[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: gnulib-tool.py: Fix pylint 'attribute-defined-outside-init' warnings
From: |
Collin Funk |
Subject: |
Re: gnulib-tool.py: Fix pylint 'attribute-defined-outside-init' warnings. |
Date: |
Thu, 18 Apr 2024 10:34:06 -0700 |
User-agent: |
Mozilla Thunderbird |
Hi Bruno,
On 4/18/24 6:05 AM, Bruno Haible wrote:
> The same holds for GCC, glibc, qemu, and many other project: They have
> many open issues because the developers prioritize them. They nevertheless
> depend on their users for reporting issues. Paul Eggert and I regularly
> report issues in GCC, for instance. Please help making pylint better,
> since we use it. (Call it a "wish", rather than a "bug", if you prefer.)
True, that is a better way to think about it. If no one submitted
bugs/requests than software would never improve. :)
I submitted a feature request here since I think a configuration
option would probably be the best fix:
https://github.com/pylint-dev/pylint/issues/9555
> Yes, please. I think it will be a benefit to follow this convention,
> even though it departs from our previous coding style.
Nice. I'll add a small note in main.py.
> Please submit patches in this direction one-class-at-a-time, not all
> together, since that would be unreviewable.
Yes, that makes sense. I think it will be best for me to go through
the obvious ones first. I'll submit seperate patches for those.
When I say obvious in GLModule we have (comments are added):
# Cache for the getter functions.
self.cache = dict()
# Raw content of description file.
self.content = ''
self.filesystem = GLFileSystem(self.config)
self.modulesystem = GLModuleSystem(self.config)
It is very clear that the cache and content are meant for internal
use. The filesystem and modulesystem are less clear to me. I think
they *could* be used from the outside, but should they be? Those will
take more consideration.
Collin