Davis Vaughan

A few questions: - Should a `.name_repair` argument be passed to `vec_c()`? - If a `.ptype` is specified and `vec_c()` fails because it cannot cast to the `.ptype`, we currently...

Other things vctrs can help with: - Using `vec_rbind()` in `map_dfr()` - Using `vec_cbind()` in `map_dfc()` - Possibly powering `map_dbl()` and friends by just doing `map_vec(.ptype = double())`? - Altering...

I see that mention of `funs::s()` [here](https://github.com/r-lib/funs/issues/23#issue-408193412). I guess it would be like `s(x, ptype = NULL)`? I am a bit worried about performance here. If purrr was being implemented...

Thought: How can we compute the prototype of the output in advance if we need all elements of the output to determine the common type? Unless, `.ptype` is supplied, we...

In #1160 we discussed just setting `allow_rename = FALSE` in the tidyselect call to disable this early on, but https://github.com/tidyverse/tidyr/issues/1160#issuecomment-928151730 reminds us of a tidyselect bug preventing this from working...

We are going to wait until we get a tidyselect release that will fix this, and then we can apply `allow_rename = FALSE` everywhere all at once. Needed in: -...

Should archive export something like `archive_formats()` to make it easier to identify if a file extension is supported? https://github.com/r-lib/vroom/blob/1b8354671b85d40e3797116be892bd298ebab850/R/path.R#L171

I'm convinced that using `group_by()` to complete or expand "within" a group is a useful part of this API that can't be represented with `nesting()` and `crossing()` alone

https://github.com/tidyverse/tidyr/pull/1300 mentions in the docs that `complete()` and `expand()` work with grouped data frames, and adds examples

I am not sure comparing tibble sub assignment behavior directly against `dplyr_col_modify()` is the right way to go here. `dplyr_col_modify()` is mainly used for mutate/summarise which always overwrite an existing...