Re: git bug/feature request

From: Peter Karlsson <peter@softwolves.pp.se>
Date: Nov 27, 2007, 14:35 UTC
Message-ID: <Pine.LNX.4.64.0711271531110.1011@ds9.cixit.se>
In-Reply-To: <200711271127.41161.gapon007@gmail.com>

gapon:

> i have discovered "weird" behaviour of git in this scenario*:

Yeah, I have run into it several times myself, and that is being both user A and B at the same time. The problem seems to be that git allows you to push into a repository which has a check-out, causing it to change states in a subtle way. That's just plain broken.

Git should either handle it somehow (perhaps by forcing the push into a new branch, which the pushee needs name), or just plainly refuse to push into a repository with a check-out.

I have learned to work around this problem by always pulling between my repositories, not pulling. I could probably have worked around it by having a master repository that is bare, but I have found that difficult because I am tracking an upstream non-Git repository, so to push and pull changes from that, I need a repository where I can have a check-out.

> * yes, i know that this scenario is "incorrect" but... it's possible
> and therefore i think it should be somehow handled - i tried a
> similar one with hg and bzr and i like their behaviour more

Yeah. It's even more irritating that recovering from the error state is difficult as well.

-- 
\\// Peter - http://www.softwolves.pp.se/