Re: Fatal: <hash> is corrupted can be caused by wrong permissions

2 messages, 2 authors, 2016-06-15 · open the first message on its own page

Re: Fatal: <hash> is corrupted can be caused by wrong permissions

From: Matthieu Moy <hidden>
Date: 2016-06-15 23:02:46

"R. David Murray" [off-list ref] writes:
On git 1.7.1 I ran into the Fatal corruption error in one of my projects.  For
reasons that are irrelevant I was operating on the repo as user X, a member of
group Z, but the files were all owned by Y:Z, with (I thought) -rw-rw---
permissions.  After a bunch of head banging I finally figured out that the
Fatal error wasn't because the object files were corrupted, but because three
objects (for reasons unknown to me that are probably lost in the early history
of the setup of this particular repo...one of the blobs was the first commit)
were -rw-----.
I can reproduce with Git 1.7.10 by doing a chmod 0 on some object files,
but recent Git's produce

Checking object directories: 100% (256/256), done.
fatal: failed to read object ab2e06e74d922268fbff8d219dad9eee63786947: Permission denied

So I guess this has already been fixed :-).

-- 
Matthieu Moy
http://www-verimag.imag.fr/~moy/

Re: Fatal: <hash> is corrupted can be caused by wrong permissions

From: Jeff King <hidden>
Date: 2016-06-15 23:02:47

On Fri, Oct 24, 2014 at 04:05:24PM +0200, Matthieu Moy wrote:
I can reproduce with Git 1.7.10 by doing a chmod 0 on some object files,
but recent Git's produce

Checking object directories: 100% (256/256), done.
fatal: failed to read object ab2e06e74d922268fbff8d219dad9eee63786947: Permission denied

So I guess this has already been fixed :-).
Probably d6c8a05 (open_sha1_file: report "most interesting" errno,
2014-05-15). The example in the commit message even uses "chmod 0". :)

It's in v2.0.1.

-Peff
Keyboard shortcuts
hback out one level
jnext message in thread
kprevious message in thread
ldrill in
Escclose help / fold thread tree
?toggle this help