Issue #1664: Unit Test support for mapserver

[Tamas Szekeres]

For C# unit tests could be utilized by tools like nunit

http://www.nunit.org/

that could be made integrated with the build process as well.

However I think many of the tests should be unified amongst theese languages.

Best Regards,

Tamas


On Tue, 14 Feb 2006 12:51:15 -0700, Sean Gillies <sgillies at FRII.COM> wrote:

>On Feb 14, 2006, at 9:23 AM, Umberto Nicoletti wrote:
>
>> I have opened issue #1664 to track comments, feedback and of course
>> criticisms on the much needed (IMHO) addition of unit tests to
>> mapserver.
>>
>> Unit tests are a completely automated way of testing code, which
>> should improve the development process, the code quality and the
>> developers' confidence in making changes to the code. I now use them
>> in almost every project I am involved and I find them invaluable. Go
>> and check out some unit tests advocacy here:
>>
>> http://www.junit.org
>>
>> For unit testing mapserver I have decided to use Cunit (version 2.0.x
>> from http://cunit.sf.net). Unit tests are located under the cunit
>> directory and must be compiled in pretty much the same way a mapscript
>> extension would. Build and running instructions are detailed in
>> bugzilla.
>> Please note that you need to install CUnit on your computer before you
>> can use it. I am currently using 2.0 on gentoo x86.
>>
>> The unit tests I have implemented so far only test the connectin
>> pooling functions and to activate them you need to configure using the
>> --with-threads argument.
>>
>> Last but not least I am working on a continuos integration project (a
>> la Cruise Control).
>>
>> Let me know what you think and let's get this into cvs HEAD!
>>
>> Umberto
>
>I had not yet heard of cunit. Seems pretty cool. IMO, this could be
>generally better, and more direct, than the unit tests under
>mapscript/python. We've all got C compilers, so no one can complain
>about how they don't have Python. On the other hand, it doesn't
>exercise the C#, Java, or Python bindings at all, so it isn't a
>complete replacement for the Python test suite (or a similar Java
>suite).
>
>Like Frank suggests, write up an RFC. I'm in support of the idea, and
>I'm sure others can also be convinced of its utility.
>
>cheers,
>Sean
>
>---
>Sean Gillies
>sgillies at frii dot com
>http://zcologia.com/news