If the JSON content contains some Unicode characters, the jsonpatch
final output will encode the Unicode character using ASCII (i.e
`\u0394`). This behaviour comes from the module `json.dump()` governed
by a flag `ensure_ascii`[1].

For example:

```json
/* patch.json */
[{
  "op": "add",
  "path": "/SomeUnicodeSamples",
  "value": "𝒞𝘋𝙴𝓕ĢȞỈ𝕵 đ áê 🤩 äÄöÖüÜß"
}]
```

After applying the patch on an empty source file `{}`, this is the
output:

```json
{"SomeUnicodeSamples": "\ud835\udc9e\ud835\ude0b...\u00fc\u00dc\u00df"}
```

This commit adds a flag `-u|--preserve-unicode` in the jsonpatch CLI
to configure the behaviour of `json.dump`'s `ensure_ascii` flag.

Using the `--preserve-unicode` flag, the cli will print the Unicode
characters as-is without any encoding.

[1]: https://docs.python.org/3/library/json.html#basic-usage

if there is no diff print nothing, and exit 1 if a diff is found

jsonpatch in a shell script like so:

cat <<EOT | jsonpatch -i a.json
[{"path": "/foo", "value": "baz", "op": "replace"}]
EOT

which will edit a.json in-place and leave a backup in a.json.orig.