Disable hugging when the argument is named

[DavisVaughan]

Follow up to #228 after running this on dplyr code and talking to @lionel- about it

We remain confident that hugging is extremely useful in general, but we think that hugging is mostly useful when the argument in question is not named. For example:

# Very useful - and you almost read this in your head as `abort_paste0()`
abort(paste0(
  "bullet",
  "info"
))

# Feels a bit awkward - the `message` is in the way.
# The presence of `message` somewhat implies that you plan to add >1 arguments, so the naming of them would be useful,
# and if >1 arguments are present then hugging would be invalid anyways.
abort(message = paste0(
  "bullet",
  "info"
))

Hugging becomes particularly not-useful in functions that accept 1 or more key/value pairs, like mutate()

# before
storms <- storms %>%
  mutate(name = if_else(str_sub(name, 1, 3) %in% c("AL0", "AL1"), name, str_to_title(name))) %>%
  mutate(
    name2 = some_thing(),
    name3 = some_other_thing()
  ) 

# after, with hugging
storms <- storms %>%
  mutate(name = if_else(
    str_sub(name, 1, 3) %in% c("AL0", "AL1"), 
    name, 
    str_to_title(name)
  )) %>%
  mutate(
    name2 = some_thing(),
    name3 = some_other_thing()
  ) 

# after, without hugging
storms <- storms %>%
  mutate(
    name = if_else(
      str_sub(name, 1, 3) %in% c("AL0", "AL1"), 
      name, 
      str_to_title(name)
    )
  ) %>%
  mutate(
    name2 = some_thing(),
    name3 = some_other_thing()
  ) 

Hugging takes away from the symmetry a little bit, making it hard to scan down the pipeline and see the new variable names of name =, name2 =, and name3 =. It also makes it a little harder to add a 2nd column expression to that first mutate() call, because you have to place your cursor between the )) and then add the , there.

We do think that hugging can be somewhat nice in other dplyr verbs, like filter(), where named arguments are not used

# before
storms <- storms %>%
  filter(any(status %in% c("hurricane", "tropical storm", "tropical depression")))

# after with hugging - in this case you read it like `filter_any()` which is kind of nice
storms <- storms %>%
  filter(any(
    status %in% c("hurricane", "tropical storm", "tropical depression")
  ))