> Map<String,Object> scalingPolicyMap = Maps.newHashMap();
> scalingPoliciesList.add(scalingPolicyMap);
> scalingPolicyMap.put("cooldown", scalingPolicy.getCooldown());
> scalingPolicyMap.put("type", scalingPolicy.getType().toString());
> scalingPolicyMap.put("name", scalingPolicy.getName());
> // A couple of different scaling policies are supported, such as
> percent or number based, or targeting specific numbers of instances
> - scalingPolicyMap.put(scalingPolicy.getTargetType().toString(),
> scalingPolicy.getTarget());
> + String targetString = scalingPolicy.getTarget();
> + Integer targetInt = Ints.tryParse(targetString);
> + Float targetFloat;
> + if (targetInt != null) {
> + scalingPolicyMap.put(scalingPolicy.getTargetType().toString(),
> targetInt);
> + } else if ((targetFloat = Floats.tryParse(targetString)) != null) {
> + scalingPolicyMap.put(scalingPolicy.getTargetType().toString(),
> targetFloat);
> + } else {
> + scalingPolicyMap.put(scalingPolicy.getTargetType().toString(),
> targetString);
> + }
Hm...is there any way we could use polymorphism here? Each `TargetType` should
_know_ how to correctly return its scaling value, no? Could we move the parsing
logic in there and then reduce this to something like:
```
scalingPolicyMap.put(scalingPolicy.getTargetType().toString(),
scalingPolicy.getTargetType().toValue(scalingPolicy.getTarget()));
```
or, indeed, have that on the scalingPolicy itself (since it should be able to
access both the target type and value):
```
scalingPolicyMap.put(scalingPolicy.getTargetType().toString(),
scalingPolicy.getTargetValue() // does the appropriate conversion);
```
?
---
Reply to this email directly or view it on GitHub:
https://github.com/jclouds/jclouds-labs-openstack/pull/39/files#r7234823
