[kforge-dev] Changeset [1248]: [test][m]: add in kforge-test-nose script to run tests using nose. * As …

Rufus Pollock rufus.pollock at okfn.org
Fri May 23 10:20:10 UTC 2008


On 21/05/08 12:11, John Bywater wrote:
> KForge: Timeline wrote:
>> [test][m]: add in kforge-test-nose script to run tests using nose.
>>   * As nosetests just pulls in all tests put in disable on tests we do 
>> not want to run normally (in particular tests that required command 
>> line interaction).   * NB: still running quite a few test we probably 
>> do not want to run by default but none that actually break automated 
>> testing in the way that cli stuff does.
>>
>> URL: http://project.knowledgeforge.net/kforge/trac/changeset/1248
>>   
> 
> I'm reluctant to complain about change simply because of the commit time 
> (3am+) ;-)

You do realise I was in the US hence 5/8 hrs behind (I only just got 
back ...)?

> How can I put this? :-)
> 
> I'm not opposed to nosetests, as such. I'm sure it makes some people happy.
> 
> However, as already expressed, I am opposed to: nosetests + KForge.

I know and I thought when we discussed at our last in person meeting 
that I'd brought you round to this but I obviously misconstrued the 
discussion. As I emphasized then nosetests *is* backwards compatible -- 
the changes made in this commit simply make it easier to exclude certain 
tests and do not (to my knowledge) in any compromise the existing test 
framework ...

> 1. We are focussing on releasing KForge 1.0. We are focussing on 
> *stablising* things. There are other, much greater priorities, such as 
> picking up the Debian package, which should receive attention first.

I don't think these are mutually exclusive. Frankly I have no interest 
in messing around with our test suite -- it is just that to do other 
things I do want to do involve writing tests and using nosetests is just 
a lot more pleasant ...

> 2. KForge hasn't been developed with nosetests and therefore switching 
> test harness framework is objectively destablising, and therefore 
> technically foolish; nosetests doesn't appear to respect the OO 
> traditions, of which unit testing suites is one aspect, and there isn't 
> considered by me to compatible with the KForge project.

A lot of people including many OOers do use test discovery tools. Should 
we write to Martin Fowler to get a ruling from 'on high'? I've long been 
wanting something like this (right from the start of the project in 
fact, when I wrote our own little bespoke test discovery script).

> 3. I already raised an objection regarding this, in writing and in 
> person. To go ahead with something I don't agree with is to go ahead 
> without me, in other words here to go ahead on your own, or in yet more 
> words inevitably to fork this project. I don't really mind if you want 
> to fork KForge, but it would unfortunate.

Stay calm! :) I'm certainly not trying to 'fork' KForge. I would note 
we've been debating this off and on on the list and in person and, as I 
said, when we last spoke I thought you were happy for me to use 
nosetests as it was 'backwards-compatible'. Clearly this was a 
misunderstanding so we have to debate this some more.

> 4. The test suite/framework/runner isn't broken. So why "fix" it?

For reasons i've already explained :)

<quote>
But why bother to continue to write our own bespoke wrapper of the
unittest runner where nosetests already exists?

...

The thing is I really find nosetests a lot more convenient:

    * No need to write special 'suite' methods
    * Easier ways to run/not run tests from command line
    * Better reporting from succeeding/failing tests (using -d switch)
    * Possibility to use nosetest plugins etc
</quote>

Anyway let's talk more, perhaps in person? :-)

~rufus




More information about the kforge-dev mailing list