[vtk-developers] Testing guidelines

[Moreland, Kenneth]

I'm one of those that gets annoyed with a bunch of tests named TestFeature because, as Marcus said, it's redundant and is tiresome to manage.  In addition to having to type more at the command line, you get other irritating features like having all the tests grouped together on the dashboard rather than divided in alphabetic sections.

It occurs to me, though, that what Bill and David (and even Marcus) have advocated is that the source code be named TestFeature.cxx, whereas Marcus and I have advocated that the test be named Feature.  Thus, we should be able to have both.  Why not make the convention to name the source code starting with Test and then have some CMake commands to strip off the Test part for the actual name of the test?

-Ken


On 8/21/10 1:54 PM, "Bill Lorensen" <bill.lorensen at gmail.com> wrote:

I still prefer TestClassName.cxx.

On Sat, Aug 21, 2010 at 2:55 PM, David Doria <daviddoria+vtk at gmail.com> wrote:
>>
>> Can I suggest we agree upon and adopt a convention moving forward? We can
>> go back and "correct" existing tests as we have time. It seems reasonable to
>> have a
>>
>> TestClassName.cxx (without the vtk- prefix) for every concrete class.
>
> I have followed this convention in the tests I have added, but in many ways
> the "Test" seems redundant. The big plus is that in my editor of choice (Qt
> Creator) I am able to type "TestS" and have every test starting with S
> listed, but to run a test I have to type,
>
> ./bin/ChartsCxxTests TestScientificPlot -I
>
> (e.g vtkPoints -> TestPoints.cxx)
>
> I doubt the one-to-one mapping of test names to classes will ever be the
> case. Some classes are very amenable to this, others are not. I agree that
> it would be good to establish a standard, and stick to it. Now would be a
> great time to get something set, and do it. I have been working with Bill a
> little to migrate some tests to use the testing interactor machinery,
> finding issues and seeing what we can do there too.
>
> Marcus
>
> (I started a new thread since the old title (about verdict) wouldn't
> probably interest many people).
> Hi all,
> I started a page here:
> http://www.vtk.org/Wiki/TestingGuidelines
> (and linked to it
> from http://www.vtk.org/Wiki/VTK/Developers_Corner#Developers_Corner )
> Please add any thoughts and ideas about testing and coverage to that page. I
> think by having a set of more systematic testing guidelines we will be able
> to keep much more of the code covered. I've also started a list of questions
> (about deprecated classes, coverage target, etc) that should be addressed.
> Thanks,
>
> David
>
> _______________________________________________
> Powered by www.kitware.com
>
> Visit other Kitware open-source projects at
> http://www.kitware.com/opensource/opensource.html
>
> Follow this link to subscribe/unsubscribe:
> http://www.vtk.org/mailman/listinfo/vtk-developers
>
>
>
_______________________________________________
Powered by www.kitware.com

Visit other Kitware open-source projects at http://www.kitware.com/opensource/opensource.html

Follow this link to subscribe/unsubscribe:
http://www.vtk.org/mailman/listinfo/vtk-developers





   ****      Kenneth Moreland
    ***      Sandia National Laboratories
***********
*** *** ***  email: kmorel at sandia.gov
**  ***  **  phone: (505) 844-8919
    ***      web:   http://www.cs.unm.edu/~kmorel

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://public.kitware.com/pipermail/vtk-developers/attachments/20100823/24fec679/attachment.html>