[Cuis-dev] Tests

[Luciano Notarfrancesco]

Hi Hernan, Juan,
Well.. I wouldn’t worry about tests that are not well written, they should
be fixed anyway. But I expected to be able to run a TestCase multiple times
with the same result. I’m using the changes that I posted since last year,
otherwise I couldn’t have repeatability in my tests. I cannot run tests
with cmd-t because the tests are not unit tests for single clases, they
have more “context” than a single class. For example I cannot just test the
class Matrix, I have to test matrices with coefficients over a certain ring
or field, so I have generic tests that have an instance variable for the
coefficients ring, and I have to set things like that when I build a test
suite.

Another way to achieve what I need is to copy the test case instead of
creating a new one. Just need to implement

TestCase>>copy
    ^ self class selector: testSelector

And in my subclasses I can make sure to copy all the additional state. Also
I’d have to implement #isSameAs: in my subclasses (why not call it #= ?).

Cheers,
Luciano

On Tue, 16 Aug 2022 at 20:44 Hernan Wilkinson <hernan.wilkinson at 10pines.com>
wrote:

> Hi Luciano, Juan,
>  using the same test instance to debug breaks the "test isolation"
> rule, can bring problems if the test is not well written, etc.
>  I think that if you have test data that you want to use to debug a test
> you could put that data in class variables or something like that, not in
> instance variables. Would that work for you? Or you can apply the change
> you suggest for the test cases that you want that behavior, I mean redefine
> #debug as you suggested in your .cs for the test case you want that
> behavior.
>  I think it is not a good idea to use that behavior for all test cases.
>
>  Regarding the test runner, I usually run the tests with cmd+t not the
> test runner, but it makes sense what you suggest to open the debugger in
> that case.
>
> Cheers!
> Hernan.
>
> On Tue, Aug 16, 2022 at 10:16 AM Juan Vuletich <juan at cuis.st> wrote:
>
>> Hi Luciano,
>>
>> On 8/15/2022 6:48 AM, Luciano Notarfrancesco via Cuis-dev wrote:
>> > I have subclasses of TestCase that hold more state besides the test
>> > selector. For example I have a subclass with a 'seed' instance
>> > variable that is set randomly in >>#initialize, and a 'random'
>> > instance variable that is set to 'Random seed: seed' in >>#setUp. This
>> > allows me to write tests that use randomness and are reproducible. But
>> > when debugging a test that failed, a new TestCase instance is created
>> > and the new TestCase is debugged instead, so the original seed is lost
>> > and my tests are not reproducible. I'm also experimenting with generic
>> > tests, for example I have a generic VectorSpaceTest with an instance
>> > variable 'scalars' for the coefficients field, so I can create a test
>> > suit with tests for vector spaces over the rationals, Z/pZ, rational
>> > functions, etc (vector spaces over different 'scalars'), but again the
>> > scalars are lost when I try to debug a test. I don't see any reason to
>> > create new instances of the test cases when debugging, see the
>> > attached change set and let me know if you agree with it.
>>
>> I'd defer to Hernán and others who may have an opinion. I don't know if
>> reusing the test instance could bring problems. Besides, any test I
>> write is repeatable, setting Random seeds as needed. I guess you must
>> have your reasons to do otherwise...
>>
>> > Also, when a test was taking too long I used to click the stop button
>> > in the test runner and a debugger would open, so I what it was doing.
>> > But now the test runner just stops without opening a debugger.
>>
>> Maybe you were doing cmd-. to interrupt the process? I don't remember
>> [stop] opening a debugger. It is easy to do, though:
>>
>> terminateRun
>>      running debugWithTitle: 'Interrupted Test'
>>
>> If people agree, I can integrate this change.
>>
>> >
>> > Thanks,
>> > Luciano
>>
>> Cheers,
>>
>> --
>> Juan Vuletich
>> cuis.st
>> github.com/jvuletich
>> researchgate.net/profile/Juan-Vuletich
>> independent.academia.edu/JuanVuletich
>> patents.justia.com/inventor/juan-manuel-vuletich
>> linkedin.com/in/juan-vuletich-75611b3
>> twitter.com/JuanVuletich
>>
>>
>
> --
> <https://10pines.com/>Hernán WilkinsonSoftware Developer & Coach
>
> Alem 896, Floor 6, Buenos Aires, Argentina
>
> +54 11 6091 3125
>
> @HernanWilkinson
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.cuis.st/mailman/archives/cuis-dev/attachments/20220816/f37b4f57/attachment-0001.htm>