Re: [PATCH v7 0/3] RUNTIME_PREFIX relocatable Git

[Daniel Jacques <dnj@google.com>]

On Mon, Mar 26, 2018 at 2:01 AM Junio C Hamano <gitster@pobox.com> wrote:

> When testing the non-relocatable (i.e. traditional) Git, we use
> GIT_EXEC_PATH and bin-wrappers/ trick to ensure that we test the
> version we just have built, not a random version that happen to be
> on the $PATH, without requiring the built product first to be
> installed before being tested.  From the diffstat for this patchset,
> I am guessing that you are using the same mechanism even when
> testing the relocatable one.

I am - the tests will run against a runtime-prefix-enabled binary, but the
overrides supplied by the test suite will still be respected, so it should
generally behave like non-runtime-prefix in the tests.

> I wonder if the relocatable Git would allow a simpler arrangement to
> test without installing.

> I am asking merely out of curiosity, not suggesting to make a trial
> install somewhere in the build area and run the built Git normally
> without GIT_EXEC_PATH trick.

RUNTIME_PREFIX resolves paths relative to the runtime path of the Git
binary. These path expectations are constructed around installation
directories, so I'd expect that installation is a prerequisite of testing.
It's also worth noting that the runtime path resolution is
platform-dependent, so some Git-supported platforms will not be able to be
built and/or tested in this manner (at the moment). I only implemented
support for platforms I / Chromium uses, both because of my ability to test
on them and my own familiarity with them.

That said, if the test suite dropped its overrides and did a local
(testing) installation, that would be expected to work on the
RUNTIME_PREFIX binary. I assumed that one of the reasons we don't test
against an installation is b/c Git hard-codes its paths, but this would not
be a problem with RUNTIME_PREFIX enabled.

-Dan