[VisruthSK]

This PR was made with assistance from Codex.

Summary

Fix failures in R-devel due to changes to base apply(). Bumped roxygen to 8.0.0 and regenerated documentation. Also normed files to lf ending.

Copyright and Licensing

By submitting this pull request, the copyright holder is agreeing to license the submitted work under the following licenses:

• Code: BSD 3-clause (https://opensource.org/licenses/BSD-3-Clause)
• Documentation: CC-BY 4.0 (https://creativecommons.org/licenses/by/4.0/)

[github-actions]

This is how benchmark results would change (along with a 95% confidence interval in relative change) if 3f6b50d is merged into master:

• ✔️as_draws_array: 140ms -> 139ms [-2.11%, +0.51%]
• ✔️as_draws_df: 65.8ms -> 65.7ms [-0.72%, +0.63%]
• ✔️as_draws_list: 152ms -> 150ms [-2.7%, +0.08%]
• ✔️as_draws_matrix: 24.3ms -> 24.4ms [-0.95%, +1.72%]
• ✔️as_draws_rvars: 112ms -> 112ms [-0.58%, +1.35%]
• 🚀summarise_draws_100_variables: 730ms -> 727ms [-0.77%, -0.01%]
• ✔️summarise_draws_10_variables: 81.5ms -> 81.6ms [-0.2%, +0.5%]
  Further explanation regarding interpretation and methodology can be found in the documentation.

[github-actions]

This is how benchmark results would change (along with a 95% confidence interval in relative change) if 93af05f is merged into master:

• 🚀as_draws_array: 145ms -> 143ms [-2.28%, -0.37%]
• ✔️as_draws_df: 69.7ms -> 69.6ms [-1.57%, +1.32%]
• ✔️as_draws_list: 158ms -> 159ms [-0.92%, +1.5%]
• ✔️as_draws_matrix: 24.7ms -> 24.9ms [-0.35%, +1.87%]
• ✔️as_draws_rvars: 116ms -> 115ms [-1.8%, +0.28%]
• ✔️summarise_draws_100_variables: 733ms -> 734ms [-0.33%, +0.48%]
• ✔️summarise_draws_10_variables: 82.1ms -> 82.3ms [-0.25%, +0.58%]
  Further explanation regarding interpretation and methodology can be found in the documentation.

[github-actions]

This is how benchmark results would change (along with a 95% confidence interval in relative change) if 59398ee is merged into master:

• ✔️as_draws_array: 130ms -> 129ms [-1.64%, +0.5%]
• ✔️as_draws_df: 71.7ms -> 71.3ms [-2.06%, +0.91%]
• ✔️as_draws_list: 144ms -> 142ms [-3.02%, +0.83%]
• ✔️as_draws_matrix: 20ms -> 19.9ms [-1.51%, +0.17%]
• ✔️as_draws_rvars: 106ms -> 106ms [-0.96%, +2.04%]
• ✔️summarise_draws_100_variables: 600ms -> 601ms [-0.19%, +0.25%]
• ✔️summarise_draws_10_variables: 67.4ms -> 67.4ms [-0.38%, +0.33%]
  Further explanation regarding interpretation and methodology can be found in the documentation.

[github-actions]

This is how benchmark results would change (along with a 95% confidence interval in relative change) if bc7f533 is merged into master:

• ✔️as_draws_array: 138ms -> 137ms [-1.44%, +0.04%]
• ✔️as_draws_df: 77ms -> 77.7ms [-1.91%, +3.73%]
• ✔️as_draws_list: 153ms -> 153ms [-0.41%, +0.79%]
• ✔️as_draws_matrix: 24.2ms -> 24ms [-1.9%, +0.25%]
• ✔️as_draws_rvars: 129ms -> 128ms [-1.87%, +0.92%]
• ✔️summarise_draws_100_variables: 825ms -> 825ms [-2.65%, +2.79%]
• ✔️summarise_draws_10_variables: 96.3ms -> 95.8ms [-2.06%, +1.08%]
  Further explanation regarding interpretation and methodology can be found in the documentation.

[paul-buerkner]

thank you! I have been in contact with the CRAN team and it may be that the commit that causes posterior to fail on R-devel may be reverted. But perhaps it would make them happy if posterior would still pass already now. And it may make posterior more robust against these changes in the future.

Do you expect any negative "side effects" of this PR? I.e. some aspects of rvar computation becoming more prone to bugs or errors? If not, I am happy to merge this PR.

[paul-buerkner]

@mjskay do you have a few minutes to review this PR? You can judge rvar code better than I can.

[VisruthSK]

I'm realizing that it would've been better to separate out the rvar changes from the other ones in this PR, but hopefully that doesn't make it too difficult to review.

I don't think there should be any major correctness or performance changes--I didn't check but I can't imagine too many people were relying on the vectrs default behaviours for as.array() and as.matrix(). I guess a rev dep check might be useful? But also might be overkill...

Also would like to get @mjskay to review if possible before merging.

[VisruthSK]

@paul-buerkner based on what you said in #473, and due to #475 I think this PR can be closed. The one thing that I think should be carried forward is the lf normalization, I could rebase the PR and just apply that fix whenever, obviously not critical but just nice to have and reduces nonsense in diffs sometimes.

[paul-buerkner]

Please go ahead and rebase the PR (or make a new one) with the remaining fixes you want to apply. Closing this one for the time being.