Merge schemata fragments to schemas. A schemata from this pass may may contain multiple mutation kinds and span over multiple files.
Save fragments to use them to build schematan.
All mutants that have been used in any generated schema.
Minimal mutants that a schema must contain for it to be valid.
Max mutants per schema.
Build scheman from the fragments.
TODO: optimize the implementation. A lot of redundant memory allocations etc.
Conservative to only allow up to <user defined> mutants per schemata but it reduces the chance that one failing schemata is "fatal", loosing too many muntats.