the same for every Spec parser in spack, so don't build it
  every time.

- This improves time to import package.py files a lot, as a Lexer
  doesn't have to be constructed for every spc in the packages.

- To concretize dealii:
  - Before: ~20 sec
  - After:  ~6 sec

lib/spack/spack/spec.py

@@ -2015,10 +2015,13 @@ def __init__(self):
             (r'\s+', lambda scanner, val: None)])
 
 
+# Lexer is always the same for every parser.
+_lexer = SpecLexer()
+
 class SpecParser(spack.parse.Parser):
 
     def __init__(self):
-        super(SpecParser, self).__init__(SpecLexer())
+        super(SpecParser, self).__init__(_lexer)
         self.previous = None
 
     def do_parse(self):