The `groups` argument of `summary_rows()` is ultra confusing and doesn't match `cells_summary()`'s interface

[rich-iannone]

There are a few issues with summary_rows() that make it difficult to use. This issue will lay out some of the biggest problems with it.

One of the most confusing parts of summary_rows() is the groups argument. Its default of NULL results in the formation of grand summary rows, which is surprising since there is a dedicated function for that grand_summary_rows(). Many functions that have a groups argument now allow tidyselect semantics (cells_summary(), cells_stub_summary(), cells_row_groups()) but summary_rows() requires a precise vector a group ID values.

Using groups = TRUE to mean all groups is something that was deprecated in gt v0.3 but lives on here.

There's another issue (will be filed separately, but related): cells_summary() seems to not work with its defaults. Here's an MRE:

library(gt)

# MRE for `cells_summary()` not working
gtcars %>% 
  gt(rowname_col = "model", groupname_col = "mfr") %>%
  summary_rows(
    groups = c("Ford", "Ferrari"),
    columns = year,
    fns = list(`latest year` = ~ max(.)),
    formatter = fmt_number,
    decimals = 0,
    use_seps = FALSE
  ) %>%
  tab_style(
    style = list(
      cell_fill(color = background_color),
      cell_text(color = text_color)
    ),
    locations = cells_summary(groups = everything()) # default for `groups`
  )

Error: `everything()` must be used within a *selecting* function.
ℹ See <https://tidyselect.r-lib.org/reference/faq-selection-context.html>.
Run `rlang::last_error()` to see where the error occurred. 

This should just style all available summary rows (less the stub parts of them).