Replies: 4 comments 7 replies
-
Usually, anything that's not 0 should be considered a faulty execution. The majority of tools works this way. @lahma or @chaquotay uses/used Bitbucket, do you have any experience here? |
Beta Was this translation helpful? Give feedback.
-
We currently use Bamboo (also from Atlassian, but as far as I know not related to Bitbucket), so I don't have any Bitbucket experience 😕 |
Beta Was this translation helpful? Give feedback.
-
We're running on bitbucket pipelines with nuke and failed tests error the pipeline. Maybe need to look into it more. Can those parallel tests affect final exit code? |
Beta Was this translation helpful? Give feedback.
-
Hey @lahma, thanks for the input. Our script call is like:
Are there any differences to yours? Only the first step (Android build) is executing UnitTests. Here the additional log where I can see that NUKE build fails: Thanks! |
Beta Was this translation helpful? Give feedback.
-
Usage Information
Nuke 6.3.0 / SDK 7.0.100 / Framework net7.0 / OS Windows 10
Description
I observed a wrong behavior of the Bitbucket Pipelines Script Runner - I'm not sure if this is a bug:
Running tests in NUKE with DotNetTest won't indicate a failing build, if one of the unit tests is failing:
The reason is the returned exit code which is negative, but it should be a positive value to indicate BB a failed build. A workaround is currently adapting the build.ps1 script as follows:
instead of the default logic:
I'm not sure if BB causes the problem or if it's possible to solve this on NUKE side.
Adapting each bootstrapper script file on creating a new build project could lead to missing this step.
Reproduction Steps
Expected Behavior
Build should fail if unit test is failing:
Actual Behavior
Regression?
No response
Known Workarounds
Could you help with a pull-request?
No
Beta Was this translation helpful? Give feedback.
All reactions