pombreda / egit

Automatically exported from code.google.com/p/egit
0 stars 0 forks source link

InvocationTargetException: Missing Tree #114

Closed GoogleCodeExporter closed 9 years ago

GoogleCodeExporter commented 9 years ago
I working against a USB Drive.

All I am doing is fetching, pushing and committing to SVN.

This is the 2nd time that I get the same issue.  

Eclipse pops up a dialog indicating "InvocationTargetException" when I
attempt to checkout the project from a USB drive.  [Which had been pushed
to previously]

Looking at the logs provides more insight.  (Missing Tree)

Questions:  

The dialog that appears should likely be a bit more descriptive than
InvocationTargetException.  (ie:  Saying your GIT repo is corrupt)
[http://osdir.com/ml/git/2009-03/msg00019.html]

I commit to my local repo on my local PC, then push to the USB Drive.  What
might be the reason I am getting such corruption?  

I am solely using EGIT via Eclipse.

Original issue reported on code.google.com by eric.gul...@gmail.com on 4 Sep 2009 at 2:53

Attachments:

GoogleCodeExporter commented 9 years ago
What steps will reproduce the problem?
No real steps.  All that was done are git pulls, pushes and commits.  

What is the expected output? What do you see instead?
Well in the fist place, my GIT repo gets corrupted.
Secondly the error dialog isn't very descriptive.

What version of the product are you using? On what operating system?
Eclipse Git Feature 
(Incubation)    0.5.0.200908282229  org.eclipse.egit.feature.group
Windows XP.

Please provide any additional information below.

Original comment by eric.gul...@gmail.com on 4 Sep 2009 at 5:35

GoogleCodeExporter commented 9 years ago
Also reported at http://bugs.eclipse.org/286653

Original comment by robin.ro...@gmail.com on 5 Sep 2009 at 9:05

GoogleCodeExporter commented 9 years ago
This is too vague of a problem description to really do anything about it.  
Yes, the 
error message can be improved, but the corruption is most likely caused by 
JGit's 
push operation.  I'd need samples of the two repositories, or a list of the 
objects 
involved, to try to better understand how JGit is omitting a tree during the 
push.

Where is the tree 573c14f2fc83f658a9f12dffb6aa1fde11e5f1c8 used by your 
repository?  
Try something like:

  git rev-list --objects --all | grep 573c14f2fc83f658a9f12dffb6aa1fde11e5f1c8

in the source (aka non-corrupt) repository to see if it is a root level tree 
(has no 
path after it), or a subtree (has a path after it).  Can you narrow down what 
commit 
introduced this tree to your project?  Where does that commit fall in relation 
to the 
push you were trying to do?  Was it the last (oldest) commit to be pushed, or 
was it 
the first, or was it somewhere in the middle of the batch being pushed?

Original comment by sop+code@google.com on 8 Sep 2009 at 2:56

GoogleCodeExporter commented 9 years ago
Problem is both my repos get corrupted.  I first commit to the hard drive, then 
push
to the USB drive.  

I will first need to find a non corrupt version...  And try to get you a list 
of the
objects involved.  My expectation is that it was the latest commit.   Since it 
was
working before?!  :)  I'll see if I can find a working repo.  
In terms of your questions about commits, do you know what commands would 
provide
that info?

Original comment by eric.gul...@gmail.com on 8 Sep 2009 at 4:18

GoogleCodeExporter commented 9 years ago
I was able to run it in the 'corrupt' repo.  
git rev-list --objects --all > er.txt

There is only ONE reference to that hash.  
It does have a path after it.  It's one of the java packages which has all my 
source
code under it.

-  Should I try doing this on the non corrupt version?  Does it mean anything 
that I
can do this on the 'corrupt' one?
-  Do you know what commands I would have to do to figure the answers to your 
commit
questions?

Original comment by eric.gul...@gmail.com on 8 Sep 2009 at 4:25

GoogleCodeExporter commented 9 years ago
I created a new REPO, same things happened.

git fsck --full
broken link from    tree 1f8e9475250ff660b4bd967d3bd227600ed42a15
              to    tree a17d7ea31681746c7b4452a5b0238d59226f5211
missing tree a17d7ea31681746c7b4452a5b0238d59226f5211

git fsck --full 1f8e9475250ff660b4bd967d3bd227600ed42a15
broken link from    tree 1f8e9475250ff660b4bd967d3bd227600ed42a15
              to    tree a17d7ea31681746c7b4452a5b0238d59226f5211
dangling commit 34b927950767bf06b994b0ddbfa21e63fdf67178
dangling commit ff2c7f8b91ca6ae5d4d7f2342684d1733fd734c5
missing tree a17d7ea31681746c7b4452a5b0238d59226f5211

git ls-tree 1f8e9475250ff660b4bd967d3bd227600ed42a15
040000 tree 7294d799874a00d03043aada204277b93730b5f2    main
040000 tree a17d7ea31681746c7b4452a5b0238d59226f5211    test

git show 34b927950767bf06b994b0ddbfa21e63fdf67178
fatal: unable to read destination tree 
(a17d7ea31681746c7b4452a5b0238d59226f5211)

git show ff2c7f8b91ca6ae5d4d7f2342684d1733fd734c5

<shows me a commit>

I read the thread http://bugs.eclipse.org/286653

Unfortunately I am new to GIT, however, I would express that this does seem to 
be a
egit/jgit issue since I created this repo from scratch as well.  

Original comment by eric.gul...@gmail.com on 20 Sep 2009 at 3:18

GoogleCodeExporter commented 9 years ago
Issue 118 has been merged into this issue.

Original comment by sop+code@google.com on 5 Oct 2009 at 10:28

GoogleCodeExporter commented 9 years ago
Probably fixed in the references bugzilla entry

Original comment by robin.ro...@gmail.com on 12 Jul 2010 at 4:57