| |
|
|
| i wish me one small experiment, that without Roland hardly goes.
I faith i'm a idea invaded, How one XProfan möglicherweise relatively simply and without Changes in the structure sau quick get.
something Tiparbeit is But already, but that "muss" Yes Roland make. *duck*
I put me the so to, roughly 2 works:
1.
plainer beginning with CLS [n]/ SetText.
and a possible previous function part only between this new real functions and calls suitable on - integrally normal Interpretermodus.
2.
The function ProcAddr could one procure, that they, z.B. if only 1 Parameter transfer watts or the 1. Parameter a Long-Constant is, a scheduler Funktionsadressen returns:
so against would it possible, To Programmbeginn these Funktionsadressen To relating as well as in API-functions and from Objektcode out aufzurufen.
naturally went instead of by ProcAddr too a Bereichsvariablenadresse or ProcAddr(0) could a Bereichsvariablenadresse zurückliefern which The Funktionsadressen holds.
I mean, on this step aufbauend are then only yet adjustments the in XProfan installed Präkompilers necessary, circa Piece for Piece the XProfan version of To Version always and always faster go To let.
moreover comes, I so by Precompiler too same offer could, not only The "natives Profan"-functions nativ available to to put, separate too The XProfan-functions.
it would imho a Verschwendung, this only over slower Wrapper To make possible what about me know nothing better, as The real XProfan-functions with nativer speed anzubieten. |
|
|
| |
|
|