First, my aplogies to this list that I volunteered to do this mod to fluid~ and probably almost a year later, haven't had the time to tackle it yet.
Well, I might actually have a few hours later on today to work on this, and just would like to see if the work would still be revelant.
The motivation in my case is to use a large piano sounfont, loaded only once, and to use multiple instances of fluid~ to seperately process each voice.
The simplest idea I came up with would be to simply compare the file strings, and if the same, to share the soundfount rather than loading it again.
If I don't hear from anyone about a better approach, I hope to go ahead and tackle this.
One related questions - is it safe to use C++ STL as is, or should I redefine the allocator to use pd_new (or whatever it's called)?
Larry