Skip to content

Conversation

silverweed
Copy link
Contributor

First PR of a series to merge the RNTuple Attributes into master. The final result will be this, although the commits will be reorganized to be more coherent and reviewable.

This first PR updates the binary format specification (introducing a new minor version) and updates the Serializer and Descriptor code to match. This is backward-compatible and no Attribute can be written yet since the writer API will be introduced later.

Checklist:

  • tested changes locally
  • updated the docs (if necessary)

Copy link

github-actions bot commented Sep 16, 2025

Test Results

    22 files      22 suites   3d 21h 51m 39s ⏱️
 3 688 tests  3 684 ✅ 0 💤 4 ❌
79 208 runs  79 204 ✅ 0 💤 4 ❌

For more details on these failures, see this check.

Results for commit bb2d966.

♻️ This comment has been updated with latest results.

@silverweed silverweed added the clean build Ask CI to do non-incremental build on PR label Sep 16, 2025
@silverweed silverweed closed this Sep 16, 2025
@silverweed silverweed reopened this Sep 16, 2025
Copy link
Contributor

@jblomer jblomer left a comment

Choose a reason for hiding this comment

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

In principle looks good to me. Some comments.

@silverweed silverweed force-pushed the ntuple_attr_1 branch 2 times, most recently from b0449c0 to 48b72ff Compare September 18, 2025 08:45
@silverweed
Copy link
Contributor Author

I updated the PR and uniformed the RNTupleAttrSetDescriptor with the other descriptor classes.

@silverweed silverweed requested a review from jblomer September 18, 2025 08:46
Copy link
Contributor

@jblomer jblomer left a comment

Choose a reason for hiding this comment

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

Very nice! Some minor comments.

Since this is changing the binary format, let's perhaps get a second approval.

Copy link
Member

@hahnjo hahnjo left a comment

Choose a reason for hiding this comment

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

some comments on the spec additions and the iterator implementations

Copy link
Contributor

@enirolf enirolf left a comment

Choose a reason for hiding this comment

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

LGTM in principle! Just some additional suggestions to the spec to improve readability and consistency.

Copy link
Contributor

@enirolf enirolf left a comment

Choose a reason for hiding this comment

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

Looks good from my side! A second approval would be good.

@silverweed silverweed force-pushed the ntuple_attr_1 branch 2 times, most recently from 323e8bb to 2940de0 Compare October 6, 2025 09:18
@silverweed silverweed requested a review from hahnjo October 6, 2025 09:19
Copy link
Member

@hahnjo hahnjo left a comment

Choose a reason for hiding this comment

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

Thanks for reorganizing the spec additions, I (personally) find them easier to follow this way. It's still mentioning "model" in some places (notably in the _userModel field name), not sure if we want to fully remove it...

Copy link
Contributor

@jblomer jblomer left a comment

Choose a reason for hiding this comment

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

I think we need one more change with this PR: the RNTuple anchor needs to update the schema version to match the new specification.

@silverweed
Copy link
Contributor Author

It's still mentioning "model" in some places (notably in the _userModel field name), not sure if we want to fully remove it...

I'll remove it from the leftover places; I guess at this point I should also rename the field to _userSchema?

@hahnjo
Copy link
Member

hahnjo commented Oct 6, 2025

It's still mentioning "model" in some places (notably in the _userModel field name), not sure if we want to fully remove it...

I'll remove it from the leftover places; I guess at this point I should also rename the field to _userSchema?

Or actually _userData since that's what the record field contains?

@silverweed silverweed force-pushed the ntuple_attr_1 branch 2 times, most recently from 3a8b73c to da574f2 Compare October 6, 2025 13:42
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
clean build Ask CI to do non-incremental build on PR in:RNTuple
Projects
None yet
Development

Successfully merging this pull request may close these issues.

5 participants