JBTM-3883 use a base64 encoder if required

[mmusgrov]

https://issues.redhat.com/browse/JBTM-3883

I re-opened JBTM-3883 to add this extra check. I added a comment to the issue and to the test in this PR that explains the details of why the base64 encoder might be needed.

CORE !AS_TESTS !RTS !JACOCO !XTS !QA_JTA !QA_JTS_OPENJDKORB !PERFORMANCE !LRA !DB_TESTS !mysql !db2 !postgres !oracle

[jbosstm-bot]

Started testing this pull request with CORE profile: https://ci-jenkins-csb-narayana.apps.ocp-c1.prod.psi.redhat.com/job/btny-pulls-narayana/PROFILE=CORE,jdk=openJDK17,label=jnlp-agent/469/

[jbosstm-bot]

Started testing this pull request with CORE profile: https://ci-jenkins-csb-narayana.apps.ocp-c1.prod.psi.redhat.com/job/btny-pulls-narayana/PROFILE=CORE,jdk=openJDK11,label=jnlp-agent/469/

[jbosstm-bot]

CORE profile tests passed - Job complete https://ci-jenkins-csb-narayana.apps.ocp-c1.prod.psi.redhat.com/job/btny-pulls-narayana/PROFILE=CORE,jdk=openJDK11,label=jnlp-agent/469/

[jbosstm-bot]

CORE profile tests passed - Job complete https://ci-jenkins-csb-narayana.apps.ocp-c1.prod.psi.redhat.com/job/btny-pulls-narayana/PROFILE=CORE,jdk=openJDK17,label=jnlp-agent/469/

[jbosstm-bot]

Started testing this pull request with CORE profile: https://ci-jenkins-csb-narayana.apps.ocp-c1.prod.psi.redhat.com/job/btny-pulls-narayana/PROFILE=CORE,jdk=openJDK17,label=jnlp-agent/470/

[jbosstm-bot]

Started testing this pull request with CORE profile: https://ci-jenkins-csb-narayana.apps.ocp-c1.prod.psi.redhat.com/job/btny-pulls-narayana/PROFILE=CORE,jdk=openJDK11,label=jnlp-agent/470/

[jbosstm-bot]

CORE profile tests passed - Job complete https://ci-jenkins-csb-narayana.apps.ocp-c1.prod.psi.redhat.com/job/btny-pulls-narayana/PROFILE=CORE,jdk=openJDK11,label=jnlp-agent/470/

[jbosstm-bot]

CORE profile tests passed - Job complete https://ci-jenkins-csb-narayana.apps.ocp-c1.prod.psi.redhat.com/job/btny-pulls-narayana/PROFILE=CORE,jdk=openJDK17,label=jnlp-agent/470/

[mmusgrov]

I've put this on hold while I think through the impact on crash recovery..

[mmusgrov]

I've removed the Hold label because Marco, or Manuel, correctly pointed out to me that SHA-224 hashing and base64 encoders are deterministic functions and as such will always produce the same output for a given input string, ie assuming that any particular integration gives us the same string we'll be able to detect records that "belong" to the recovery manager.

Note that hashing and slicing, although increasing the probability of collisions, are acceptable because the chances of that happening are extremely low (there was some analysis about probabilities in the associated quarkus issue and PR.

[marcosgopen]

Thanks Mike for your PR, LGTM!