Re: Allowing override of the default "origin" nickname

From: Johannes Schindelin <johannes.schindelin@gmx.de>
Date: Jan 11, 2008, 13:52 UTC
Message-ID: <alpine.LSU.1.00.0801111348230.31053@racer.site>
In-Reply-To: <30e4a070801110506h19b77488sbfa6ae48701f30f9@mail.gmail.com>

Hi,

On Fri, 11 Jan 2008, Mark Levedahl wrote:

Show 14 quoted lines

> On Jan 11, 2008 7:03 AM, Johannes Schindelin <Johannes.Schindelin@gmx.de> wrote:
>
> > IIUC your patch only replaces the term "origin" by something 
> > user-specifiable.  I fail to see the use of it; care to explain what 
> > workflow is easier with your patches than without?
> 
> Consider a project with several servers, each of which is *supposed* is 
> host the same project but due to lags in mirroring across airgaps, etc., 
> are never quite in sync. Now, we get on a teleconference and discuss 
> issues, find differences, and everyone reports that "origin" has x, but 
> in fact "origin" means different servers to almost everyone. Also 
> consider that only a small subset of the group really understands git, 
> most just follow cook-book recipes to get their work done and don't 
> understand what is going on.

Okay, I now understand your intention much better than before.

> This is my problem: "origin" is an abstraction that hides the different 
> server names in use and makes communication difficult, having everyone 
> use nicknames related to their particular upstream server reduces the 
> confusion.

Unfortunately, I think this will just lead to even more confusion. Because those people following recipes without thinking will now use a name that does not even say the role, let alone the server.

IOW I think that your patch worsens the situation you describe.

IMHO you should optimise the communication by agreeing on one origin, or alternatively not talk about a server at all (which is made easy by the global uniqueness of commit names; just say "my tip is ac9b7192").

Ciao, Dscho