Repository navigation
sqlite3.Blob indexing is not conformant with existing norms #92019
Description
Yesterday, @JelleZijlstra enlightened me that sqlite3.Blob indexing operate with bytes objects of length 1. This is a deviation from how bytes, bytearray, and memoryview work:
>>> bytes(b"0")[0]
48
>>> bytearray(b"0")[0]
48
>>> memoryview(b"0")[0]
48
>>> b = bytearray(b"0")
>>> b[0] = 1
>>> b[0] = b"1"
Traceback (most recent call last):
File "<stdin>", line 1, in <module>
TypeError: 'bytes' object cannot be interpreted as an integer
>>> m = memoryview(b)
>>> m[0] = 1
>>> m[0] = b"1"
Traceback (most recent call last):
File "<stdin>", line 1, in <module>
TypeError: memoryview: invalid type for format 'B'
>>>
Suggesting to align sqlite3.Blob indexing behaviour so it expects ints and returns ints.
Activity
pythongh-92019: Make sqlite3.Blob indexing conform with the norm
JelleZijlstra commentedon Apr 28, 2022
Credit to @Akuli who brought this up in python/typeshed#7684 (comment)
serhiy-storchaka commentedon Apr 29, 2022
There is also a similar deviation in
mmap.JelleZijlstra commentedon Apr 29, 2022
We probably can't fix that though, since mmap has already been in the stdlib for a long time.
gh-92019: Make sqlite3.Blob indexing conform with the norm (#92020)