2to3 doesn't handle byte comparison well

[vdupras]

BPO	9043
Nosy	@terryjreedy, @benjaminp

^{Note: these values reflect the state of the issue at the time it was migrated and might not reflect the current state.}

Show more details

GitHub fields:

assignee = None
closed_at = <Date 2010-06-21.14:59:11.300>
created_at = <Date 2010-06-21.14:46:13.908>
labels = ['type-bug', 'expert-2to3']
title = "2to3 doesn't handle byte comparison well"
updated_at = <Date 2010-06-25.21:19:57.080>
user = 'https://bugs.python.org/vdupras'

bugs.python.org fields:

activity = <Date 2010-06-25.21:19:57.080>
actor = 'terry.reedy'
assignee = 'none'
closed = True
closed_date = <Date 2010-06-21.14:59:11.300>
closer = 'benjamin.peterson'
components = ['2to3 (2.x to 3.x conversion tool)']
creation = <Date 2010-06-21.14:46:13.908>
creator = 'vdupras'
dependencies = []
files = []
hgrepos = []
issue_num = 9043
keywords = []
message_count = 3.0
messages = ['108272', '108273', '108633']
nosy_count = 3.0
nosy_names = ['terry.reedy', 'benjamin.peterson', 'vdupras']
pr_nums = []
priority = 'normal'
resolution = 'wont fix'
stage = 'resolved'
status = 'closed'
superseder = None
type = 'behavior'
url = 'https://bugs.python.org/issue9043'
versions = ['Python 2.7']

[vdupras]

If we run 2to3 on the following code:

s = b' '
print s[0] == s

we end up with this:

s = b' '
print(s[0] == s)

However, the first code, under python2 prints True while the converted code, under python3 prints False.

Shouldn't 2to3 convert this code to:

s = b' '
print s[0:1] == s

instead?

[benjaminp]

2to3 can't guess which slices are from bytes.

[terryjreedy]

Replacing string indexes with length-1 slices in the 2.x codebase, whether or not one one is running 2to3, seems to be a 'known' technique for Python3 portability and unicode-bytes inter-operability. (It was discussed this week on py-dev list .) But I do not see it mentioned in PyWiki, such as on
http://wiki.python.org/moin/PortingPythonToPy3k