From 89e8b31bb23412c9d601cb50e83fa582cb921141 Mon Sep 17 00:00:00 2001
From: kgriffs <kurt.griffiths@rackspace.com>
Date: Tue, 14 Jan 2014 14:59:30 -0600
Subject: [PATCH] perf: Further optimizations for falcon.uri

---
 falcon/bench/bench.py |  3 ++-
 falcon/util/misc.py   |  4 +---
 falcon/util/uri.py    | 29 ++++++++++++++++-------------
 3 files changed, 19 insertions(+), 17 deletions(-)

diff --git a/falcon/bench/bench.py b/falcon/bench/bench.py
index 298a5c4..b38296b 100755
--- a/falcon/bench/bench.py
+++ b/falcon/bench/bench.py
@@ -128,7 +128,8 @@ def queues_env():
     path = ('/v1/852809/queues/0fd4c8c6-bd72-11e2-8e47-db5ebd4c8125'
             '/claims/db5ebd4c8125')
 
-    return helpers.create_environ(path, query_string='limit=10&thing=a%20b',
+    qs = 'limit=10&thing=a%20b&x=%23%24'
+    return helpers.create_environ(path, query_string=qs,
                                   headers=request_headers)
 
 
diff --git a/falcon/util/misc.py b/falcon/util/misc.py
index 85d0879..bb57794 100644
--- a/falcon/util/misc.py
+++ b/falcon/util/misc.py
@@ -104,7 +104,7 @@ def http_date_to_dt(http_date):
 
 
 def to_query_str(params):
-    """Converts a dict of params to afaln actual query string.
+    """Converts a dict of params to an actual query string.
 
     Args:
         params: dict of simple key-value types, where key is a string and
@@ -129,8 +129,6 @@ def to_query_str(params):
         elif v is False:
             v = 'false'
         elif isinstance(v, list):
-            # PERF(kgriffs): map is faster than list comprehension in
-            # py26 and py33. No significant different in py27
             v = ','.join(map(str, v))
         else:
             v = str(v)
diff --git a/falcon/util/uri.py b/falcon/util/uri.py
index aa5e6c2..ec9b895 100644
--- a/falcon/util/uri.py
+++ b/falcon/util/uri.py
@@ -190,21 +190,20 @@ if six.PY2:  # pragma: no cover
             # do, let's encode in a non-lossy format.
             decoded_uri = decoded_uri.encode('utf-8')
 
-        # PERF(kgriffs): Use a closure instead of a class.
-        only_ascii = [True]
+        only_ascii = True
 
-        def unescape(matchobj):
-            # NOTE(kgriffs): Strip '%' and convert the hex number
-            char, byte = _HEX_TO_BYTE[matchobj.group(0)[1:]]
-            only_ascii[0] = only_ascii[0] and (byte <= _UTF8_MAX)
+        tokens = decoded_uri.split('%')
+        decoded_uri = tokens[0]
+        for token in tokens[1:]:
+            char, byte = _HEX_TO_BYTE[token[:2]]
+            decoded_uri += char + token[2:]
 
-            return char
-
-        decoded_uri = _ESCAPE_SEQUENCE.sub(unescape, decoded_uri)
+            if only_ascii:
+                only_ascii = (byte <= 127)
 
         # PERF(kgriffs): Only spend the time to do this if there
         # were non-ascii bytes found in the string.
-        if not only_ascii[0]:
+        if not only_ascii:
             decoded_uri = decoded_uri.decode('utf-8', 'replace')
 
         return decoded_uri
@@ -253,10 +252,14 @@ else:  # pragma: no cover
         # do, let's encode into a non-lossy format.
         decoded_uri = decoded_uri.encode('utf-8')
 
-        # Replace escape sequences
-        decoded_uri = _ESCAPE_SEQUENCE.sub(_unescape, decoded_uri)
+        # PERF(kgriffs): This was found to be faster than using
+        # a regex sub call or list comprehension with a join.
+        tokens = decoded_uri.split(b'%')
+        decoded_uri = tokens[0]
+        for token in tokens[1:]:
+            decoded_uri += _HEX_TO_BYTE[token[:2]] + token[2:]
 
-        # Back to str
+        # Convert back to str
         return decoded_uri.decode('utf-8', 'replace')