Broken commit 5e6ecdad9f69b1ff789a17733b8edc6fd7091bd8
anatolyborodin opened this issue · comments
"git fsck" thinks this commit is broken:
error in commit 5e6ecdad9f69b1ff789a17733b8edc6fd7091bd8: invalid author/committer line - bad time zone
commit 5e6ecdad9f69b1ff789a17733b8edc6fd7091bd8
tree 0d1971c86d047f9530aee85a50b8fc1ca12723ba
parent 9471b0ab889a4684f87952cb951f7f4dc3f59dda
author Shrikant Sharat Kandula <...> 1313584730 +051800
committer Shrikant Sharat Kandula <...> 1313584730 +051800
Would it be possible to repair the history?
No it isn't possible to repair the history. That will change the SHA value of every commit after it and update the timestamps on ALL of them. It will invalidate every tag since that commit and it will cause far too many headaches.
To be clear, it is technically possible to repair this, also if repairing it didn't affect the repository's history from that point on, I would be very much in favor of it, but it irrevocably breaks expectations around our history and will cause more trouble than not. You appear to be the first person to have checked this and found it to be an issue.
Even if git fsck reports the commit as being damaged, git continues to work just fine. This doesn't realistically affect the use of the repository or any person's ability to contribute to it. I appreciate you reporting this, but it simply isn't something we're going to fix.
You appear to be the first person to have checked this and found it to be an issue.
Oh, really? I thought I'm just too lazy to find an old and closed issue.
I just happen to have "transfer.fsckObjects = true" in my ~/.gitignore...
@sigmavirus24, yep, this looks like a very popular repo. Forking just for one commit would be a solution, but... Ok, I'll just disable the checks in .git/config after the checkout.
@anatolyborodin Recent versions of git (master, but not yet in any released version) have support for configurable fsck checks. So you will be able to set receive.fsck.badtimezone = warn, for example, if you would like the usual checks, but would prefer to ignore this particular (minor) problem.