This is the mail archive of the
mailing list for the Cygwin project.
Re: [cygwin] Re: git fails to push
- From: Peter Rosin <peda at lysator dot liu dot se>
- To: cygwin at cygwin dot com, adam at dinwoodie dot org
- Date: Thu, 6 Aug 2015 22:46:15 +0200
- Subject: Re: [cygwin] Re: git fails to push
- Authentication-results: sourceware.org; auth=none
- References: <5502264B dot 3050803 at lysator dot liu dot se> <C31449BA-B14C-4047-981B-6980AB755309 at etr-usa dot com> <5508284B dot 1040304 at lysator dot liu dot se> <E043FEB21B394020B5EA3AFC2ED69308 at black> <55083FC1 dot 4010503 at lysator dot liu dot se> <20150317171011 dot GG20741 at dinwoodie dot org>
On 2015-03-17 18:10, Adam Dinwoodie wrote:
> On Tue, Mar 17, 2015 at 03:52:49PM +0100, Peter Rosin wrote:
>> Interesting. But the patch may not be needed. The easy thing to
>> do seems to be to either use core.createobject = rename
>> or to convince some git :-) that Cygwin also needs
>> OBJECT_CREATION_USES_RENAMES = UnfortunatelyNeedTo
>> in its section in config.mak.uname.
> With my Git maintainer hat on, I'm not going to include that change in
> the Cygwin distributed version of Git unless and until I hear reports of
> `git config core.createObject rename` helping more people.
> This is clearly not a common problem (it's the first I've heard of this
> sort of problem, and I've been using Git on Cygwin on NTFS for years
> without hitting these issues), so I'm very reluctant to inflict the
> disadvantages of using the "rename" object creation on everyone unless
> it's going to fix problems for enough people to make those disadvantages
> (As it is, we haven't had verification that this change does anything
> anyway, but if it doesn't, the entire point is irrelevant...)
Sorry to reply this late to this old thing...
...but anyway, here are some further evidence of problems caused
by not always going with rename.
It seems like some people are migrating to msysgit over this issue.
Problem reports: http://cygwin.com/problems.html
Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple