`check_data()` fails with long expressions because `check_dim()` does not accept character vectors for `x_name`

[StefanoMezzini]

check_data(data.frame(fac = factor(c("a", "b", "z"), levels = c("a", "z", "b"))),
                      list(fac = factor(c("a", "b", "z"), levels = c("a", "z", "b"))))

gives the error

Error in `check_dim()` at chk/R/check-data.R:41:3:
! `x_name` must be a string (non-missing character scalar).
Run `rlang::last_trace()` to see where the error occurred.

because backtick_chk(deparse(substitute(x)))  returns the character vector of length 2

[1] "`data.frame(fac = factor(c(\"a\", \"b\", \"z\"), levels = c(\"a\", \"z\", `"
[2] "`    \"b\")))`"`

Despite this bug being a likely uncommon edge case, I think the bug is worth fixing because the error is quite uninterpretable without debugging.

We could:

• only use first string (and append "..." to the end of it?)
• collapse the elements with p0(..., collapse = "")
• indenting of the second element can probably be dropped, since the first line ends with a space, but this may not always be the case