vec_ptype2.character.character issue

[BurninLeo]

The vctrs package asked me to report an error:

Please report it at https://github.com/r-lib/vctrs/issues with a reprex and the full backtrace.

Here we go:

Error in `stop_native_implementation()`:
! `vec_ptype2.character.character()` is implemented at C level.
  This R function is purely indicative and should never be called.
ℹ This is an internal error that was detected in the vctrs package.
  Please report it at <https://github.com/r-lib/vctrs/issues> with a reprex and the full backtrace.
Backtrace:
 1. ... %>% bind_rows(mz %>% cleanup1() %>% cleanup2())
 2. dplyr::bind_rows(., mz %>% cleanup1() %>% cleanup2())
 3. vctrs::vec_rbind(!!!dots, .names_to = .id, .error_call = current_env())
 4. vctrs (local) `<fn>`()
 5. vctrs:::vec_ptype2.character.character(...)
 6. vctrs:::stop_native_implementation("vec_ptype2.character.character")
Error in stop_native_implementation("vec_ptype2.character.character") :

This R function is purely indicative and should never be called.
ℹ This is an internal error that was detected in the vctrs package.
Please report it at <https://github.com/r-lib/vctrs/issues> with a reprex and the full backtrace.
10.
stop(fallback)
9.
signal_abort(cnd, .file)
8.
rlang::abort(message, ..., call = call, use_cli_format = TRUE,
.frame = .frame)
7.
cli::cli_abort(c("{.fn {fn}} is implemented at C level.", ` ` = "This R function is purely indicative and should never be called."),
.internal = TRUE)
6.
stop_native_implementation("vec_ptype2.character.character")
5.
vec_ptype2.character.character(x = x, y = y, x_arg = x_arg, y_arg = y_arg,
call = call, `vctrs:::s3_fallback` = `vctrs:::s3_fallback`)
4.
(function ()
vec_ptype2.character.character(x = x, y = y, x_arg = x_arg, y_arg = y_arg,
call = call, `vctrs:::s3_fallback` = `vctrs:::s3_fallback`))()
3.
vec_rbind(!!!dots, .names_to = .id, .error_call = current_env())
2.
bind_rows(., mz %>% cleanup1() %>% cleanup2())
1.
merge(m1 %>% cleanup1(), m2 %>% cleanup1(), by.x = "R103", by.y = "SERIAL",
all.x = T, all.y = T) %>% cleanup2() %>% bind_rows(mz %>%
cleanup1() %>% cleanup2())

And here's a replication code I built

add_class = function(x) {
    x = data.frame(
      lapply(x, function(xx) {
        structure(xx, class = c("avector", class(xx)))
      })
    )
    return(x)
}
drop_class = function(x) {
    x = data.frame(
      lapply(x, function(xx) {
        structure(xx, class = setdiff(class(xx), c("avector")))
      })
    )
    return(x)
}

x1 = rbind(c(1,2,"c"), c(2,3,"d"))
x2 = rbind(c(3,4,"d"), c(4,5,"f"))

bind_rows(
    x1 %>% add_class() %>% drop_class(),
    x2 %>% add_class() %>% drop_class()
)

No need to discuss the add_class () / drop_class() functions, this issue is only upon request to report problematic error reporting :)