Details
-
Bug
-
Status: Closed (View Workflow)
-
Minor
-
Resolution: Fixed
-
None
-
-
pacific, 1.0-b05/b-06
Description
Steps to reproduce
Try this example Pipeline
node {
try {
sh 'exit 1'
} catch (Throwable e) {
sh 'echo "Eat all exceptions"'
} finally {
sh 'sleep 1000'
sh 'sleep 1000'
sh 'sleep 5000'
sh 'sleep 1000'
sh 'sleep 1000'
sh 'sleep 5000'
sh 'sleep 1000'
sh 'sleep 1000'
sh 'sleep 5000'
sh 'echo "Succeeds"'
}
}
Run the Pipeline. When executing this pipeline looks like:
It is not possible to close the failing shell step - it opens again after a second or two
What it should look like
When this condition is hit and the run is executing:
- The proceeding failing step should be closed
- Running step should be open and tailing output
- If the user opens or closes the failing step the UI should never change the expand state of the step.
Attachments
Activity
this is really weird, I don't think that sequence of "states" was anticipated. It looks bad that there is a "red failure" as part of normal operation, but certainly yes this isn't correct behavior of expanding at least.
Please have a look on the video I took.
The current situation is a bit different however we have a "rule" that failing steps are expanded, if you want I can limit that to not-running jobs, but the rest I think is working fine now.
I implemented now that in running mode the failure nodes are not expanded only the currently running step.
This will be a common state to get into when people start using Kyoto and Blue Ocean. The examples in Kyoto use Docker.
The first step is inspecting a local container called "java". If this command fails (it will the first time because it doesn't exist locally) then it will fetch it from the docker repository.