[Gtk-sharp-list] [gtk-sharp] Matrix.Scale and Context.Scale anomaly?

[Hywel]

Using a Cairo.Matrix.Scale (sx, sy) with a zero value for either sx or 
sy gives the correct result (the Xx or the Yy component of the matrix is 
zero'd), but using Context.Scale with either zero value for sx or sy 
gives the wrong result for the Context Transformation Matrix.  There 
seems to be no valid reason why these should differ (mathematically they 
are valid, though perhaps unlikely, operations).  I'm not sure how the 
two functions call back into the Cairo 'C' code - can anyone help or 
shed any light on this apparent anomaly?