Hi Tim --
This sounds like a bug / incomplete design to us (and probably an easy one
to fix). Would you be open to filing a GitHub issue against it?
-Brad
On Fri, 7 Jun 2019, Stavenger, Timothy J via Chapel-developers wrote:
I'm working towards having our Travis-CI build run 'mason test' instead
of 'start_test'. Ultimately the intent is to have our code a package in
mason, the hope here is that running these tests will verify the same
thing mason would be doing. If this isn't a good approach, perhaps my
question below is misguided - is there a better or more correct way?
When running 'mason test' the exit code when there are failures is 0,
which wouldn't make an automated build fail. Is this expected behavior?
--- Summary: 23 tests run ---
-----> 15 Passed
-----> 8 Failed
$ echo $?
0
_______________________________________________
Chapel-developers mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/chapel-developers
