Hadley Wickham

Need to do after PRs mostly closed * Ensure all use `.R` * Ensure `tests/` and `R/` align

Remove lazy eval fallbacks (>= Sept 2023)

5

They shouldn't be needed anymore, and they make the error messages unnecessarily confusing. (Have been deprecated since 0.7.0 released June 2017)

## Pre 1.0.0 * [x] `funs()` to defunct. Since 0.8.0 (Feb 2019). * [x] `select_vars()` and friends to defunct. Since 0.8.4 (Oct 2019) ## Probably low-risk 1.0.0 * [x] Move...

Move database generics to dbplyr

6

Deprecate database generics, following strategy in https://github.com/tidyverse/design/issues/106. Then remove tests, since they should live in dbplyr. https://github.com/tidyverse/dbplyr/issues/385 needs to happen first. Remove ```R setOldClass(c("sql", "character")) setOldClass(c("ident", "sql", "character")) ```

Next round of database defunctions

3

#4663

Automatically unpack in `across()`

1

To support this sort of functionality: ```R library(dplyr, warn.conflicts = FALSE) quantile_df summarise(across(x:y, quantile_df)) ```

Fixes #6309

Add dplyr_sticky_cols() or similar

4

So that when you implement a tibble subclass with sticky columns, you don't need to add a new `select()` method. Likely to involve renaming `ensure_group_vars()`

Consider bind_cols() alternatives

13

So we can have experimental solution without forcing logic into vctrs.

Mention project search and nav

1

In scripts/projects