You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
While I was working on #552, I realized skipping the Qt-specific test case using the skipIf(toolkit_object.toolkit != "qt4", ...) was not enough, I also had to add a skipIf(GuiTestAssistant.__name__ == "Unimplemented", ...).
This suggests that tests that require Qt, e.g. because the test body contain Qt specific code, are skipped relying on the fact that GuiTestAssistant is only implemented for Qt. This is not ideal, because as soon as GuiTestAssistant is implemented for other toolkit (e.g. wx, see #266), these tests will be run erroneously, causing build errors.
The text was updated successfully, but these errors were encountered:
That was my stupidity: In the case of when the GuiTestAssistant is not implemented, GuiTestAssistant.__init__ is overridden to raise NotImplemenetedError. The test case needs to place GuiTestAssistant after unittest.TestCase in the MRO so that the test loader can still instantiate the test case before it is skipped. Nothing is wrong here. Sorry for the noise.
While I was working on #552, I realized skipping the Qt-specific test case using the
skipIf(toolkit_object.toolkit != "qt4", ...)
was not enough, I also had to add askipIf(GuiTestAssistant.__name__ == "Unimplemented", ...)
.This suggests that tests that require Qt, e.g. because the test body contain Qt specific code, are skipped relying on the fact that
GuiTestAssistant
is only implemented for Qt. This is not ideal, because as soon asGuiTestAssistant
is implemented for other toolkit (e.g. wx, see #266), these tests will be run erroneously, causing build errors.The text was updated successfully, but these errors were encountered: