[vtk-developers] TestStackedPlot failures

[Bill Lorensen]

Marcus,

I agree. This was a tough one to catch. Not a typical situation.

It will be difficult to replace a human being. Job security. But I don't
need a job.

That said, at least we can say that since the gerrit, continuous and
nightly builds are so clean we caught this quickly. A month ago, this would
have just accumulated as another failure.

Bill

On Thu, Mar 7, 2013 at 3:43 PM, Marcus D. Hanwell <
marcus.hanwell at kitware.com> wrote:

> On Thu, Mar 7, 2013 at 1:51 PM, Marcus D. Hanwell
> <marcus.hanwell at kitware.com> wrote:
> > On Thu, Mar 7, 2013 at 8:10 AM, Bill Lorensen <bill.lorensen at gmail.com>
> wrote:
> >> Marcus,
> >>
> >> The Cxx test was changed but the Python test was not.
> >>
> > Yeah, sorry I missed that (and I guess we didn't spot it in the
> > review). I approved and merged your topic earlier today.
> >
> For posterity...we missed this because the Python test uses the same
> baseline image as the C++ test. When I pushed the topic the C++ test
> failed as expected (and we could see the new baseline), but the Python
> test continued to pass (as it matched the old baseline). I pushed in
> the new baseline, Chris merged the topic and then the Python test we
> hadn't spotted began to fail. Bill changed its code to match the C++
> test and all is well.
>
> I don't think this will happen much, and despite the timing I really
> didn't do it on purpose to demonstrate a point ;-) Merging
> master/rebasing etc wouldn't have helped here.
>
> Marcus
>



-- 
Unpaid intern in BillsBasement at noware dot com
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://public.kitware.com/pipermail/vtk-developers/attachments/20130307/232d7f5c/attachment.html>