If I have a test class like
public class Tests
{
[TestCaseSource(nameof(TestSource))]
public void Test(string c)
{
Assert.Pass();
}
public static IEnumerable<object> TestSource()
{
for (char c = '\0'; c < 255; c++)
{
yield return c.ToString();
}
}
}
then the test runner in the Visual Studio can discover all the tests but doesn't run all of them.
these are the ascii character ranges that are not running:
U_0001 - U_0006
U_000E - U_001F
U_007F - U_0084
U_0086 - U_009F
We currently use some of these characters like U_0002
for network work related test which are included in the test but currently do not run
This is the output for the test run
---------- Starting test discovery for requested test run ----------
NUnit Adapter 3.17.0.0: Test discovery starting
NUnit Adapter 3.17.0.0: Test discovery complete
========== Test discovery finished: 255 Tests found in 4,8 sec ==========
---------- Starting test run ----------
NUnit Adapter 3.17.0.0: Test execution started
Running selected tests in C:\Users\WEIK014\source\repos\TestSourceTestProject\TestSourceTestProject\bin\Debug\netcoreapp2.1\TestSourceTestProject.dll
NUnit3TestExecutor discovered 199 of 199 NUnit test cases
NUnit Adapter 3.17.0.0: Test execution complete
========== Test run finished: 199 Tests run in 8,6 sec (199 Passed, 0 Failed, 0 Skipped) ==========
as you see it doesn't even recognize that it skips tests.
Versions in the project:
Well that sounds like it's probably a bug somewhere in Visual Studio or the test adapter with displaying Unicode control characters. As a workaround, you can override the test display name for control characters:
public static IEnumerable<TestCaseData> TestSource()
{
for (char c = '\0'; c < 255; c++)
{
var data = new TestCaseData(c.ToString());
if (char.IsControl(c))
{
data.SetArgDisplayNames($"{(int)c:X4}");
}
yield return data;
}
}
This will preserve the display for printable characters (e.g. Test("a")
), and the control characters will display as Test(0001)
, etc.