Populating /usr/tests

Julio Merino julio at meroh.net
Sun Oct 20 02:52:14 UTC 2013

On Sat, Oct 19, 2013 at 4:49 PM, Garrett Cooper <yaneurabeya at gmail.com> wrote:
> You're missing a lot of code in etc/, some in Makefile.inc1, and other pieces in share/mk as well with bsd.own.mk. Also, your approach with tests/Makefile is ok, but it doesn't partition well with prove.test.mk on my branch, etc.

Hello Garrett,

It's not obvious to me what you mean by "missing a lot of code" so I'm
going to have to guess. Here are the two possibilities that come to

The first is: the patches are missing something because they won't
behave in the intended way (where intended = what their description
says). If that's the case, I really don't know what's missing and I'd
like you to elaborate a bit more please. The patches seem to work and
they do what their description says (no more and hopefully no less).
Keep in mind that the scope of these patches is quite restricted and
their only goal is "plug existing tests into /usr/tests".

The second is: the patches are missing something because they don't do
everything you already have on your branch. This is certainly the
case. I'm not intending to cover all cases at once: it's easier to go
step by step with small and easy to review proposals. Additional
features can be built on top later and existing code can still be very
easily tweaked.

Now, because I think the second case is the most likely of the two...
I'm still not really sure about what exactly you are referring to! I
took a look at your branch and I can't see major differences in the
basic code to bring /usr/tests up and running. One thing I found is
that you have a tests-specific mtree file, which is a nice idea and I
have brought that into the patchset. The other things that I could
see... well, they may be necessary later on, but they don't seem to be
required for this specific patchset. (In other words: they have to be
looked at, but they can probably be looked at separately on their

Along these lines: do you have any suggestions on how to extract the
set of changes in your branch that are not in head? I'm not talking
about a diff: I'm talking about a collection of patches (or commits)
describing the differences between HEAD and your branch in a
semantical manner. Looking at a snapshot of the tree and comparing
differences is really difficult and does not give any idea on how the
changes were built. (Sorry, I don't know git enough nor your
repository to do this easily; I believe it's possible, but it escapes
me how.)


Julio Merino / @jmmv

More information about the freebsd-testing mailing list