Qute suffix issue on hot reload

[gsmet]

@mkouba from time to time, when the Quarkus GitHub Bot is reloading with live reload, I end up having the following error:

Caused by: io.quarkus.qute.TemplateException: [WorkflowReportFormatter/checkRunReport.md] match the path WorkflowReportFormatter/checkRunReport but the file suffix is not configured via the quarkus.qute.suffixes property
    at io.quarkus.qute.deployment.QuteProcessor.collectCheckedTemplates(QuteProcessor.java:383)
    at java.base/jdk.internal.reflect.NativeMethodAccessorImpl.invoke0(Native Method)
    at java.base/jdk.internal.reflect.NativeMethodAccessorImpl.invoke(NativeMethodAccessorImpl.java:77)
    at java.base/jdk.internal.reflect.DelegatingMethodAccessorImpl.invoke(DelegatingMethodAccessorImpl.java:43)
    at java.base/java.lang.reflect.Method.invoke(Method.java:568)
    at io.quarkus.deployment.ExtensionLoader$3.execute(ExtensionLoader.java:849)
    at io.quarkus.builder.BuildContext.run(BuildContext.java:256)
    at org.jboss.threads.ContextHandler$1.runWith(ContextHandler.java:18)
    at org.jboss.threads.EnhancedQueueExecutor$Task.run(EnhancedQueueExecutor.java:2513)
    at org.jboss.threads.EnhancedQueueExecutor$ThreadBody.run(EnhancedQueueExecutor.java:1538)
    at java.base/java.lang.Thread.run(Thread.java:840)
    at org.jboss.threads.JBossThread.run(JBossThread.java:501)

I have:

quarkus.qute.suffixes=md
quarkus.qute.content-types."md"=text/markdown

in the application.properties.

I wonder if we should make .md part of the default suffixes but still this is weird.

It is not reproducible always but when it does happen, you can't recover from it by pushing a new request.

[quarkus-bot[bot]]

/cc @mkouba (qute)

[mkouba]

Hm, this is weird. We could enhance the error message to also contain the TemplateFilePathsBuildItem.filePaths. But to be honest I have no idea what's going on...

[gsmet]

@mkouba yeah it's as if the config was somehow lost after hot reloading.

[mkouba]

@gsmet Is this still a problem?

It is not reproducible always but when it does happen, you can't recover from it by pushing a new request.

Are there any steps to reproduce?