- disable "-Wsuggest-override" in tests

* Solves #11156

* Check file size for all file format. Disable APPEND if file is empty

* Add test for APPEND mode

In case of regression trees, node risk is computed as sum of squared
error. To get a meaningfull value to compare with it needs to be
normalized to the number of samples in the node (or more generally to
the sum of sample weights in this node). Otherwise the sum of squared
error is highly dependend on the number of samples in the node and
comparision with `regressionAccuracy` parameter is not very meaningful.

After normalization `node_risk` means in fact sample variance for all
samples in the node, which makes much more sence and seams to be what
was originaly intended by the code given that node risk is later used as
a split termination criteria by
```
sqrt(node.node_risk) < params.getRegressionAccuracy()
```

- Recommended compiler checks:
  - GCC: CV_GCC
  - Clang: CV_CLANG
- fixed problem with CMAKE_CXX_COMPILER_ID=Clang/AppleClang mess on MacOSX
  Details: cmake --help-policy CMP0025
- do not declare Clang as GCC compiler