The new API for content attributes is too strict while reading an attribute.

If the attribute is a long and is stored as a string, we get an error.

We should return "true" on isCompatible if this is a string and it's parsable to a long.

Also, a specific debug log would be benefit to identify problematics attributes.