Tests are not taken into account when analyzing unused methods

I am TDDing small application and have just noticed production methods called just from test code are grayed as never used method. I don't think it is ideal behaviour, because it suggests I could delete it and being surprised later. On the other hand it is correct from target point of view - it really is not used (yet) from production target.

Could we have either:

• taking all callers into account and avoid incorrect unused method warning or
  
• have configurable behaviour so that everybody can decide what suits them better or
  
• having it as it is, but being able to see some hint it is called just from tests?