# Bad object pointed under refs/head/ 2 messages from 2015-02-13 to 2015-02-14. Participants: Zheng Zhang, Michael Haggerty. Thread: https://gitlist.dev/t/38561 ## Zheng Zhang, 2015-02-13 10:06 Subject: Bad object pointed under refs/head/ Message-ID: URL: https://gitlist.dev/e/CAN6kTyfiw6QPk469xV6gyL-%2BYB5Ca3vNHhEX7daiyn7j536SXQ%40mail.gmail.com ``` Hi, I was running some test with Git 1.8.4.5, then I accidentally met a problem that leaded to the following error, > > error: refs/heads/develop does not point to a valid object! Turns out that the sha in refs/heads/develop is a bad object id, this happened after merging a branch X to branch develop, but packed-refs is updated to a corrected sha. No other merges at that point. The fix is easy, just removed refs/heads/develop. So there were two sha created, one is updated to refs/heads/develop, and the other one which is corrected, updated to packed-refs, Weird. I am wondering if there is a way to prevent this happening? Is this an ancient bug? Many thanks -- GitCafe.com Share a cup of open source Zhang Zheng @simsicon ``` ## Michael Haggerty, 2015-02-14 06:19 Subject: Re: Bad object pointed under refs/head/ Message-ID: <54DEE8DE.9060501@alum.mit.edu> URL: https://gitlist.dev/e/54DEE8DE.9060501%40alum.mit.edu In-Reply-To: ``` On 02/13/2015 11:06 AM, Zheng Zhang wrote: > I was running some test with Git 1.8.4.5, then I accidentally met a > problem that leaded to the following error, > >>> error: refs/heads/develop does not point to a valid object! > > Turns out that the sha in refs/heads/develop is a bad object id, this > happened after merging a branch X to branch develop, but packed-refs > is updated to a corrected sha. No other merges at that point. > > The fix is easy, just removed refs/heads/develop. > > So there were two sha created, one is updated to refs/heads/develop, > and the other one which is corrected, updated to packed-refs, Weird. > > I am wondering if there is a way to prevent this happening? Is this an > ancient bug? If you can find and document a way to reproduce this problem, then you will have a much better chance of finding somebody who is willing to look into it. Otherwise it's hard to know how to even get started. Michael -- Michael Haggerty mhagger@alum.mit.edu ```