[marcelm]

A couple of commits that are somewhat related:

• Bumped ruff to a newer version, which caused a lot more issues to appear
• Fixed most issues either with ruff check --fix or by hand
• Disabled the remaining lints
• Dropped support for Python 3.10
• Use Python 3.14 instead of 3.10 also for some steps in the GitHub Actions workflow
• Add changelog entry for lz4 support (PR Add optional LZ4 frame support (refresh #168) #181)
• Bump codecov-action version
• Enable dependabot (but only for GitHub Actions)

CI now fails on Windows (due to its tests being run on Python 3.14), but I don’t have time to investigate now.

Edit:

• Remove tell/seek from _PipedCompressionProgram

[rhpvorderman]

CI now fails on Windows (due to its tests being run on Python 3.14), but I don’t have time to investigate now.

From a quick check I see a lot of resource warnings. So something is probably not going as it should on all operating systems, it is just in windows where it leads to noticable failures.

[rhpvorderman]

I removed a rogue read statement (in a close operation? No idea why it got there.) That removed quite a lot of errors but does not fix it. It seems specific to the seeking behavior that only worked on windows. I guess that since 3.14 it does not anymore. I currently have no time to investigate further.

[marcelm]

The test is only enabled on Windows and fails on Linux anyway. Should we disable seeking on Windows as well for now?

[rhpvorderman]

Seeking in a stream is always a bit odd isn't it? Should we set pipedcompressionreader and writer to always return seekable() -> false ?

[marcelm]

I agree.

I checked why we added this at all: The PR that added seek() was #83, and I added it because it was needed in dnaio. That PR is from January 2022, but then in November 2024, we made a change in marcelm/dnaio#148 to not rely on seekable().

What do you think, should I just revert PR #83?

[rhpvorderman]

What do you think, should I just revert PR #83?

Yes, that makes sense. And maybe add a seekable -> false function? Or is that included by default from the IObase class?

[marcelm]

I’ve added the changes to this PR.

[rhpvorderman]

All tests pass! Looks great.