Improve tokenization performance by removing a single character.

strip() instead of lstrip() makes the loop terminate earlier, saving
lots of calls. Shaves a few % of parsing the x86 Kconfigs.

1 files changed, 5 insertions, 1 deletions

diff --git a/kconfiglib.py b/kconfiglib.py
index fbeaad5..4679126 100644
--- a/kconfiglib.py
+++ b/kconfiglib.py
@@ -614,7 +614,11 @@ class Config(object):
         for_eval -- True when parsing an expression for a call to
                     Config.eval(), in which case we should not treat the first
                     token specially nor register new symbols."""
-        s = s.lstrip()
+
+        # lstrip() would work here too, but removing the '\n' at the end leads
+        # to earlier termination in the 'while' loop below, saving lots of
+        # calls
+        s = s.strip()
         if s == "" or s[0] == "#":
             return _Feed([])