RE: Git'ing a non-labeled set of sources

-----Original Message-----
From: Sparks, Sam 
Sent: Wednesday, August 08, 2007 3:46 PM
To: Peter Baumann
Cc: git@vger.kernel.org
Subject: RE: Git'ing a non-labeled set of sources

quoted

-----Original Message-----
From: Peter Baumann [mailto:waste.manager@gmx.de] 
Sent: Wednesday, August 08, 2007 3:17 PM
To: Sparks, Sam
Cc: git@vger.kernel.org
Subject: Re: Git'ing a non-labeled set of sources

On Wed, Aug 08, 2007 at 02:52:14PM -0500, Sparks, Sam wrote:

quoted

quoted

-----Original Message-----
From: Peter Baumann [mailto:waste.manager@gmx.de] 
Sent: Wednesday, August 08, 2007 2:37 PM
To: Sparks, Sam
Cc: git@vger.kernel.org
Subject: Re: Git'ing a non-labeled set of sources

On Wed, Aug 08, 2007 at 01:59:38PM -0500, Sparks, Sam wrote:

quoted

Hello All,

Please excuse me if this is an ignorant question; I'm new to 

git and my

quoted

have overlooked something in the documentation.

I'm attempting to obtain a snapshot of source code from an 

unlabeled git

quoted

branch in a public repository. I've found in the 

documentation that a

quoted

quoted

quoted

timestamp cannot be used to specify a particular version of 

source code,

quoted

but I believe I can work with the commit value as 

returned by 'git

quoted

quoted

quoted

quoted

show'.

However, I have been unsuccessful in my attempts to use this 

identifier

quoted

to clone or checkout the associated source tree. Has anyone been
successful in using git to successfully replicate an 

unlabeled version

quoted

of sources in a repository?

Here is my latest attempt:
/dir_i_want_to_replicate $ git show --pretty=short
commit 5b1313fb2758ffce8b624457f777d8cc6709608d
Author: ....

/replication_dir $ git clone git://www.denx.de/git/u-boot.git
u-boot-mpc83xx
Blah blah blah..
 100% (4378/4378) done
/replication_dir/u-boot-mpc83xx/ $ git checkout
5b1313fb2758ffce8b624457f777d8cc6709608d
error: pathspec '5b1313fb2758ffce8b624457f777d8cc6709608d' 

did not match

quoted

any. 

Because there is no 5b1313fb2758ffce8b624457f777d8cc6709608d in the
repo.

 $ git clone git://www.denx.de/git/u-boot.git
 Initialized empty Git repository in /tmp/u-boot/.git/
 remote: Generating pack...
 remote: Done counting 40938 objects.
 remote: Deltifying 40938 objects.
 remote:  100% (40938/40938) done
 Indexing 40938 objects...
 remote: Total 40938 (delta 32545), reused 39302 (delta 31040)
 100% (40938/40938) done
 Resolving 32545 deltas...
 100% (32545/32545) done

 $ cd u-boot; git-rev-list --all|grep ^5b1
 5b1d713721c3ea02549940133f09236783dda1f9

-Peter

Hmmm. There must be something I am not understanding about git.
When I run the same command on my existing git repository, I get the
following:
$ git-rev-list --all | grep ^5b1
5b1313fb2758ffce8b624457f777d8cc6709608d
5b1d713721c3ea02549940133f09236783dda1f9

Does git allow the removal of commits? I don't understand 

how I was able

quoted

to download it a couple of months ago, but it is no longer 

accessible.

quoted

Thanks for the quick reply,
Sam

If the branch was rebased, then yes. But this is not nice for
contributers, so it is normally the rule to not rebase a published
branch. But e.g. the branch 'pu' on git.git is a branch which will be
rebased. This is mentioned by Junio in his notes from the maintainer.

What made me wonder is, if you have the commit localy, why can't you
check it out? Please try a git-fsck --full run, to see if you have a
corrupt repo.

-Peter

I had to upgrade from git v1.4.4.4 to v1.5.0.7 to have access 
to the git-fsck command. 'git-fsck --full' printed out 
nothing, and 'git-fsck --full run' printed out an error 
message (error: invalid parameter: expected sha1, got 'run'). 
Was I doing something wrong?

--Sam

Peter,

I had an epiphany and discovered the problem: When I created the
repository I was trying to replicate, I didn't download it using git; I
simply copied it from another u-boot repository on the same server. 

When trying to use the commit number on a known good build, I didn't
have any problem replicating that branch.

Thanks for the help, and sorry it was a dumb error on my part.
--Sam