Needed for merging detector information from different
detectors banks

Quite a simple but important test, because these functions would not work via NormaliseByDetector unless they were factory registered. Also provides checking that they are types of IFunction, since the test downcasts the boost shared_ptr.

Processes the peaks in the rhs workspace, removing the first match
it finds (in Q, within the given tolerance) from the peaks in the
lhs workspace.

Follows exactly what's written in the ticket. Peaks in the second
workspace that are within the given Q tolerance in all directions
don't make it into the output.
Note that it's quite possible for a given peak in the first workspace
to be matched to multiple peaks in the second, if the tolerance
is large enough/the peaks are close together.

Pixel ID should follow time-of-flight in the event structure - not the
other way around.

Refs #5367

Also added a diagnostic print to show off what the difference in the
number of masked spectra versus detectors.

Fixes the problem of only ignoring OutputWorkspace_ properties so it was
returning the additional MonitorWorkspace properties. It now does
not return any property with an underscore in the name.

One of the loaders that SNS supports, LoadEventPreNexus, does not use the
Filename property. Therefore, I've made a way to set that.