More failing tests...

[William Reade]

On Thu, 2013-02-28 at 17:26 +1300, Tim Penhey wrote:
> Can anyone help with these?
> 
> https://bugs.launchpad.net/juju-core/+bug/1134959
> 
> I'm getting them failing most of the time, and I have no idea why.

It's r943, I'm afraid; it seems that dimiter did not run the full test
suite before submitting. However, we're now up to r947, and 3 more
people (ian, tim, gustavo) have apparently been blithely checking in
further code without running the full test suite.

This is probably my fault for not making my expectations clear, so let's
be explicit about this:

 1) Do *not* merge into trunk without a full passing test suite.
    Frankly, you shouldn't even lbox propose without the justified
    belief that you haven't broken anything, and the only way to get
    that is to RUN THE TESTS.

 2) If trunk is broken, the first person to discover it is responsible
    for reverting it to the latest non-broken revision *before* doing
    any further work; coordinate in IRC so only one person does it,
    mail juju-dev to ensure that the perpetrator sees what's happened,
    and then carry on from the non-broken state.

On the other hand, I'm disappointed that I need to say this. You've all
worked with other developers before, so you should know that by not
taking the 10 minutes to run the tests, you're imposing a cost of at
*least* 10 minutes on *everyone* else who gets your potentially-broken
code and needs to figure out what happened; this is inefficient, and
more than a little rude, so please don't do it any more.

I'm reverting to r942 as we speak, and dimiter's looking into the
original problems. Tim/ian/gustavo, it's your responsibility to get your
work landed again: feel free to do this without further supervision, so
long as:

 1) Your original diff, as reviewed and accepted, applies cleanly.

 2) The full test suite passes on your machine.

Does anyone have anything to add here?

Cheers
William