Closed UnixJunkie closed 11 years ago
Good idea, should be doable easily (more or less depending on how easy oUnit makes it to get stats; don't remember).
(general note: I won't be implementing anything for a while, though -- thesis to finish and all that -- July at the earliest)
On 05/14/2013 11:55 AM, Vincent Hugot wrote:
Good idea, should be doable easily (more or less depending on how easy oUnit makes it to get stats; don't remember).
(general note: I won't be implementing anything for a while, though -- thesis to finish and all that -- July at the earliest)
No problem.
If you label the issue and give some guidelines, it gives a chance to someone else to implement the feature.
There are quite some hackers who already commited things in qtest I see.
I am also quite busy these days, but may contribute to qtest in the near future as it looks so cool and useful.
Best regards, Francois.
Guideline:
probably as simple as replacing the ignore in qtest.mll/generate
out "let () = ignore (Runner.run (\"\" >::: List.rev !___tests))\n";
I cannot test but I sent a patch for this.
Please tell me how it goes for you
This can be close now
Closed by #11.
(forgot to do that in my merge comment...)
Hello,
I think any test failing should give a $? to 1 on unix.
Then, we could do things like (on Bash): ./test_suite && make install
Which means only install if the tests were OK.
Regards, F.