On Thu, Jan 15, 2004 at 04:17:19PM +0000, Steve Harris wrote: > On Thu, Jan 15, 2004 at 04:58:26PM +0100, Alfons Adriaensen wrote: > > On Thu, Jan 15, 2004 at 03:35:05PM +0000, Steve Harris wrote: > > > > > the difference is that it makes the plugin > > > logic, for 95% of plugins, much simpler. > > > > I fail to see that. Only plugins that specifically requested NULL pointers > > by setting the hint bit have to test for them. The 95% others, including all > > existing ones, can just ignore the whole issue, they will always get valid > > pointers. > > Yes, OK, maybe its not a significant as I think, I still think it is > still more complex - you introduce a new hint, and the plugin can no > longer assume that all buffer pinters are valid - though it has to > be aware that it might not be. > > So I think this would be an acceptable solution for LADSPA, and it > would involve fewer changes to existing hosts than adding refernce > buffers.
I agree. It was not clear at all how the plugin should obtain the dummy pointer value to compare with -- that would probably require more invasive changes. OK, now that we agree :-), I'll leave to you the honour of designating the new hint bit. If you like I'll add the necessary comments to the header file and send it for approval this weekend. -- Fons
