Allow removal of config_entries via main class

[ekohl]

This allows removal of config entries by setting their value to undef. In doing so, users can remove values via hiera. One example is checkpoint_segments which was removed in PostgreSQL 9.5, but commonly used as a tuning parameter in older versions.

I'd like some verification on whether undef values really couldn't be passed in for some other use case, but I can't imagine any.

[puppet-community-rangefinder]

postgresql::server is a class

Breaking changes to this file WILL impact these 36 modules (exact match):

• bltavares-baseline
• nicopaez-dockerserver
• jmkeyes-razor
• jjulien-trac
• atmesh-deploy_railsapp
• dmexe-rails
• landcareresearch-pidservice
• eschiller-trac
• driebit-zotonic
• katello-gutterball
• landcareresearch-postgis
• chedi-django
• nicopaez-alfred
• nicopaez-triage
• mjhas-mailserver
• lwo-dataverse
• urgi-galaxy_roles_profiles
• hetzner-roundcube
• stackstorm-st2
• mukaibot-bamboo
• sensson-powerdns
• maestrodev-maestro
• npwalker-pe_external_postgresql
• zleslie-bacula
• puppetlabs-puppetdb
• SchnWalter-happydev
• groupbuddies-gb
• blackknight36-bacula
• conzar-ckan
• conzar-pidservice
• treydock-keycloak
• theforeman-foreman
• katello-candlepin
• theforeman-pulpcore
• nnutter-testdb
• Ginja-puppet_stack

Breaking changes to this file MAY impact these 14 modules (near match):

• jmkeyes-netbox
• lboynton-gitlab
• knowshan-phppgadmin
• Firebladee-moodle
• fiddyspence-zabbix
• binford2k-drupal
• dmexe-deploy
• 4geeks-odoo
• simp-foreman
• maestrodev-continuum
• vshn-uhosting
• ash-netbox
• Lavaburn-cabot
• cleo-harmony

This module is declared in 71 of 575 indexed public Puppetfiles.

---

These results were generated with Rangefinder, a tool that helps predict the downstream impact of breaking changes to elements used in Puppet modules. You can run this on the command line to get a full report.

Exact matches are those that we can positively identify via namespace and the declaring modules' metadata. Non-namespaced items, such as Puppet 3.x functions, will always be reported as near matches only.

[david22swan]

@ekohl
Making a quick change to point your PR at the newly default main branch.
Apologies if this is inconvenient in any way.

[sanfrancrisko]

I may circulate your query around the IAC Team @ekohl :

I'd like some verification on whether undef values really couldn't be passed in for some other use case, but I can't imagine any.

But, in principle, it seems OK to me.

It does appear as though the test you added is failing. From what I can see, in this loop, we're receiving:

$entry = fsync
$value = 'off'

...four times. I'm not seeing any of the additional params being evaluated.

[ekohl]

It looks like nil isn't mapped to undef in the tests. Trying with :undef to see if that's mapped to undef.

[ekohl]

Spec tests pass now. The acceptance tests fail, but those look unrelated to my changes.

[daianamezdrea]

Hi @ekohl, indeed, the failures from acceptance tests are not from this PR, thank you for submitting the PR! I'll have a look to see what is happening here. Cheers!

[sanfrancrisko]

Thanks for the enhancement @ekohl and addressing the test failures.

I've rebased with puppetlabs:main to pull in a workaround for the acceptance test failures. I'll merge the PR when the tests complete (and go green).