Dear George, You made a really good point. Unfortunately, this new feature does not make any difference to virtual function. However, similar feature is already provided as STUB class compilation. Please look into demo/makecint/Stub2 for example. You can precompile part of derived class definition so that you can call virtual function correctly accross the interpreted-compiled class partition. The STUB class has to define all the data members and virtual function interface. Interpreter can give implementation to the virtual function and non-virtual member function. I will investigate if it is possible to combining protected member stub and virtual function stub. Thank you Masaharu Goto >Masa, > >This is wonderful! Great! > >If I understand right, this functionality could also provide for >virtual functions being called correctly in interpreted derived >classes, which is arguably more important. Does this change already do >that or is more work required to get that? > >Many thanks for this change, though -- I'll be most pleased once we >actually get recent versions of CINT to propagate into our standard >Root. > >George > >
This archive was generated by hypermail 2b29 : Tue Jan 02 2001 - 11:50:28 MET