Current DER_cert_to_PEM_cert() uses textwrap like this:

    def DER_cert_to_PEM_cert(der_cert_bytes):
        """Takes a certificate in binary DER format and returns the
        PEM version of it as a string."""

        f = str(base64.standard_b64encode(der_cert_bytes), 'ASCII', 'strict')
        return (PEM_HEADER + '\n' +
                textwrap.fill(f, 64) + '\n' +
                PEM_FOOTER + '\n')

But textwrap is designed to break lines at word boundary.
So textwrap.fill for base64 text may be slow.

And `import textwrap` is little slow too.

We can use simple slicing instead of textwrap.

For the record, this is part of `import requests`.
`import ssl` took 11 ms and `textwrap` weights 2.4 ms of it.

                    - ipaddress 2048 us (self 2048 us)
                    - textwrap 2402 us (self 2402 us)
                    - _ssl 2744 us (self 2744 us)
                        - _struct 232 us (self 232 us)
                      - struct 411 us (self 180 us)
                      - binascii 272 us (self 272 us)
                    - base64 997 us (self 315 us)
                  - ssl 11158 us (self 2969 us)

New changeset b75a228af8c0553aef44e4e03763af90fbc8737f by INADA Naoki in branch 'master':
bpo-31659: Use simple slicing to format PEM cert (GH-3849)
https://github.com/python/cpython/commit/b75a228af8c0553aef44e4e03763af90fbc8737f