Re: [bug] Possible Windows 'git mv' bug

[Johannes Schindelin <Johannes.Schindelin@gmx.de>]

[-- Attachment #1: Type: text/plain, Size: 2232 bytes --]

Hi Aaron,

On Sun, 31 Jan 2016, Torsten Bögershausen wrote:

> On 2016-01-31 15.03, Aaron Gray wrote:
> > 
> > I think I have found a possible difference in behaviour between
> > Windows git commandline distro and Linux git
> > 
> > basically If I do a :-
> > 
> >     git mv logger.h Logger.h
> > 
> > I get the following :-
> > 
> >     fatal: destination exists, source=lib/logger.h, destination=lib/Logger.h
> > 
> > It looks and smells like a bug to me !

Thanks for your bug report. Having said that, I would like to suggest a
couple improvements for future reference:

Git for Windows' home page (https://git-for-windows.github.com/#contribute)
asks explicitly to please open tickets on Git for Windows' bug tracker
(https://github.com/git-for-windows/git/issues).

Also, at least one crucial bit of information is missing: the Git for
Windows version. Typically this is enough, but the Windows version is
often also crucial. To prevent crucial bits of information from missing, I
highly recommend volunteering information up front (in my mind, a good
rule of thumb is: spend as much care crafting your bug report and thinking
about what to include as you would wish the combined readership would
spend on helping you). If I may say so, I provided a good list of advices
to help crafting bug reports:
https://github.com/git-for-windows/git/wiki/Issue-reporting-guidelines

> Which version of Git are you using ?
> Because it is fixed in the latest version in Git and Git for Windows.

Another piece of advice I detail in our issue reporting guidelines is to
search the existing bug tracker (one purpose of the tracker is to keep
track of what has been reported and what has not yet been reported, after
all).

And indeed, I opened this ticket:

	https://github.com/git-for-windows/git/issues/419

... whose initial report incidentally also lacks the crucial information
mentioned above! So I have to step up my own game, too.

This report *also* includes detailed information, though, in particular
that Torsten fixed this issue a long time ago, which suggests to me that
you are somehow still stuck with the 1.x train (run `git version` to find
out).

Ciao,
Johannes