Ignore failing tests#39
Ignore failing tests#39pzygielo wants to merge 1 commit intojavaee-samples:masterfrom pzygielo:ignore-failing-tests
Conversation
|
@pzygielo In this specific project, tests are actually allowed to fail, as they indicate the project which they test against is faulty and the vendor should fix the issue. The test is not broken, but for that test Payara is. It agree that is does make it difficult to see if new tests fail or not. We've all been trained (me too) to strive for passing builds, but this project is thus a little different. |
|
I think it would be better to have such test in vendor(s) suite, and not in samples. From samples I'd expect to pass rather than show faults in used components (regardless of ci-job disclaimer). With failing test, one always has to check the reason for failure. And having two failing just on start (that are not announced to be known to fail by default in any way in code), although somehow educational, isn't quite nice experience to begin with.
|
I propose to convert failing travis/default build of master, to new issue and
@Ignorefailing tests.Tests were broken in #34, with explanation provided there. IMO new issue that would remind to remove this exclusion once [not sure about condition] is met, would be better than failing build.
(Or perhaps warning for skipped tests is better than new issue, I do not know.)
Although #34 declares
I checked that for
-P glassfish-embedded,!payara-ci-managed(i.e. for glassfish) the same happens.