Hi Junio,

On Mon, 28 Jan 2019, Junio C Hamano wrote:

Johannes Schindelin [off-list ref] writes:

quoted

quoted

Two questions that would come to the readers' minds are

 - Why "it fails to detect the system prefix correctly"?  Is that a
   bug waiting to hurt end users?

I recall explaining that already in a different thread (one that actually
*was* about the RUNTIME_PREFIX feature, where you asked why we cannot test
for it in the test suite): to verify that we are in a valid Git
installation location,...
... But that was clarified already to your satisfaction in the RUNTIME_PREFIX
patches. Or do you want to reopen that case now?

This is not about me.  It is about those who read _this_ commit ---
if the change is justified to them clearly enough, then we'd be
happy.  I did not think it is with the proposed log message, and
that is where my comment came from.

You know what, I finally start to see your point.

Will adjust the commit message accordingly,
Dscho