Look at the "LDAP: Schema for User Applications" RFC-4915. It defines 'uid'/'userid' being a "Directory String" (LDAP syntax identifier 1.3.6.1.4.1.1466.115.121.1.15). "Directory String" is defined in RFC-4517 (LDAP: Syntaxes and Matching Rules).

Basically, a user name must be at least one character, with the characters being any from the UCS. UCS is equivalent to the the characters and char code points defined in Unicode without all the metadata and rules Unicode assigns to various characters (such as combining rules, for example).

That's all that LDAP defines. (Any further restrictions on uid/userid could potentially be imposed by the concrete directory service/server you are contacting, but that's outside the scope of LDAP itself.)

I am right now not having the energy to write a proper clean and self-contained answer. Feel free to look up the mentioned RFCs, and put the information together into a coherent answer, if you or anyone else feels like... ;-)

P.S.: The ldapwiki page you linked to is clearly stating that that page is about SAM user names, and not about ldap user names (uid/userid). SAM is a Windows-specific thing (as the page also mentions multiple times), not an LDAP-specific thing. You are going down the wrong road here with respect to the question you posed.

elgonzo‭ You're probably right there. I'm using the protocol Ldap to connect to an active directory. When I google it, I'm not the first one to mix this up. :D

Okay, if you are interfacing with an AD service, i would suggest to go with what the https://ldapwiki.com/wiki/SamAccountName page says. Although, you could probably simplify the regex by using look-ahead/look-behind assertions (depending on the capabilities of Java's regex engine). If you have a Windows box or VM at hand, you should be able to relatively easily verify the statements the page makes about the logon format name -- simply try creating a new user account, and try both exceeding the max. logon name length and using illegal characters (the Windows dialog for creating new user accounts should not allow you to create user names longer than 20 chars, and complain about illegal characters kinda like this: https://software.codidact.com/uploads/3sqPwnCHejcBwysfCbUggUeV)

And for good measure, here the "horse's mouth": https://docs.microsoft.com/en-us/windows/win32/adschema/a-samaccountname :-)

With regard to any library offering functions doing this format check for you, i honestly don't know. But it should be relatively simple and straightforward to implement yourself if you can't find some ready-to-use function, i guess.

(Note that Active Directory also maintains uid alongside samaccountname. uid is the user name, and not necessarily required to be equal to the NTLM logon name samaccountname.)