Bauke Scholtz

It's indeed a pita but it's what it is. It has been reverted from 4.1 now.

This should be reported at https://github.com/eclipse-ee4j/mojarra/issues

What's the stack trace of the `@PostConstruct` call? Have you been able to boil down a reproducer?

I wanted to check if the bean is initializing itself or via another bean in a loop which would then suggest a cyclic dependency and this seems indeed to be...

Ah there is finally the root cause I was looking for: ``` Caused by: java.lang.IllegalStateException: Cannot create a session after the response has been committed at org.apache.catalina.connector.Request.doGetSession(Request.java:3088) at org.apache.catalina.connector.Request.getSession(Request.java:2508) at...

Summarized: an ideal error page shouldn't reference any beans which potentially throw uncaught exceptions. Otherwise you could get caught in an awkward loop. Closing off.

Norepro on demo section of https://showcase.omnifaces.org/validators/validateUnique Can you provide a reproducer?

Ok, took a closer look at source code, this is reproducible when following 2 conditions are met: - validator is placed *after* all the components - INTERPRET_EMPTY_STRING_SUBMITTED_VALUES_AS_NULL context param is...

Fixed in Mojarra side. Work around for now, put validator *before* the target components, as shown in showcase.

Fix has been revered in Mojarra side because a TCK failed. I need to take a second look.