ndiff reports incorrect location when diff strings contain tabs

[asottile]

BPO	37810
Nosy	@tim-one, @rhettinger, @asottile
PRs	bpo-37810: Fix difflib ? hint in diff output when dealing with tabs #15201 [3.8] bpo-37810: Fix difflib ? hint in diff output when dealing with tabs (GH-15201) #15365

^{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 = 'https://github.com/tim-one'
closed_at = <Date 2019-08-21.20:26:10.221>
created_at = <Date 2019-08-10.18:59:04.158>
labels = ['3.8', 'library', '3.9']
title = 'ndiff reports incorrect location when diff strings contain tabs'
updated_at = <Date 2019-08-21.20:26:10.220>
user = 'https://github.com/asottile'

bugs.python.org fields:

activity = <Date 2019-08-21.20:26:10.220>
actor = 'tim.peters'
assignee = 'tim.peters'
closed = True
closed_date = <Date 2019-08-21.20:26:10.221>
closer = 'tim.peters'
components = ['Library (Lib)']
creation = <Date 2019-08-10.18:59:04.158>
creator = 'Anthony Sottile'
dependencies = []
files = []
hgrepos = []
issue_num = 37810
keywords = ['patch']
message_count = 3.0
messages = ['349353', '349354', '349391']
nosy_count = 3.0
nosy_names = ['tim.peters', 'rhettinger', 'Anthony Sottile']
pr_nums = ['15201', '15365']
priority = 'normal'
resolution = 'fixed'
stage = 'resolved'
status = 'closed'
superseder = None
type = None
url = 'https://bugs.python.org/issue37810'
versions = ['Python 3.8', 'Python 3.9']

[asottile]

Here's an example

from difflib import ndiff


def main():
    x = '\tx\t=\ty\n\t \t \t^'
    y = '\tx\t=\ty\n\t \t \t^\n'
    print('\n'.join(
        line.rstrip('\n') for line in
        ndiff(x.splitlines(True), y.splitlines(True)))
    )


if __name__ == '__main__':
    exit(main())

Current output:

$ python3.8 t.py
  	x	=	y
- 	 	 	^
+ 	 	 	^
? 	     +

Expected output:

$ ./python t.py
  	x	=	y
- 	 	 	^
+ 	 	 	^
? 	 	 	 +

Found this while implementing a similar thing for flake8 here: https://gitlab.com/pycqa/flake8/merge_requests/339 and while debugging with pytest

[rhettinger]

This seems like a reasonable suggestion to me. Am not sure whether it should be backported.

Tim, what do you think?

[asottile]

That's actually a good point, I don't think this should land in python3.7 since it changes outuput -- I'm removing that from the versions (though the bug does affect every version I have access to)