I've found what it is actually, if anybody else ever bumps into it. Having a "growable" set of grouped ports at the top of the def will grow its types into subsequent ports instead of moving the stack upwards from the bottom. It's a bit hard to explain, but all I can say for now is that if you have port groups that can grow and multiple groups, they can interact in nasty ways, so make sure you always sort growables bottom-most. It might be me not doing the right thing or not knowing another prescribed way to do it, but nonetheless this is insidious as hell.
On Tue, Feb 17, 2015 at 12:50 PM, Raffaele Fragapane < [email protected]> wrote: > Thanks Matt, > It's not related in this case, I managed to suppress the database error, > but I still have problems with mis-typed and stale typed ports n the ice > context. > > At this point it's got some really bizarre connotations. > > For the record the above is all persisting even if I not only restart the > app, but also nuke the preferences and burn incense at the four corners of > the desk. > > On Tue, Feb 17, 2015 at 12:16 PM, Matt Lind <[email protected]> wrote: > >> I have experienced that behavior with PPG logic for custom plugins and >> shaders. Usually happens when performing some illegal command and >> Softimage ducks for cover to protect itself from crashing. Whatever >> internal data Softimage was managing gets corrupted/deleted and you must >> restart the application to rebuild it to continue work. Whatever the >> issue, the internal data seems to be dynamically built. On the rare chance >> you have to deal with .spdl and related bits, you might want to flush your >> .spdlindex file in your user profile as a last resort, but don’t touch it >> if you don't have to. >> >> Matt >> >> >> >> >> Date: Mon, 16 Feb 2015 14:39:35 -0000 >> From: "raff" <[email protected]> >> Subject: ICE Node Database Invalid Error >> To: <[email protected]> >> >> >> Hey folks, >> >> I'm getting some very weird behaviour with a node where XSI issues a >> warning on reload that the database for ICE Nodes is invalid followed by >> the name of the node I'm prototyping. >> >> The node definition behaves erratically, with me having had to change >> things such as the names of the callbacks or the port group ID for >> parameters to actually reflect changes, otherwise they would do weird stuff >> like appearing in the PPG as the right type (or not appear at all), but >> actually internally be flagged as a different parameter (with related error >> about mis-matching CDataArray type, such as a parameter that's clearly a >> float insisting it's a matrix). >> >> Anybody ever encountered anything like it and has an idea if and how this >> database can be purged or what would lead to these issues? It's making for >> dev hell. >> >> Cheers, >> >> Raff >> > > > > -- > Our users will know fear and cower before our software! Ship it! Ship it > and let them flee like the dogs they are! > -- Our users will know fear and cower before our software! Ship it! Ship it and let them flee like the dogs they are!
