Slicing a bytes object returning <int> with single index

[maxpat78]

Bug report

Bug description:

Consider the first 'if' in the following piece of code.
It used to run nice in Python 2.x (where it was developed):

	if n_offset == 1: # ...BYTE
		offset = struct.unpack_from("<b", offset[0])[0] # offset[0]
	elif n_offset == 2: # ...WORD
		offset = struct.unpack_from("<h", offset[:2])[0]
	elif n_offset in (3,4):  # ...DWORD
		offset = struct.unpack_from("<i", offset[:4])[0]
	else: # ...QWORD
		offset = struct.unpack_from("<q", offset)[0]

Actually, it raises the exception TypeError: a bytes-like object is required, not 'int' and needs an awkward reconversion:

offset = struct.unpack_from("<b", offset[0].to_bytes())[0] # offset[0] -> int ! Must convert back!

Isn't it a contradiction (and a bug) that slicing returns bytes except in the case of a single slicing index???

CPython versions tested on:

3.13

Operating systems tested on:

Windows

[TeamSpen210]

It’s not actually, strings are the weird ones. Consider a list - mylist[4] gives the 5th element, but mylist[4:5] gives a 1-element list. Bytes objects are conceptually a sequence of 0-255 integers. If you do want 1 element, do a slice like offset[0:1].

[skirpichev]

Isn't it a contradiction (and a bug) that slicing returns bytes except in the case of a single slicing index???

Are you about this:

offset = struct.unpack_from("<b", offset[0])[0] # offset[0]

?

It's not slicing, but subscription.

[maxpat78]

Call it how you like, indexing, subscription or slicing-of-one-item: but one expects that what you extract is what there is inside an object - bytes -> byte

[skirpichev]

but one expects that what you extract is what there is inside an object - bytes -> byte

Python works here as expected, similar you can see just with ordinary lists:

>>> xs = [1, 2, 3]
>>> type(xs[0])
<class 'int'>
>>> type(xs[0:1])
<class 'list'>

Problem is definitely with your expectations.

[sergey-miryanov]

It is documented behavior since 3.3:

Since bytes objects are sequences of integers, for a bytes object b, b[0] will be an integer, while b[0:1] will be a bytes object of length 1. (This contrasts with text strings, where both indexing and slicing will produce a string of length 1)

[skirpichev]

In fact, that was documented since beginning in py3, see 98297ee.