housekeeping tasks

[skshetry]

Some of the things in addition to "move to GHA" and changelog:

CI

• Cache pre-commit's caches on CI.
• Run tests using lower bounds of requirements.
• Stop building packages for PR.
• Check for licensing issues and vulnerabilities for dependencies.

Tests

• Look on how we can speedup tests. Maybe, we can remove a lot of redundant tests.
  Also, since these days we use RepoTree for external repo scenario, lots of tests involving ls/update/import/get could be reduced (which by the way are the slowest). At least mark them as slow.
• Provide a way to selectively disable certain remotes when testing (or, make it optional by default, but run all on CI)?
  Something like pytest --disable-hdfs?
• Cover branches with coverage. Display coverage report on the GHA log and measure branch coverage #4760
• Track failing/flakey tests, mark them as flakey.
• Why is docker running on unit tests? 😸
• Configurable tests for remotes.
• Consider using nox/tox for tests and pylint as these things are quite slow.
• Add type hints (Add type-hints #4697).
• Improve coverage to above 90% codecov: add target coverage, and threshold for patch and project #4873

Packaging

• Move to PEP 517/518? use setuptools_scm and build_requires #6775
• Auto-update tests requirements. Either use dependabot or https://pyup.io.
• Auto-update pre-commit.
• update restyler? How to keep that in-sync? Should we host our own docker images that are kept fresh?
• Python 3.9 support "soon", actually we could start testing the support. Previous attempt: [TEST-PR] dvc: test python3.9 support #3492
• Deploy on test.pypi.org

These were the tasks I have been keeping tracks of on a separate list, thought I'd just dump. Not everything needs to fixed though. I could take them when I am free as well.

[efiop]

update restyler? How to keep that in-sync? Should we host our own docker images that are kept fresh?

Could consider disabling it. Seems like our linter job is doing the thing and we don't actually use restyler's PRs that often, as people tend to fix their patches when they see that linter complains.

[shcheklein]

Should we make some of them hacktoberfest tickets for the upcoming event?

[skshetry]

@shcheklein, sure, we can split this ticket.

[DiPaolo]

I've just spotted the missed year and copyright owner in LICENSE file:

dvc/LICENSE

Line 189 in 11382cc

Copyright {yyyy} {name of copyright owner}

Should it be added as a housekeeping item as well?

[skshetry]

@DiPaolo, it is a boilerplate to copy in your work (or, source files) that you want to specifically apply Apache License (and, replace that syntax with information).

Here's what Apache License says:

To apply the Apache License to specific files in your work, attach the following boilerplate declaration, with the fields enclosed by brackets "[]" replaced with your own identifying information. (Don't include the brackets!)

Which is close to what's in the LICENSE:

dvc/LICENSE

Lines 180 to 184 in 11382cc

	To apply the Apache License to your work, attach the following
	boilerplate notice, with the fields enclosed by brackets "{}"
	replaced with your own identifying information. (Don't include
	the brackets!) The text should be enclosed in the appropriate
	comment syntax for the file format. We also recommend that a

To emphasize the last line:

The text should be enclosed in the appropriate comment syntax for the file format.

[isidentical]

Cache pre-commit's caches on CI.

I guess we can simply use the official pre-commit action, pre-commit/action. It automatically handles the caching

[skshetry]

Closing as stale.