The array_inplace_repeat is inefficient for small arrays and a high number of repeats. This can be improved by using the same approach as in https://bugs.python.org/issue47005

I'd bet we could add a couple of utility functions that could be used in multiple places, to keep the "trick" all in one place. Something like

void
_PyBytes_RepeatInPlace(char **buffer, size_t start_len, size_t end_len)
{
    // Repeatedly double.
    size_t copied = start_len;
    while (copied < end_len) {
        size_t to_copy = Py_MIN(copied, end_len - copied);
        memcpy(buffer + copied, buffer, to_copy);
        copied += to_copy;
    }
}

void
_PyBytes_Repeat(char *dest, size_t len_dest,
                const char *src, size_t len_src)
{
    // XXX maybe handle zero lengths
    // XXX maybe use memset for len_src == 1
    memcpy(dest, src, len_src);
    _PyBytes_RepeatInPlace(dest, len_src, len_dest);
}

New changeset 850687df47b03e98c1433e6e70e71a8921eb4454 by Pieter Eendebak in branch 'main':
bpo-47070: Add _PyBytes_Repeat() (GH-31999)
https://github.com/python/cpython/commit/850687df47b03e98c1433e6e70e71a8921eb4454