Pylint no longer support separating options from positional arguments with --

[phst]

Bug description

With Pylint 2.12, pylint -- file.py works as expected, with Pylint 2.14 it raises an error like "pylint: error: Unrecognized option found:". I guess that's related to the argparse migration.

Configuration

No response

Command used

pylint -- file.py

Pylint output

pylint: error: Unrecognized option found:

Expected behavior

The -- should separate options from positional arguments. In this case the invocation should have identical results to pylint file.py.

Pylint version

pylint 2.14.3
astroid 2.11.6
Python 3.9.12 (main, May  6 2022, 20:15:26) 
[GCC 11.2.0]

OS / Environment

Debian GNU/Linux

Additional dependencies

No response

[Pierre-Sassoulas]

Thank you for opening the issue.

I added the discussion label because I don't know if we want to deviate from what argparse is doing and if it's worth it. In your example it's possible to use pylint file.py directly, right ? What is the standard we're talking about here, is it just something that optparse was doing or is it something well established for CLI ?

[phst]

Thank you for opening the issue.

I added the discussion label because I don't know if we want to deviate from what argparse is doing and if it's worth it.

Argpase supports this syntax. From https://docs.python.org/3/library/argparse.html#arguments-containing:

you can insert the pseudo-argument '--' which tells parse_args() that everything after that is a positional argument

So this should work out of the box, and I'm wondering why it doesn't.

In your example it's possible to use pylint file.py directly, right ?

yes, the real use case is e.g. a script that wraps Pylint and receives arbitrary filenames, which might start with a dash.

What is the standard we're talking about here, is it just something that optparse was doing or is it something well established for CLI ?

It's quite established, see e.g. the GNU Coding Standards (https://www.gnu.org/prep/standards/html_node/Command_002dLine-Interfaces.html) or POSIX (https://pubs.opengroup.org/onlinepubs/9699919799/basedefs/V1_chap12.html#tag_12_02, guideline 10).

[phst]

I don't think this is fixed.

$ pylint --version
pylint 2.15.2
astroid 2.12.9
Python 3.10.6 (main, Aug 30 2022, 04:58:14) [Clang 13.1.6 (clang-1316.0.21.2.5)]

$ pylint -- build.py 
************* Module --
--:1: [F0001(fatal), ] No module named --

That is, Pylint now thinks that it should lint a file named --. Instead, it should skip the -- argument.

[DanielNoord]

Hm, that's weird. My test might not be good enough.

Putting this on my TODO list.