> On 16 Apr 2019, at 16:45, Boudewijn Rempt <boud@valdyas.org> wrote:
>=20
>> On dinsdag 16 april 2019 21:38:04 CEST Ben Cooksley wrote:
>>=20
>> This hook was implemented in the first place to ensure that people had
>> correctly setup Git on their local machine.
>>=20
>> On some versions of Git (maybe all?) it will automatically use the local
>> user account name as the name.
>>=20
>> This leads to people committing as "me", "user" and "nobody" without
>> meaning to, but which still leads to a situation in which the metadata of=
 a
>> commit has ended up being useless.
>>=20
>> I'd rather maintain a small list of exceptions for those who do have name=
s
>> without a space in them to ensure that for the vast majority of our users=

>> do correctly get informed they need to fix their local setup.
>=20
> You could do a small blacklist of known wrong names like the ones you cite=
 -- but you will never be able to implement a hook that identifies a string a=
s a name correctly. I'm sorry -- but it's just impossible.=20
>=20
> It's also just not good manners to tell people their name isn't a real nam=
e: and I think that allowing ourselves to accept every name except for thing=
s like me, user, nobody, admin, root is more important than making sure we'v=
e got correct metadata.=20
>=20
> Because none of us can actually be sure the metadata is correct anyway: th=
ere's not just the impossibility of creating that identifies a string as a n=
ame correctly, humans cannot do that either.

A blacklist won't work. The common wrong name due to misconfiguration would b=
e if I commit as "nicolas" or "nalvarez" instead of "Nicol=C3=A1s Alvarez". T=
hat's a more common error than "user" or "root".

It seems easier to whitelist legitimate mononyms on request...

--=20
Nicol=C3=A1s=