separate_wider_regex(), names_sep, and too_few

[jennybc]

Moving from Slack.

I just used separate_wider_regex() on a tibble created with parsermd on the section headings from a chapter from R Packages.

It turns out I needed to specify names_sep and too_few, but those user errors interact with each other in a way that makes it harder to figure out what's wrong.

library(tidyverse)

x <- tribble(
  ~ sec_h1, ~ sec_h2,
  "Intro {intro}", NA,
  "First cool thing {first-cool-thing}", NA,
  "First cool thing {first-cool-thing}", "First cool details {first-cool-details}",
  "Second cool thing", NA
)
x |> 
  separate_wider_regex(
    everything(),
    c("text" = "[^\\{]*", " [{]", label = ".*", "[}]"),
    too_few = "debug"
  )
#> Warning: Debug mode activated: adding variables `sec_h1_ok`, `sec_h1_matches`, and
#> `sec_h1_remainder`.
#> Warning: Debug mode activated: adding variables `sec_h2_ok`, `sec_h2_matches`, and
#> `sec_h2_remainder`.
#> Error in `unpack()`:
#> ! Names must be unique.
#> ✖ These names are duplicated:
#>   * "text" at locations 1 and 7.
#>   * "label" at locations 2 and 8.
#> ℹ Use argument `names_repair` to specify repair strategy.

I think I should be getting scolded first and foremost for not specifying names_sep, which is actually required since cols = everything(). That is the root cause of the unpack() error but the suggested remedy is off-base.

I'd probably also be happy with some default choice of names_sep (related #1367).

Until you get the naming stuff straightened out, you can't see the cool too_few = "debug" diagnostics.