Re: [PATCH] user.default: New config to prevent using the default values for user.*

From: Junio C Hamano <gitster@pobox.com>
Date: Mar 8, 2008, 02:08 UTC
Message-ID: <7v7igeynl5.fsf@gitster.siamese.dyndns.org>
In-Reply-To: <alpine.LSU.1.00.0803080223030.3975@racer.site>

Johannes Schindelin <Johannes.Schindelin@gmx.de> writes:

Show 24 quoted lines

> On Sat, 8 Mar 2008, Santi Béjar wrote:
>
>> On Fri, Mar 7, 2008 at 5:41 PM, Alex Riesen <raa.lkml@gmail.com> wrote:
>> > Santi Béjar, Thu, Mar 06, 2008 22:45:43 +0100:
>> >
>> > > On Wed, Mar 5, 2008 at 9:44 PM, Alex Riesen <raa.lkml@gmail.com> wrote:
>> >  > > Santi Béjar, Wed, Mar 05, 2008 20:18:04 +0100:
>> >  > >
>> >  > >  > @@ -171,7 +171,7 @@ static const char au_env[] = "GIT_AUTHOR_NAME";
>> >  > >  >  static const char co_env[] = "GIT_COMMITTER_NAME";
>> >  > >  >  static const char *env_hint =
>> >  > >  >  "\n"
>> >  > >  > -"*** Your name cannot be determined from your system services (gecos).\n"
>> >  > >  > +"*** Your name cannot be determined.\n"
>> >  > >
>> >  > >  Why not?
>> >  >
>> >  > Is this important? Or in another way, is this useful? The important
>> >  > thing is how you can fix it. But others think otherwise I'll change it
>> >  > to explain the reason.
>> >
>> >  It is precise explanation of what happened. It could be a hint to fix
>> >  gecos field to someone. It is considered useful not to hide
>> >  information, even if you have no idea of how useful it is.

I'd agree with that reasoning on one condition, that is, msysgit folks would adjust this to whatever gecos equivalent is called in the Windows world ;-).

I do not personally think this part of the patch is so bad to waste your brain cells and time arguing about. I would even suggest removing the whole "cannot be determined" and just say "Please tell me who you are."

Yes, if you tell people that a lacking GECOS is the underlying reason that you are getting this message, it would hint them an alternative solution of talking to their sysadmin to get it fixed, and that one-shot fix will last for all repositories on that machine (and machines that consult the same NIS/LDAP for name information). However, the "config --global" suggested in that message is also a valid one-shot fix, and hopefully one-shot across machines that NFS mounts the home directories. The latter knowledge can hopefully be reused when you are forced to use git on Windows, but the fix based on GECOS would probably not.

So I'd even argue that Santi's change to the message is an improvement that removes geekspeak that is not particularly useful to the end user in this context.

However, I would also say that this message improvement does not belong to user.default or whatever the configuration variable is called in this round of proposal. It should be a separate patch, and it would be a much easier sell than the rest of the patch series.