+ //
+ // Note: Doing this, this way, triggered a nasty bug in
+ // wxTopLevelWindowGTK::GtkOnSize which took -1 literally once
+ // either of maxWidth or maxHeight was set. This symptom has been
+ // fixed there, but it is a problem that remains as long as we allow
+ // unchecked access to the internal size members. We really need to
+ // encapuslate window sizes more cleanly and make it clear when -1 will
+ // be substituted and when it will not.
+