...

...

I notice many modules no longer have a ".git" subdirectory.

That would cause problems, but I would suspect a haywire script or similar on your end has simply deleted those directories. Running "git submodule init" and "git submodule update" from within the main tde/ directory should restore those directories and reset your copy to the latest GIT HEAD.

Please let me know if you run into any further issues; for obvious reasons I am concerned at even the slightest hint that the main GIT repository could possibly be corrupt!

I have not noticed any problems with GIT. I even tried to clone a complete tree and no problems I noticed (except for the usual slow speed). Maybe you have just broken your local copy.

You can try: git fsck && git submodule foreach "git fsck"

Note: Submodules may not have .git folder, but may have .git file that points to the location git folder in the top level object (in .git/modules).

I have tried all of the above to no avail. The git fsck commands results in several "dangling blob," dangling tree," and "dangling tag" messages.

When I attempt to resync my entire local tree, the process always hangs at kmplayer. When I enter that directory and run 'git pull' I receive the following message:

You are not currently on a branch, so I cannot use any 'branch.<branchname>.merge' in your configuration file. Please specify which remote branch you want to use on the command line and try again (e.g. 'git pull <repository> <refspec>'). See git-pull(1) for details.

I get the same error message from several modules, such as tdeadmin.

This all started last night when I wanted to push a simple editorial patch for tdeadmin for the kcontrol network module to recognize the newly released Slackware 14.

I have tried to enter each defective module and run 'git checkout master' but that does not help. The process just hangs with no messages at all.

I've tried following the directions in the wiki. I ran 'git clone' for the entire tree and that has not helped.

What do I try next?

Darrell

Have you checked your hard drive recently? Any disk issues could cause a mess in GIT, and I would argue could be likely to show up in GIT before anything else simply because the entire GIT tree uses many tens of thousands of small files.

Nothing wrong with the drive. Just yesterday when booting the scheduled fsck ran on that partition.

I was able to sync the tree after moving the contents of the kmplayer directory. Sort of --- nothing filled in to replace the missing files. I still get the "You are not currently on a branch..." messages with many modules.

I don't want to waste 4+ GB of bandwidth recreating a whole new local tree, but I don't know what else to do. Not to mention that with a new tree none of the modules will have my user name in the config files.

Darrell

Check for filenames that might have accidentally had a carriage return inserted, or space entered at the front or end of the file name. That has caused me very similar problems to what you describe in the past when using CLI tools.

Konqueror can search for a leading or trailing space: ( *) or (* ), don't use the (). I don't know (other than visually) how to search for a carriage return.

Keith

On Sat, Oct 6, 2012 at 2:15 PM, Darrell Anderson humanreadable@yahoo.com wrote:

...

Have you checked your hard drive recently? Any disk issues could cause a mess in GIT, and I would argue could be likely to show up in GIT before anything else simply because the entire GIT tree uses many tens of thousands of small files.

Nothing wrong with the drive. Just yesterday when booting the scheduled fsck ran on that partition.

I was able to sync the tree after moving the contents of the kmplayer directory. Sort of --- nothing filled in to replace the missing files. I still get the "You are not currently on a branch..." messages with many modules.

I don't want to waste 4+ GB of bandwidth recreating a whole new local tree, but I don't know what else to do. Not to mention that with a new tree none of the modules will have my user name in the config files.

Darrell

---

To unsubscribe, e-mail: trinity-devel-unsubscribe@lists.pearsoncomputing.net For additional commands, e-mail: trinity-devel-help@lists.pearsoncomputing.net Read list messages on the web archive: http://trinity-devel.pearsoncomputing.net/ Please remember not to top-post: http://trinity.pearsoncomputing.net/mailing_lists/#top-posting

On 6 Oct 2012, Darrell Anderson uttered the following:

I have tried all of the above to no avail. The git fsck commands results in several "dangling blob," dangling tree," and "dangling tag" messages.

These are not signs of fs corruption unless you get lots of them: they just mean that some content is inaccessible. This can happen legitimately, e.g. a rebase followed by the usual automatic gc prune of the reflog.

When I attempt to resync my entire local tree, the process always hangs at kmplayer. When I enter that directory and run 'git pull' I receive the following message:

You are not currently on a branch, so I cannot use any 'branch.<branchname>.merge' in your configuration file. Please specify which remote branch you want to use on the command line and try again (e.g. 'git pull <repository> <refspec>'). See git-pull(1) for details.

You should probably be using 'git pull' in the top-level followed by 'git submodule update'. When a submodule is checked out in its parent, it is pointing at a single specific commit, thus is not on any branch (it's at a point in the past history of some branch instead). So 'git pull' isn't going to be happy when it reaches its merge phase and tries to merge its recent changes into what it has already seen.

This is harmless, and is indeed expected when you are not on a branch. The hang.. the hang is *not* expected.

I get the same error message from several modules, such as tdeadmin.

It's not an error -- not exactly, anyway. If you just want to fetch and not update, you should be using 'git fetch' or 'git remote update'; if you want to update a submodule, you should be using 'git submodule update'. There is little reason to do a 'git pull' in a submodule.