filecmp.cmpfiles w/ absolute path names

[bersbersbers]

BPO	46512
Nosy	@terryjreedy, @bersbersbers

^{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 = None
created_at = <Date 2022-01-25.09:46:10.637>
labels = ['type-bug', 'library', '3.9']
title = 'filecmp.cmpfiles w/ absolute path names'
updated_at = <Date 2022-01-29.08:21:26.034>
user = 'https://github.com/bersbersbers'

bugs.python.org fields:

activity = <Date 2022-01-29.08:21:26.034>
actor = 'bers'
assignee = 'none'
closed = False
closed_date = None
closer = None
components = ['Library (Lib)']
creation = <Date 2022-01-25.09:46:10.637>
creator = 'bers'
dependencies = []
files = []
hgrepos = []
issue_num = 46512
keywords = []
message_count = 3.0
messages = ['411570', '412043', '412058']
nosy_count = 2.0
nosy_names = ['terry.reedy', 'bers']
pr_nums = []
priority = 'normal'
resolution = None
stage = None
status = 'open'
superseder = None
type = 'behavior'
url = 'https://bugs.python.org/issue46512'
versions = ['Python 3.9']

[bersbersbers]

It is very easy to use filecmp.cmpfiles incorrectly by passing absolute path names. This is because

1. the documentations does not say that relative path names have to be passed, and
2. filecmp.cmpfiles does not issue a warning when absolute path names are passed.

Consider this example code, which does look sensible at first glance:

    files = dir_a.glob("*")
    (equal, _, _) = filecmp.cmpfiles(dir_a, dir_b, files, shallow=False)
    print("equal:", *equal)

However, in the full example below, you will see that this code fails to detect that two files are actually different.

"""Demo behavior of filecmp.cmpfiles with absolute path names."""
import filecmp
import tempfile
from pathlib import Path

with tempfile.TemporaryDirectory() as tmpdirname:
    # prepare two different files
    tmpdir = Path(tmpdirname)
    dir_a = tmpdir / "a"
    dir_b = tmpdir / "b"
    file_a = dir_a / "foo.txt"
    file_b = dir_b / "foo.txt"

    dir_a.mkdir()
    dir_b.mkdir()
    file_a.write_text("A")
    file_b.write_text("B")

    # actually diff the files
    files = dir_a.glob("*")
    # filecmp should issue a warning here!
    (equal, _, _) = filecmp.cmpfiles(dir_a, dir_b, files, shallow=False)
    # otherwise, this result is easy to misinterpret - files are reported as equal
    print("equal:", *equal)

[terryjreedy]

https://docs.python.org/3/library/filecmp.html#filecmp.cmpfiles
I consider not working for absolute path names to be a bug. Did your example work with relative paths?

[bersbersbers]

Did your example work with relative paths?

Yes, it does. Just append the following to my example code:

    # actually diff the files - correctly!
    files = [f.relative_to(dir_a) for f in dir_a.glob("*")]
    (_, different, _) = filecmp.cmpfiles(dir_a, dir_b, files, shallow=False)
    print("different:", *different)

Output then is

equal: C:\Users\bers\AppData\Local\Temp\tmp1p6jh4rg\a\foo.txt
different: foo.txt