As noted a few times recently, I’ve cycled over into “demo mode” and have been working on… well, something that I’ll be keeping quiet about for a little while yet until it properly gets past that stage where it could be changed or indeed completely reinvented on a whim. Generally speaking though, my intention is to have something ready for X’2018 in early November – assuming we have a representative to take it along – and have a cluster of prototype routines along with some other ideas I’ve been meaning to try and a folder of logos because everybody loves logos… right?
Things have been going slowly but surely so far – today alone I somehow managed to metaphorically paint myself into a corner on two different occasions with what should’ve been relatively simple code, either choosing a less than sensible place to store things or simply running out of memory altogether – and, because it’s been a painfully long time since I worked on anything that wasn’t a onefiler, last week saw me sitting down and trying to get my head around cruncher and loader integration with some of the more recent iterations of popular tools with very little joy… at least until I discovered lft’s Spindle that is.
Spindle is, in essence, a back engine for demos which deals with loading and decompressing of data to the point where the user doesn’t need to be involved; no fighting assemblers to build loader or decruncher code, no messing around with building disk images of data, just document the files being used and the order they’re needed via a script file which is then run through Spindle to automagically get a working disk image out the far end. There are caveats of course – a couple of places in memory have to be avoided, something else I bumped into today – but having to do things just a teensy bit differently is worth it for the results.
Main project aside, there’s also been a little time spent generally doodling both with graphics and ideas for effects and a couple of forays onto other platforms; something might come of the latter if they survive testing on real hardware – in both cases the code was for a platform I don’t actually own right now, so that’s something of an issue – but at the very least it was all good practice.