Try running with latest PCT

[jglick]

Fixes #657. Going to see what, if anything, jenkinsci/plugin-compat-tester#276 was actually needed for. Also wondering if jenkinsci/plugin-compat-tester#299 or other recent changes will break things.

[jglick]

Something weird in https://ci.jenkins.io/job/Tools/job/bom/job/PR-659/1/pipeline-console/?selected-node=263 and something else in jquery3-api; I suspect there is some sort of regression affecting plugins with no src/test/java/ (just InjectedTest).

https://github.com/jenkinsci/bom/pull/659/checks?check_run_id=3733673007 seems to reproduce the need for jenkinsci/plugin-compat-tester#276:

java.lang.NoSuchMethodError: io.jenkins.plugins.casc.ConfigurationAsCode.toYaml(Lio/jenkins/plugins/casc/model/CNode;)Lio/jenkins/plugins/casc/snakeyaml/nodes/Node;
	at io.jenkins.plugins.casc.misc.Util.toYamlString(Util.java:152)
	at hudson.plugins.view.dashboard.ConfigurationAsCodeBackCompatTest.testOldImagePortletUrl(ConfigurationAsCodeBackCompatTest.java:63)

[jetersen]

if dashboard is the only plugin with this issue why not just fix that plugin instead of creating an hack?

[jglick]

TBD if only dashboard is affected. If the problem is just an old JCasC dep, then fine. But AFAICT it is using something newish.

[jglick]

Hmm, failFast not working any more? (In this case, good.)

[jglick]

850f995 might have broken failFast, TBD.

[jglick]

Well, maybe it does still work, just more slowly.

[timja]

if dashboard is the only plugin with this issue why not just fix that plugin instead of creating an hack?

from what i remember it's not a hack as such, the configuration as code plugin and test harness are not correctly update by PCT, only the plugin is updated and not the test harness.

It creates a nightmare when they aren't updated together.

not such an issue now while there's not many changes but still has been a huge pain before

[jglick]

Perplexed by why https://ci.jenkins.io/job/Tools/job/bom/job/PR-659/2/testReport/hudson.plugins.view.dashboard/ConfigurationAsCodeTest/ would be passing against the latest line yet failing against the oldest line. The Jenkins version should not make any difference, and we are using the same configuration-as-code version in all lines. (And same for snakeyaml-api.)

[jglick]

There is also a PCT regression affecting workflow-scm-step (log):

 [ERROR]     'dependencies.dependency.version' for org.jenkins-ci.plugins.workflow:workflow-step-api:jar is missing. @ line 66, column 17

[jglick]

In sum, there are at least three PCT bugs visible in this project:

• configuration-as-code test harness issue, as already suspected
• library wrapper plugins broken
• something involving workflow-scm-step

[imonteroperez]

In sum, there are at least three PCT bugs visible in this project:

• configuration-as-code test harness issue, as already suspected
• library wrapper plugins broken
• something involving workflow-scm-step

Regarding workflow-scm-step I can confirm it was failing with the same issue before jenkinsci/plugin-compat-tester#299 gets merged

[imonteroperez]

Suggested change

	version=0.5.1-rc1106.47640a4e521e
	version=0.5.1

https://github.com/jenkinsci/plugin-compat-tester/releases/tag/plugins-compat-tester-aggregator-0.5.1

[jglick]

We still need jenkinsci/plugin-compat-tester#276, or something like it.

[jglick]

jenkinsci/plugin-compat-tester#320 ought to speed things up a bit.

[jglick]

#1336