[Pharo-project] SUnit design question
I'm not sure if I asked that question before, but why is there no dedicated TestRunner model in SUnit? (detailed question further down) There is a TestRunner but that is clearly a UI part. We have right now 2 places in the system where that would come in handy: HDTestReport and the CommandLineTestRunner. Both places modify the output of the test runner, so to say instrument the test runs. see http://stackoverflow.com/questions/16315726/how-do-i-instrument-test-runners...
Indeed it would be nice to have. On May 1, 2013, at 11:29 AM, Camillo Bruni <camillobruni@gmail.com> wrote:
I'm not sure if I asked that question before, but why is there no dedicated TestRunner model in SUnit? (detailed question further down)
There is a TestRunner but that is clearly a UI part.
We have right now 2 places in the system where that would come in handy: HDTestReport and the CommandLineTestRunner.
Both places modify the output of the test runner, so to say instrument the test runs.
see http://stackoverflow.com/questions/16315726/how-do-i-instrument-test-runners...
In Amber I created such thing, for the same reasons, but there SUnit is lot simpler than Pharo's one. stephane ducasse wrote:
Indeed it would be nice to have.
On May 1, 2013, at 11:29 AM, Camillo Bruni<camillobruni@gmail.com> wrote:
I'm not sure if I asked that question before, but why is there no dedicated TestRunner model in SUnit? (detailed question further down)
There is a TestRunner but that is clearly a UI part.
We have right now 2 places in the system where that would come in handy: HDTestReport and the CommandLineTestRunner.
Both places modify the output of the test runner, so to say instrument the test runs.
see http://stackoverflow.com/questions/16315726/how-do-i-instrument-test-runners...
Yeah, so maybe this is the way to go, for now I think we simply add a bit more extended MOP to the TestResult so we can rewrite it from there on. For instance we only have #addSkip: which does not give you the user-defined message from #sip: as an argument. Same goes for #addError: and #addFailure: On 2013-05-02, at 10:08, Herby VojÄÃk <herby@mailbox.sk> wrote:
In Amber I created such thing, for the same reasons, but there SUnit is lot simpler than Pharo's one.
stephane ducasse wrote:
Indeed it would be nice to have.
On May 1, 2013, at 11:29 AM, Camillo Bruni<camillobruni@gmail.com> wrote:
I'm not sure if I asked that question before, but why is there no dedicated TestRunner model in SUnit? (detailed question further down)
There is a TestRunner but that is clearly a UI part.
We have right now 2 places in the system where that would come in handy: HDTestReport and the CommandLineTestRunner.
Both places modify the output of the test runner, so to say instrument the test runs.
see http://stackoverflow.com/questions/16315726/how-do-i-instrument-test-runners...
participants (3)
-
Camillo Bruni -
Herby VojÄÃk -
stephane ducasse