The current test generator (matcher generator) is appropriate for small objects, but does not really scale up:
- unprecise test failure locations
- hard to understand mismatch messages
Yet this problem seems to be partly caused by the complexity of the problem. We have to support following data types
- primitive data types (easy)
- simple composite data types (easy)
- primitive data types
- simple composite data types
- list composite data types (harder)
- containing primitive data types
- simple composite data types
- list or array types
- map composite data types (hard)
- containing primitive data types
- simple composite data types
- list or array types
- map or set types
The assertions of these types are
- immediate assertions on the value in case of primitive data types
- assertions on (possibly long) path expressions in case of simple composite data types
- assertions on path expressions with index selectors in case of list composite data types
- no idea for map composite data types
While simple composite types can be asserted through long path expressions
The current test generator (matcher generator) is appropriate for small objects, but does not really scale up:
Yet this problem seems to be partly caused by the complexity of the problem. We have to support following data types
The assertions of these types are
While simple composite types can be asserted through long path expressions