Skip to content

Conversation

@N-Boutaib
Copy link

Which issue does this PR close?

Rationale for this change

This PR switches from xz2 to liblzma to reduce duplicate dependencies.

What changes are included in this PR?

  • Bump async-compression from 0.4.19 to 0.4.32, which uses liblzma instead of xz2
  • Patch apach-avro to use this revision which switches to liblzma in avro-rs
  • Switch from xz2 to liblzma in datafusion-core and datafusion-datasource

Are these changes tested?

Yes, via unit tests

Are there any user-facing changes?

None. Getting rid of deprecated carte

@github-actions github-actions bot added documentation Improvements or additions to documentation core Core DataFusion crate datasource Changes to the datasource crate labels Oct 28, 2025
@N-Boutaib N-Boutaib marked this pull request as ready for review October 28, 2025 15:44
Comment on lines +186 to +189
# Temporary override: pull apache-avro from upstream to include unreleased fixes.
# TODO: remove once the next version of apache-avro is published to crates.io and includes commit 3b202c5.
[patch.crates-io]
apache-avro = { git = "https://github.com/apache/avro-rs", rev = "3b202c58f12bd1217eccf8a0028e4176ee4aadf9" }
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

It would be nice to see if a new proper release of apache-avro can be shipped to crates.io in order to let this just be a normal dependency. There seems to be a PR that tries to do this already #17509.

I see that @timsaucer moved apache-avro from xz2 to libzma here apache/avro-rs@3b202c5 (10 Sep).

If https://github.com/apache/avro-rs in fact follows a 2 month release cadence, we should be seeing a new version on crates.io very soon (last release was 25 Aug, 2 months and 3 days ago)

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Opened a ticket there apache/avro-rs#323

@timsaucer
Copy link
Member

I think this is a duplicate of #17509 (as @gabotechs points out)

Also arrow-avro is now getting ready to merge in which will also resolve this issue and is within the arrow-rs family: #17861

My guess is that our best long term solution is to get #17861 merged in and then close both this PR and #17509

@timsaucer
Copy link
Member

At the very least I think this needs to move to Draft. We cannot merge it in with a [patch.crates-io] as then datafusion will not be publishable. I don't think we want the repo in that state.

Thank you for the work on this @N-Boutaib .

@N-Boutaib N-Boutaib marked this pull request as draft October 29, 2025 08:53
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

core Core DataFusion crate datasource Changes to the datasource crate documentation Improvements or additions to documentation

Projects

None yet

Development

Successfully merging this pull request may close these issues.

Dependency conflict with rquest due to async-compression and xz2 linking to lzma

3 participants