Parsing of mail addresses with personal name not correctly encoded

Nobody has claimed this yet.

**To Reproduce**
The following code demonstrates the issue:

```
var address = new InternetAddress("Mäd Mäx");
System.out.println(address);
```
**Actual behaviour**
The output is not encoded:
```
"Mäd Mäx"
```
The description of `toString()` says:

> Convert this address into a RFC 822 / RFC 2047 encoded address.
> The resulting string contains only US-ASCII characters, and hence is mail-safe.

**Expected behavior**
The output should be encoded:
`=?UTF-8?Q?M=C3=A4d_M=C3=A4x?=
`

**Additional context**
I have added a workaround like that:
`address.setPersonal(address.getPersonal());`
It looks redundant, but it isn't. Afterwards it worked.

I took a look at the code, and it seems that `jakarta.mail.internet.InternetAddress.parse(String, boolean, boolean)` assumes that the personal name is already encoded. That is true for headers, not for ordinary text.

Seems to be similar to #379.

Start at jakarta.mail.internet.InternetAddress.parse(String, boolean, boolean), which the issue identifies as treating the personal name as already encoded. Compare that behavior with setPersonal(address.getPersonal()) and the documented toString() contract. Done means parsing the reproduced address produces an RFC 2047-encoded personal name while preserving the email address.

Written by the indexing model from the issue text.