nunit-core team mailing list archive

[Charlie Poole <charlie@xxxxxxxxx>]

Hi Milan,

No, I am talking about the console runner - or even about an independent program
we might write to list tests. The status values are intrinsic to the tests, once loaded,
so any program has to start with a spec that indicates which status values are needed.

Imagine, how can I write a test if I don't know what you want?

I probably confused you by using the word "categories" (with small c) in my last note.
The status we are concerned with is called RunState and it is a property of each test.
The possible values are Runnable, NotRunnable, Explicit, Skipped and Ignored.

You should decide which of these you want. The simplest, of course, is to include
all of them, but that means youir runner, which uses the list, must know how to
deal with all the types.

Considerations for you:

Does your runner recognize [Ignore] [Explicit]?

Can your runner handle non-runnable (invalid) tests or will it crash?

Will your runner handle [Platform], [Culture] and other attributes that
cause skipping a test?

Will your runner always execute on the same machine that is used
to generate the list? Under the same culture? On the same OS?

What do you want done with regard to /other/ options on the command
line?

Charlie

-- 
Feature: Extraction of unit tests from NUnit test asembly and calling appropriate one
https://bugs.launchpad.net/bugs/531873
You received this bug notification because you are a member of NUnit
Developers, which is subscribed to NUnit V2.

Status in NUnit V2 Test Framework: Incomplete

Bug description:
Hi all, 

in order to integrate NUnit better in the existing solution it would be good that 
by using NUnit(GUI or command line version) over a given Test.exe (where all 
necessery test tags/unit tests are allready compiled) to retreive list 
of all Unit test names together with its argumens.
 
In particular case one have Test.exe which is written in order to test some 
library Library.dll and now by using command line one call 
NUNit over the Test.exe and to get back list of all Unit tests which are 
ready to be tested. Also one could as command line argument submit which 
test class or which cathegory is imjortant so that list of the returned tests is filtered.

Once the list of all Unit test cases is present (it can be displayed in some other 
industry integrated tool) tester can decide what test case will be executed by calling 
again using command line specific Unit test. It would be also very good that every test 
has its testresult file, having everything in one file makes problem in accessing only 
important test and it is not in accordance with other HW/SW test tool where every test
has its testresult file.

Br, 
Milan.