fix: check nils in ingressClass.Parameters

[randmonkey]

What this PR does / why we need it:

fix panics when some field is nil in ingressClass.spec.parameters

Which issue this PR fixes:

fixes #2958.

Special notes for your reviewer:

PR Readiness Checklist:

Complete these before marking the PR as ready to review:

• the CHANGELOG.md release notes have been updated to reflect any significant (and particularly user-facing) changes introduced by this PR

[rainest]

Initial failures in integration-tests-dbless were for UDPRoute and TLSRoute:

UDPRoute failed because the listener had no valid Kong port to associate with. Somehow this port was present on the deployment, but was TCP instead of UDP. Not sure how KTF would have that temporarily, but the code there does look at least partially incorrect. UDP stream listens should live in a different config section because of LoadBalancer inability to handle mixed TCP/UDP services:

09:51:05-0700 esenin $ grep -r 9999
pkg/clusters/addons/kong/vars.go:	DefaultUDPServicePort = 9999
09:51:08-0700 esenin $ grep -r DefaultUDPServicePort
pkg/clusters/addons/kong/vars.go:	// DefaultUDPServicePort indicates the default open port to be found on the Kong proxy's UDP service.
pkg/clusters/addons/kong/vars.go:	DefaultUDPServicePort = 9999
pkg/clusters/addons/kong/addon.go:	return urlForService(ctx, cluster, types.NamespacedName{Namespace: a.namespace, Name: DefaultUDPServiceName}, DefaultUDPServicePort)
pkg/clusters/addons/kong/addon.go:		"--set", fmt.Sprintf("proxy.stream[1].containerPort=%d", DefaultUDPServicePort),
pkg/clusters/addons/kong/addon.go:		"--set", fmt.Sprintf("proxy.stream[1].servicePort=%d", DefaultUDPServicePort),
pkg/clusters/addons/kong/addon.go:		Port:     DefaultUDPServicePort,

However, elsewhere it tries to set the udp parameter manually, so that should satisfy what Gateway wants to check:

 562 func exposePortsDefault() []string {                                                                                    
  563     return []string{                                                                                                    
  564         "--set", fmt.Sprintf("proxy.stream[0].containerPort=%d", DefaultTCPServicePort),                                
  565         "--set", fmt.Sprintf("proxy.stream[0].servicePort=%d", DefaultTCPServicePort),                                  
  566         "--set", fmt.Sprintf("proxy.stream[1].containerPort=%d", DefaultUDPServicePort),                                                
  567         "--set", fmt.Sprintf("proxy.stream[1].servicePort=%d", DefaultUDPServicePort),                                  
  568         "--set", "proxy.stream[1].parameters[0]=udp",                                                                   
  569         "--set", "proxy.stream[1].parameters[1]=reuseport",                                                             
  570         "--set", fmt.Sprintf("proxy.stream[2].containerPort=%d", DefaultTLSServicePort),                                
  571         "--set", fmt.Sprintf("proxy.stream[2].servicePort=%d", DefaultTLSServicePort),                                  
  572         "--set", "proxy.stream[2].parameters[0]=ssl",                                                                   
  573         "--set", "proxy.stream[2].parameters[1]=reuseport",                                                             
  574     }                                                                                                                   
  575 }

TLSRoute unsure. Its listener is fine, but the route itself has no status and was apparently never acknowledged. Controller logs appear to simply not see it.

[pmalek]

Those would better use require instead to prevent continuing execution, so that when an error occurs we don't continue the test as it would be pointless.