ReplicationGroup engineVersion check invalid - 0.49.1

[dhumphries-sainsburys]

What happened?

This PR #1985 i believe results in a bug where existing redis clusters with engineVersion: 6.x set are failing to sync due to the check for version trying to cast x to an int in 6.x despite it being a completely valid option. Curiously creating a new replicationgroup with 6.x doesn't have any issues as of yet

How can we reproduce it?

1. Have an existing replicationgroup as so (obviously switch out the bits required for your test env)

apiVersion: cache.aws.crossplane.io/v1beta1
kind: ReplicationGroup
metadata:
  name: test-redis
spec:
  deletionPolicy: Delete
  forProvider:
    applyModificationsImmediately: true
    atRestEncryptionEnabled: true
    authEnabled: true
    automaticFailoverEnabled: true
    cacheNodeType: cache.t3.micro
    cacheParameterGroupName: default.redis6.x
    cacheSubnetGroupName: lab-data
    engine: redis
    engineVersion: 6.x
    multiAZEnabled: false
    numCacheClusters: 2
    preferredMaintenanceWindow: wed:22:00-wed:23:00
    region: eu-west-1
    replicationGroupDescription: test
    securityGroupIdRefs:
      - name: lab-elasticache-sg
    securityGroupIds:
      - sg-0f0506ae5e242fa6fe
    snapshotRetentionLimit: 0
    snapshotWindow: 01:00-02:00
    transitEncryptionEnabled: true
  managementPolicies:
    - '*'
  providerConfigRef:
    name: aws
  writeConnectionSecretToRef:
    name: test
    namespace: flux-system

2. Upgrade to 0.49.1 (in theory anything after 0.47.0 probably has this issue)

3. Observe existing elasticache resource has error as so

    - lastTransitionTime: '2024-08-21T09:15:00Z'
      message: >-
        update failed: unable to parse version number: unable to parse version
        number: strconv.Atoi: parsing "x": invalid syntax
      reason: ReconcileError
      status: 'False'
      type: Synced

What environment did it happen in?

Crossplane version: 1.16.0
Provider version: 0.49.1
K8s Distro: EKS
K8s Version: 1.30