Accept const refs instead of const unique_ptr refs in reduce and scan APIs. - #11960
Conversation
|
For whoever ends up reviewing this for the Java side (perhaps @jlowe) I don't think this breaking change has any impact on the Spark plugin, it only affects the JNI and I have made the necessary changes. |
bdice
left a comment
There was a problem hiding this comment.
Approving C++ and Cython changes.
Co-authored-by: David Wendt <45795991+davidwendt@users.noreply.github.com>
|
Can you add a real description to this PR? It will be added to the changelog at the end of the release. |
Only the PR title is added to the changelog. The description is used for the merge commit message. (Still needs a real description.) |
Codecov ReportBase: 87.40% // Head: 86.88% // Decreases project coverage by
Additional details and impacted files@@ Coverage Diff @@
## branch-22.12 #11960 +/- ##
================================================
- Coverage 87.40% 86.88% -0.53%
================================================
Files 133 133
Lines 21833 21982 +149
================================================
+ Hits 19084 19099 +15
- Misses 2749 2883 +134
Help us with your feedback. Take ten seconds to tell us how you rate us. Have a feature suggestion? Share it here. ☔ View full report at Codecov. |
|
Added a description |
|
@gpucibot merge |
Description
There is almost never a good reason to pass arguments as
unique_ptr<T> const&. Since those arguments cannot be modified, the only use case is accessing the underlying pointer, at which point the function better communicates its intent by accepting the underlying pointer/reference as an argument instead and is also more flexible as a result.Resolves #10393
Checklist