-
Bug
-
Resolution: Unresolved
-
Minor
-
None
-
Windows
Jenkins ver. 1.651.2
conditional-buildstep 1.3.3
"Conditional build step-single" shows nice field:
"On evaluation failure" with options "Fail the build" or mark build as unstable.
Unfortunatly this is not working.
Example 1
Run? = Never
On evaluation failure = "Mark the build unstable"
Builder = "Execute windows batch command"
Command = "echo run builder"
It should mark all builds as unstable and not run the builder.
It does not run builder but does not mark build as unstable.
Example 2 - more reliable
Run? = Execute Windows batch Command
Command = which notinpath
On evaluation failure = "Mark the build unstable"
Builder = "Execute Windows batch Command"
Command = echo builder
notinpath is not in my path, so evaluation should fail - it does.
Build should be marked as unstable - it is not.
Echo builder should not be runned - it is not.
It Shouls mark all builds as unstable - but it does not.
- blocks
-
-
- Resolved
-
- duplicates
-
-
- Resolved
-
- is related to
-
-
- Closed
-
[JENKINS-34971] Conditional build step plugin does not mark build a s unstable
Roman Wieczorek
added a comment - Currently docs says:
"If the evaluation of a run condition fails, should the build fail, be marked unstable, run the build step ...
A run condition evaluation may fail to run cleanly - especially if it is dependent on expanding tokens provided by the Token Macro Plugin and the values are expected to be present or look like a certain type i.e. be a number."
Maybe it could be:
"Action to take when can not evaluate condition - this is not same as evaluating to false.
Should the builder run? Build marked as failed or unstable?
A run condition evaluation may fail to run cleanly - especially if it is dependent on expanding tokens provided by the Token Macro Plugin and the values are expected to be present or look like a certain type i.e. be a number."
"
Roman Wieczorek
added a comment - Currently docs says:
"If the evaluation of a run condition fails, should the build fail, be marked unstable, run the build step ...
A run condition evaluation may fail to run cleanly - especially if it is dependent on expanding tokens provided by the Token Macro Plugin and the values are expected to be present or look like a certain type i.e. be a number."
Maybe it could be:
"Action to take when can not evaluate condition - this is not same as evaluating to false.
Should the builder run? Build marked as failed or unstable?
A run condition evaluation may fail to run cleanly - especially if it is dependent on expanding tokens provided by the Token Macro Plugin and the values are expected to be present or look like a certain type i.e. be a number."
" Experiencing the same issue (see attached screenshot and logs) where the build is not marked failed even if the condition fails.
10:14:23 [test] $ /bin/bash /apps/instances/jenkins/temp/CommandInterpreterCondition4832653641132345461.sh 10:14:23 ERROR: Please remove '.en' at the end of SOLUTION name platform.en 10:14:23 Run condition [Not] preventing perform for step [Copy Job] 10:14:23 Started calculate disk usage of build 10:14:23 Finished Calculation of disk usage of build in 0 seconds 10:14:23 Started calculate disk usage of workspace 10:14:23 Finished Calculation of disk usage of workspace in 0 seconds 10:14:23 Finished: SUCCESS
Ajay Kumar
added a comment - - edited Experiencing the same issue (see attached screenshot and logs) where the build is not marked failed even if the condition fails.
10:14:23 [test] $ /bin/bash /apps/instances/jenkins/temp/CommandInterpreterCondition4832653641132345461.sh
10:14:23 ERROR: Please remove '.en' at the end of SOLUTION name platform.en
10:14:23 Run condition [Not] preventing perform for step [Copy Job]
10:14:23 Started calculate disk usage of build
10:14:23 Finished Calculation of disk usage of build in 0 seconds
10:14:23 Started calculate disk usage of workspace
10:14:23 Finished Calculation of disk usage of workspace in 0 seconds
10:14:23 Finished: SUCCESS
I linked in JENKINS-24242 because it has a suggested workaround that might help until this issue is resolved.
Even with the workaround, I vote for this being fixed. The workaround is a bit of a hack.
I think documentation is missleading. What does mean that evaluation failures?
Does not it return a bolean value allways?