[RFC] depend on testtools for testing?

[Martin Pool]

2009/10/29 Robert Collins <robertc at robertcollins.net>:
> Are there folk who object to adding a hard dependency on testtools for
> 'bzr selftest', given the above? Please speak up - I don't want to do
> anything that makes developing or using bzr harder [but I also want to
> reduce duplicated code :)]

As I said to Robert yesterday, to me, given the technical environment,
having the code factored out is approximately in balance with the
hassle a new dependency creates.

(I feel like this is going to be bait for someone to argue that it is
technically possible to have an external dependency without problems.
Sure.  But given the state of the art for Python, and for cross-OS
packaging, it does in practice cause problems.  If we guaranteed that
you could always install bzr and all its dependencies on any platform
using virtualenv (or whatever), then it would be an easier call - if
the stuff you need is not packaged, just use that.)

I do think it's a good sign that the proposed dependency is actively
maintained and (one hopes) regularly released by a sensible person.
On the other hand jml just got a demanding new job.  So I would be
more willing to do patches to testtools trusting that they'd get out
soon.

On the other hand, even for subunit which has a moderately sane
maintainer, getting a feature in did take noticeably longer than it
might have in bzr.

I agree with John's summary of some of the specific problems, and am
also ambivalent about taking such a change, defaulting to no.

Things that would move me away from that default:

 1- folks like Alexander who will be bitten about this saying "it's
really not a problem and a great idea" (or the opposite)

 2- the new dependency having actual compelling feature improvements
beyond what's in Bazaar

-- 
Martin <http://launchpad.net/~mbp/>