When function under test calls some other function, which we don't want to be called, we can specify a stub. testIDEA supports two types of stubs - normal stubs and user stubs. This section describes normal stubs, while user stubs are described in the next section.
When we define a normal stub, the stubbed function will not be called, but target will be stopped and then output parameters and return value will be set according to test specification.
When script extension function is specified, assignments given in testIDEA are executed before the script function is called.
The detailed description of test specification entries is given in tool-tips.
Return value assignment
Because of calling conventions compilers often use the same register for input parameter and return value. Therefore we should first assign values to parameters, and only then to assign a return value in the Assignments table. testIDEA will issue a warning if this is not the case. If we assign values in Python scripts, the same rule applies, but this time testIDEA can not detect such faults.
Using hardware breakpoints
When the tested code is located in flash memory, setting of software breakpoints may be very slow. One possible optimization is setting of option Hardware | CPU Options ... | Debugging | Set/Clear SW BP before Run in winIDEA. If set, all software breakpoints are not written into flash until before the target is started. Usually this means one flash write, instead of several.
To make testing even faster, hardware breakpoints should be used. Unfortunately the number of these breakpoints is limited, so if there are more stubs and test points then there are HW breakpoints available, we can not use them. However, if we know the order in which stubs or test points will be hit, we can use custom breakpoint activation with setting Is custom act in sections stubs and test points. If stub or test point has this setting set to Yes, then testIDEA will not set breakpoint for this stub or test point. Instead, we have to write a script function, which removes existing breakpoints and sets next ones depending on target state.
For details about writing scripts please see also section Writing script extensions.
Suppose we have only one HW breakpoint available, and want to test a function with three stubs and two test-points:
int numItems = 7;
numItems += stubbedFuncInt();
numItems *= 2; // tp_tcsa_1
numItems *= 2; // tp_tcsa_2
numItems += (int)stubbedFuncFloat();
Stub for function stubbedFuncInt() should be activated by testIDEA, then next stubs and test points will be activated by script function testCustomBPActivation()
Script parameters above are recommended - the first one is test case specification needed to get stub or test point data, the next one indicates to the script which stub or test point has been hit. Instead of this we could write new script function for each stub, or the script function could evaluate target variable to decide which stub/test point to activate next.
def testCustomBPActivation(self, testSpec, stubOrTpId):
print('testCustomBPActivation(), Test case ID: ', testSpec.getTestId(),
' stubId:', stubOrTpId)
testCtrl = self.__getTestCaseCtrl()
testCase = ic.CTestCase(self.connectionMgr)
if stubOrTpId == 'Int':
testCase.setEnableStub(testSpec, 'stubbedFuncInt', False)
testCase.setEnableTestPoint(testSpec, 'tp_tcsa_1', True)
elif stubOrTpId == 'tp_1':
testCase.setEnableTestPoint(testSpec, 'tp_tcsa_1', False)
testCase.setEnableStub(testSpec, 'stubbedFuncIntPtr', True)
elif stubOrTpId == 'IntPtr':
testCase.setEnableStub(testSpec, 'stubbedFuncIntPtr', False)
testCase.setEnableTestPoint(testSpec, 'tp_tcsa_2', True)
elif stubOrTpId == 'tp_2':
testCase.setEnableTestPoint(testSpec, 'tp_tcsa_2', False)
testCase.setEnableStub(testSpec, 'stubbedFuncFloat', True)
elif stubOrTpId == 'Float':
# in this case it is the last stub - it will be deleted by testIDEA
raise Exception('Unknown stub or test-point Id: ' + stubOrTpId)
If we use stubs only (no test points), then instead of the second script parameter we can get the currently hit stub name by the following code snippet:
def testCustomBPActivation(self, testSpec):
testCase = ic.CTestCase(self.connectionMgr)
currentAddress = self.debug.getCPUStatus().getExecutionPoint()
addressCtrl = self.debug.getAddressController()
stubName = addressCtrl.getSymbolAtAddress(ic.IConnectDebug.sFunctions,
if stubName == 'readADC':
testCase.setEnableStub(testSpec, 'readADC', False)
testCase.setEnableStub(testSpec, 'writeADC', True)
Note: One breakpoint is used by testIDEA for run control. For example, on system with 4 hardware breakpoints, at most 3 stubs and test points may be active at one moment if hardware breakpoints are used. See Run | Configuration ... dialog for selection of breakpoint type.
Since stubbed functions do not have the same effect for each call during normal run, it is also possible to specify assignments for each hit of the stub. For each stub hit we can specify the following:
Indices of steps are zero based and are shown in the leftmost column in the assignments table.