pythongh-95865: Further reduce quote_from_bytes memory consumption (python#96860)

on large input values. Based on Dennis Sweeney's chunking idea.

Author: gpshead

Lib/test/test_urlparse.py

@@ -985,6 +985,10 @@ def test_quote_from_bytes(self):
  self.assertEqual(result, 'archaeological%20arcana')
  result = urllib.parse.quote_from_bytes(b'')
  self.assertEqual(result, '')
+ result = urllib.parse.quote_from_bytes(b'A'*10_000)
+ self.assertEqual(result, 'A'*10_000)
+ result = urllib.parse.quote_from_bytes(b'z\x01/ '*253_183)
+ self.assertEqual(result, 'z%01/%20'*253_183)
 
  def test_unquote_to_bytes(self):
  result = urllib.parse.unquote_to_bytes('abc%20def')

Lib/urllib/parse.py

@@ -29,6 +29,7 @@
 
 from collections import namedtuple
 import functools
+import math
 import re
 import types
 import warnings
@@ -906,7 +907,14 @@ def quote_from_bytes(bs, safe='/'):
  if not bs.rstrip(_ALWAYS_SAFE_BYTES + safe):
  return bs.decode()
  quoter = _byte_quoter_factory(safe)
- return ''.join(map(quoter, bs))
+ if (bs_len := len(bs)) < 200_000:
+ return ''.join(map(quoter, bs))
+ else:
+ # This saves memory - https://github.com/python/cpython/issues/95865
+ chunk_size = math.isqrt(bs_len)
+ chunks = [''.join(map(quoter, bs[i:i+chunk_size]))
+ for i in range(0, bs_len, chunk_size)]
+ return ''.join(chunks)
 
 def urlencode(query, doseq=False, safe='', encoding=None, errors=None,
  quote_via=quote_plus):

Misc/NEWS.d/next/Library/2022-09-16-07-53-29.gh-issue-95865.oHjX0A.rst

@@ -0,0 +1,3 @@
+Reduce :func:`urllib.parse.quote_from_bytes` memory use on large values.
+
+Contributed by Dennis Sweeney.