Skip to content

Commit

Permalink
[3.11] pythongh-93575: Use correct way to calculate PyUnicode struct …
Browse files Browse the repository at this point in the history
…sizes (pythonGH-93602)

* pythongh-93575: Use correct way to calculate PyUnicode struct sizes

* Add comment to keep test_sys and test_unicode in sync

* Fix case code < 256.
(cherry picked from commit 5442561)

Co-authored-by: Christian Heimes <christian@python.org>
  • Loading branch information
tiran committed Jun 8, 2022
1 parent abd0223 commit 93c34da
Show file tree
Hide file tree
Showing 3 changed files with 25 additions and 11 deletions.
1 change: 1 addition & 0 deletions Lib/test/test_sys.py
Original file line number Diff line number Diff line change
Expand Up @@ -1538,6 +1538,7 @@ class newstyleclass(object): pass
samples = ['1'*100, '\xff'*50,
'\u0100'*40, '\uffff'*100,
'\U00010000'*30, '\U0010ffff'*100]
# also update field definitions in test_unicode.test_raiseMemError
asciifields = "nnbP"
compactfields = asciifields + "nPn"
unicodefields = compactfields + "P"
Expand Down
31 changes: 20 additions & 11 deletions Lib/test/test_unicode.py
Original file line number Diff line number Diff line change
Expand Up @@ -2370,20 +2370,19 @@ def test_expandtabs_optimization(self):
self.assertIs(s.expandtabs(), s)

def test_raiseMemError(self):
if struct.calcsize('P') == 8:
# 64 bits pointers
ascii_struct_size = 48
compact_struct_size = 72
else:
# 32 bits pointers
ascii_struct_size = 24
compact_struct_size = 36
asciifields = "nnbP"
compactfields = asciifields + "nPn"
ascii_struct_size = support.calcobjsize(asciifields)
compact_struct_size = support.calcobjsize(compactfields)

for char in ('a', '\xe9', '\u20ac', '\U0010ffff'):
code = ord(char)
if code < 0x100:
if code < 0x80:
char_size = 1 # sizeof(Py_UCS1)
struct_size = ascii_struct_size
elif code < 0x100:
char_size = 1 # sizeof(Py_UCS1)
struct_size = compact_struct_size
elif code < 0x10000:
char_size = 2 # sizeof(Py_UCS2)
struct_size = compact_struct_size
Expand All @@ -2395,8 +2394,18 @@ def test_raiseMemError(self):
# be allocatable, given enough memory.
maxlen = ((sys.maxsize - struct_size) // char_size)
alloc = lambda: char * maxlen
self.assertRaises(MemoryError, alloc)
self.assertRaises(MemoryError, alloc)
with self.subTest(
char=char,
struct_size=struct_size,
char_size=char_size
):
# self-check
self.assertEqual(
sys.getsizeof(char * 42),
struct_size + (char_size * (42 + 1))
)
self.assertRaises(MemoryError, alloc)
self.assertRaises(MemoryError, alloc)

def test_format_subclass(self):
class S(str):
Expand Down
Original file line number Diff line number Diff line change
@@ -0,0 +1,4 @@
Fix issue with test_unicode test_raiseMemError. The test case now use
``test.support.calcobjsize`` to calculate size of PyUnicode structs.
:func:`sys.getsizeof` may return different size when string has UTF-8
memory.

0 comments on commit 93c34da

Please sign in to comment.