find_resultrow: handle awkward section names better (#4)
    
    Issue #4 highlights an awkward section naming case in the desktop
    matrix, where we have sections with very similar names where one
    is a substring of the other. Without this change it's impossible
    to actually get `find_resultrow` to distinguish between the two
    sections. This makes it possible in two ways: by using the exact
    section name (unfortunately including the awkward HTML-translated
    bold tags...), or the differing case of the "r" character. We do
    this by (as with test names) trying an *exact* name match if the
    substring match didn't get us down to a single row, and also
    trying a non-normalized match if we still didn't get down to a
    single row.
    
    Signed-off-by: Adam Williamson <awilliam@redhat.com>