Re: n_map "freq" -1

> > On Aug 25, 2004, at 11:38 AM, Julian Rohrhuber wrote:
> >
> > >   if I set consecutive controls by giving a name and then a 
> > > number, my only intention could reasonably be to set the 
> > > consecutive _unnamed_ controls.
> > >   Setting other named controls that follow after these is not 
> > > really what a message like /n_mapn "freq" 0 5  expresses.
> >
> > Yes it IS what that expresses. That was what it was meant to 
> > express originally.
> > If I want to map \gate, \freq, \amp, \pan on consecutive buses, then I can do
> >
> > \n_mapn, id, \gate, bus, 4
> >
> > as long as these are the same rate controls.
>
> sorry to be so persistent.
>
> \n_mapn, id, \gate, bus, 4
>
> does not really express very well the mapping of the four controls 
> \gate, \freq, \amp, \pan.
>
> it is practical, but not expressive. Do you always remember the 
> order of args in a synthdef?
> or regarding to the parallelism to argument passing:
> as soon as using keywords, all other args have to be keywords, too.


ok, I see now. a ParamSpec only keeps its own index, and I guess it 
would need some other methods to make a node iterate over its 
controls, finding whether a name exists for it. It would mean a 
different osc message.
It would be very handy, but I don't know if it is reasonable.

something like this:

void Graph_MapNControl(Graph* inGraph, int32 inHash, int32 *inName, 
uint32 inNumChannels, uint32 inBus)
{
        ParamSpecTable* table = GRAPH_PARAM_TABLE(inGraph);
        ParamSpec *spec = table->Get(inHash, inName);
        if (spec) {
                uint32 index;
                Graph_MapControl(inGraph, spec->mIndex, inBus);
                for (uint i=1; i<inNumChannels-1; ++i) {
                        index = inBus + i;
                        int32 j = table->IndexFor(inHash, index);
                        *spec = table->mItems[j];
			if(*spec->mName >= 0x80000000) return // stop 
at next named control
                        Graph_MapControl(inGraph, spec->mIndex, inBus + i);
                };
        }
}
(this code is not correct, because I don't know how to derive a key 
from a value in the HashTable)

the command could be called n_mapnc (clip) maybe.
But please tell me if it is a bad idea, I appreciate that the server 
should be kept simple.
--








.