Re: Git Stability

[Thomas Adam <thomas.adam22%gmail.com@localhost>]

On Wed, Jan 13, 2010 at 09:34:21PM +0100, S.P.Zeidler wrote:
> Thus wrote Curt Sampson (cjs%netbsd.org@localhost):
> 
> > On 2010-01-13 12:25 +0100 (Wed), S.P.Zeidler wrote:
> > > The update process got interrupted and left refs/heads empty.
> > 
> > Err, *what* update process? You're rather parsimonious with information,
> > which is probably why I don't understand your problems most of the time.
> 
> http://mail-index.netbsd.org/tech-repository/2010/01/11/msg000227.html

So -- you're referring to having to update this Git repository on a daily
basis I assume, that's adding this load?  Can you elaborate on exactly
what's going on here, as not only do I personally not understand what's
happening, but I suspect a lot of other people don't either.

> http://mail-index.netbsd.org/tech-repository/2009/10/18/m
> 
> > So I'm mystified by your comment that you lose the git repo if the
> > "index" is corrupted.
> 
> -an- index.
> How would you describe the function of the files in refs/heads if not
> as an index (specifically, to the head of the respective branch)?

But it's not the head of a respective branch at all, it's "global" to the
repsository, which is why you can do thinks like:

% git checkout foo
[hack, hack, hack]

# Ah, no, I meant to be on bar:
% git checkout bar
% git commit -am "...."

But, semantics aside, you wouldn't get an index in a bare repository --
which is the respository I assume we're talking about here -- all that is
doing is receiving pack files.

Again -- I suspect this part will become more clear as we get more
information, as I don't want to get side-tracked with semantics when the
problem isn't well defined.

-- Thomas Adam

-- 
"It was the cruelest game I've ever played and it's played inside my head."
-- "Hush The Warmth", Gorky's Zygotic Mynci.