On Mon, Oct 03, 2016 at 01:06:10PM +0200, Johannes Schindelin wrote: > Hi, > > On Tue, 27 Sep 2016, Jeff King wrote: > > > On Tue, Sep 27, 2016 at 10:05:37AM -0700, Junio C Hamano wrote: > > > > > The subject says it all. Would it be bad if we introduce an > > > environment variable, GIT_SYSTEM_CONFIG=/etc/gitconfig, that names > > > an alternative location of the system-wide configuration file? > > > > > > That would supersede/deprecate GIT_CONFIG_NOSYSTEM that we > > > introduced primarily so that we can run our tests without getting > > > affected by the configuration that happens to be effective on the > > > host that the test is being run. > > > > I can't think of a reason it would be bad. > > I cannot think of any reason right now, either, but my gut tells me that > this needs to simmer a while in the backs of our minds, to give potential > reasons a chance to come forward. > > What would be the use case, BTW? IOW what would it solve that cannot > already be solved by using XDG_CONFIG_HOME? The patches Junio posted later use it for the test suite (and I also have had to skip some tests in the past related to system config because of its lack). I would also use it when doing git experiments on GitHub servers. We keep several relevant config settings in /etc/gitconfig, so if I were to say, build a new version of git and test how it repacked torvalds/linux, I need to make sure it picks up the same config. Usually I do it by baking in the right /etc/gitconfig at build time, but it would be less annoying to be able to override it at run-time. I admit both of those are uses for git _developers_, though, not git _users_. -Peff