This is correct basically all correct.
The FreeBSD people in particular have a particularly strong reason to
prefer having a distinction between build_requires and test_requires.
The impression I got was that this was a good idea, but because it
wasn't immediately pressing, it needed to go through the same process as
configure_requires:
That is, to be chewed over, and be test-implemented by a couple of the
pieces, and to have sufficient time to digest the problem and let all
the parts of the toolchain be sure they can support it properly.
Personally, I'd like to see configure_requires: finalised first before
we start dealing with test_requires, so that we have the option of
solutions that themselves leverage configure_requires:
Adam K
We had a number of package maintainers at Oslo from Debian and FreeBSD
so actually there was a lot of discussion about situations when
running tests is useful versus not. There was substantial, thoughtful
debate about different "contexts" for testing -- when certain kinds of
tests would be run and others not (e.g. development testing, release
testing, packaging testing, install testing).
So it wasn't at all a knee-jerk, "always test" mentality.
I would characterize the outcome as "lack of consensus" on whether
"test_requires" (etc) benefits outweigh costs, and the general
(consensus) view was not to recommend changes unless there was a
consensus in favor.
Regards,
David
