Skip to content

Conversation

@kou
Copy link
Member

@kou kou commented Sep 21, 2018

No description provided.

@kszucs
Copy link
Member

kszucs commented Sep 22, 2018

Great! Thanks @kou!

How should we handle parquet versioning? Will we have distinct git tags for parquet as well as separate release?

@kou
Copy link
Member Author

kou commented Sep 22, 2018

Do you mean that Apache Arrow and Apache Parquet C++ use different versions?
I thought that "Apache Arrow's release archive ships Apache Parquet C++" means "Apache Arrow and Apache Parquet C++ use the same version"...

If we use the different versions for Apache Arrow and Apache Parquet C++, we should create two debian/ directories and two *.spec files for Apache Arrow and Apache Parquet C++.

@wesm
Copy link
Member

wesm commented Sep 22, 2018

How should we handle parquet versioning? Will we have distinct git tags for parquet as well as separate release?

This is up to the Parquet community. To avoid complications we may just want to bundle the Parquet libraries in the same package as libarrow.so

@kszucs
Copy link
Member

kszucs commented Sep 24, 2018

Just asking because the two most recent parquet-cpp version numbers are 1.4.1, 1.5.0pre (at least on conda-forge) and incrementing it to 0.11.0 can cause confusion.
I'd favor the same version number too, but seen different practice for javascript packaging.

@wesm
Copy link
Member

wesm commented Sep 24, 2018

My thought was to bundle everything as a unit, to avoid having to worry about release coordination between Apache Arrow and Apache Parquet

@kou
Copy link
Member Author

kou commented Sep 28, 2018

I rethought about this.

We should use the same version as Apache Arrow for Apache Parquet C++ package. Because both .deb/.rpm use one version for one source archive.

We put Apache Arrow and Apache Parquet C++ into one source archive regardless of Apache Parquet C++ version. So the approach in this pull request is reasonable.

If there are no more concerns, I'll merge this pull request.

Depends:
${misc:Depends},
${shlibs:Depends}
Description: Apache Parquet is a columner storage format
Copy link
Member

@kszucs kszucs Sep 29, 2018

Choose a reason for hiding this comment

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

There is a typo columnAr.

We might want to highlight, that it's the cpp implementation, e.g:

C++ library to read and write Apache Parquet columnar storage format

Copy link
Member Author

Choose a reason for hiding this comment

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

Oh! Thanks for fixing the typo.

C++ information will be useful. We'll add it when we add packages for GLib bindings.

Copy link
Member

@kszucs kszucs left a comment

Choose a reason for hiding this comment

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

Thanks @kou!

@kszucs kszucs closed this in 7157a74 Sep 29, 2018
@kou kou deleted the packages-linux-parquet branch September 30, 2018 11:53
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

3 participants