Make SimpleDataBinder handle multiple indexers

### Feature description

Hello,

SimpleDataBinder uses this regex to parse the indexer and it's value from a value it wants to map:
`static final INDEXED_PROPERTY_REGEX = /(.*)\[\s*([^\s]*)\s*\]\s*$/`

But this regex will always choose the last indexer when there are multiple present. Example:
`propName = 'columns[0]['data']'`
will be matched as
```
indexedPropertyName = 'columns[0]'
index = 'data'
```

I'd argue that this is wrong. A groovy property is not allowed to be named `columns[0]` anyways (right?) so matching that makes no sense. It should instead consider the first indexer as the indexer that matters. It should be:
```
indexedPropertyName = 'columns'
index = '0'
```

It would be easy to change the regex to pick the first indexer instead like so:
`^(.*?)\[\s*([^\s]*)\s*\]\s*`
but I'm not sure if that would break anything and if that's enough to properly support nested indexers. Is this a feature the SimpleDataBinder should support? Since fields that have [] in their name are probably very unlikely to exist, making this change may not be that breaking in real usage.

Start by locating SimpleDataBinder and its INDEXED_PROPERTY_REGEX definition, then trace how the regex groups become the indexed property name and index. Check the existing binder coverage before deciding how nested indexers should behave; done means multiple indexers are parsed according to the agreed semantics without breaking existing property binding.

Written by the indexing model from the issue text.