PropagateToMirror - correcting for material budget and proper treatment of covariance

[miranov25]

The current implementation is ignoring the material crossed in PropagateToMirror

To be done in local linear approximation:

Details in JRA following comment:

• https://alice.its.cern.ch/jira/browse/PWGPP-613?focusedCommentId=292222&page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#comment-292222

[miranov25]

Relates to #9