This issue tracker has been migrated to GitHub, and is currently read-only.
For more information, see the GitHub FAQs in the Python's Developer Guide.

Author dankreso
Recipients dankreso
Date 2019-12-12.11:15:23
SpamBayes Score -1.0
Marked as misclassified Yes
Message-id <1576149324.34.0.405331329456.issue39030@roundup.psfhosted.org>
In-reply-to
Content
I've found what looks like a corner case bug. Specifically, the behaviour that looks suspicious is when a ctypes union has bit field members, where the members have bit widths that are smaller than the size types:

class BitFieldUnion(Union):
    _fields_ = [("a", c_uint32, 16), ("b", c_uint32, 16)]

buff = bytearray(4)

bitfield_union = BitFieldUnion.from_buffer(buff)
bitfield_union.a = 1
bitfield_union.b = 2

print("a is {}".format(bitfield_union.a)) # Prints "a is 1"
print("b is {}".format(bitfield_union.b)) # Prints "b is 2"
print("Buffer: {}".format(buff)) # Prints "Buffer: b'\x01\x00\x00\x00'".

(Example of this script can be found at https://rextester.com/XJFGAK37131. I've also tried it on my system which is Ubuntu 16.04.2 LTS with Python 3.6.)

Here I would expect both 'a' and 'b' to be set to 2, and for the buffer to look like '\x02\x00\x00\x00'. Here's the equivalent code in C which behaves as expected: https://rextester.com/HWDUMB56821.

If at least one of the bitwidths in the above example are changed from 16 to 32, however, then 'a', 'b', and the buffer look as expected.

I've also attached some further examples of weird behaviour with unions with bitfield members - online version can be found at https://rextester.com/VZRB77320.
History
Date User Action Args
2019-12-12 11:15:24dankresosetrecipients: + dankreso
2019-12-12 11:15:24dankresosetmessageid: <1576149324.34.0.405331329456.issue39030@roundup.psfhosted.org>
2019-12-12 11:15:24dankresolinkissue39030 messages
2019-12-12 11:15:23dankresocreate