-
Bug
-
Resolution: Won't Fix
-
Minor
-
None
There is "Run Always" checkbox under "Advanced" for Warnings plugin. It allows count error/warnings in case when build fails. However, the trend graph does not display graph for failed builds except last one. On my trend graph I got builds 1-22 and then 39. Builds 23-39 were failed builds but 23-38 did not make it to the graph. I can see that errors were in fact counted when inspecting individual builds.
- is related to
-
JENKINS-10682 Option for Trendgraph to ignore failed builds
-
- Resolved
-
[JENKINS-14867] Warnings plugin Trends graph does not include failed builds except for the last one
Andrew Gvozdev
added a comment - In reply to comment #1:
> Can you please clarify which trend graph is affected? On the project page, on
> the dashboard? Are all graphs affected? Or only the priority graph?
The project page. I have only priority graphs (for different parsers). On "Configure the trend graphs" all graphs except "Difference between new and fixed" exhibit this behavior.
We are migrating our codebase to a new compiler and with that having many errors. We are tracking our progress with Jenkins and this is really unfortunate that graphs are useless in this scenario. I could force the build to return 0 even if it fails but that makes it to show that build passes, which is misleading and it makes it hard to tell which builds actually fail and which pass.
Andrew Gvozdev
added a comment - In reply to comment #1:
> Can you please clarify which trend graph is affected? On the project page, on
> the dashboard? Are all graphs affected? Or only the priority graph?
The project page. I have only priority graphs (for different parsers). On "Configure the trend graphs" all graphs except "Difference between new and fixed" exhibit this behavior.
We are migrating our codebase to a new compiler and with that having many errors. We are tracking our progress with Jenkins and this is really unfortunate that graphs are useless in this scenario. I could force the build to return 0 even if it fails but that makes it to show that build passes, which is misleading and it makes it hard to tell which builds actually fail and which pass. Ok, thanks. I found the affected code piece. It was introduced due to a feature request in order to remove build results from the graph that are wrong because of build failures. I'm not sure if it safe to change that part of the code without breaking something else. Maye the code needs to be duplicated for graph display and results aggregation.
Please reopen if you have a solution that makes both behaviors possible.
Can you please clarify which trend graph is affected? On the project page, on the dashboard? Are all graphs affected? Or only the priority graph?