Jarl again

[maelle]

No description provided.

[github-actions]

This is how benchmark results would change (along with a 95% confidence interval in relative change) if d0a4a39 is merged into deprecations:

• ✔️as_adjacency_matrix: 744ms -> 749ms [-0.8%, +2.09%]
• ✔️as_biadjacency_matrix: 746ms -> 742ms [-1.76%, +0.63%]
• ✔️as_data_frame_both: 1.62ms -> 1.61ms [-3.48%, +2.72%]
• ✔️as_long_data_frame: 4.05ms -> 4.03ms [-2.18%, +1.36%]
• ✔️es_attr_filter: 2.65ms -> 2.63ms [-1.55%, +0.2%]
• ✔️graph_from_adjacency_matrix: 119ms -> 119ms [-1.56%, +1.96%]
• ✔️graph_from_data_frame: 3.69ms -> 3.67ms [-2.2%, +0.98%]
• ✔️vs_attr_filter: 1.49ms -> 1.49ms [-0.79%, +1.56%]
• ✔️vs_by_name: 976µs -> 973µs [-2.97%, +2.45%]
  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 9c03eec is merged into deprecations:

• ✔️as_adjacency_matrix: 795ms -> 793ms [-1.72%, +1.06%]
• ✔️as_biadjacency_matrix: 794ms -> 800ms [-0.97%, +2.54%]
• ✔️as_data_frame_both: 1.77ms -> 1.77ms [-3.05%, +2.87%]
• ✔️as_long_data_frame: 4.28ms -> 4.3ms [-2.39%, +3.06%]
• ✔️es_attr_filter: 3.02ms -> 2.96ms [-5.98%, +2.57%]
• ✔️graph_from_adjacency_matrix: 125ms -> 125ms [-1.31%, +1.97%]
• ✔️graph_from_data_frame: 3.88ms -> 3.89ms [-2.17%, +2.36%]
• ✔️vs_attr_filter: 1.67ms -> 1.68ms [-2.01%, +3.69%]
• ✔️vs_by_name: 1.05ms -> 1.07ms [-1.88%, +5.13%]
  Further explanation regarding interpretation and methodology can be found in the documentation.

[schochastics]

out of curiosity: why does jarl insist on using stringsAsFactors = FALSE

[maelle]

I think it's in lintr too? Maybe we could depend on a more recent R version actually?