On Mon, 26 May 2008, Chen Kedem wrote: Hi Chen,
> > It means that we do not have to remove additional parameters from HVM stack > > but > > internally we can use all functions like functions with variable number of > > parameters > Such move will not interfere with future strong type checking? It won't. We are talking about values which cannot be accessed directly by variable name so it will not interact with grammar rules. It's only RT behavior. > How about performance penalty? In practice nothing what can give noticeable performance difference. Probably full cleanup of redundant code will allow to eliminate few conditional jumps so it will be even a little bit faster but it can be detected only in very precise test strongly oriented to measure exactly this aspect. IMHO it's not wroth to invest time for such full modification now though in the future I'll make such cleanup. It's always good to eliminate unnecessary exceptions. Now I only want to restore original Clipper behavior for PCOUNT() without breaking or introducing anomalies to HB_PVALUE() and HB_APARAMS() best regards, Przemek _______________________________________________ Harbour mailing list Harbour@harbour-project.org http://lists.harbour-project.org/mailman/listinfo/harbour
