After getting the map code working, I rewrote the code to load the attributes into the memory buffer. This involved updating the asset tool to properly remap the metatiles to properly allow for attribute value lookup. In the end, we’re still running with too many cycles in NMI, but we’ll get that fixed on Thursday. Thanks for watching!
Broken, fixed, and better. The map loading code was changed so that the asset tool can dump the entire map directly in to the file and then we seek into it and decrement our index vs our prior approach where it had the first screen at the beginning and then subsequent rows following. That went pretty quick and then we talked about what we would do for the next stream to get palettes and attributes working again with this new format, furthering our goal of finally getting to tile collision. Thanks for watching! See you on Monday
Christmas. Star Wars. And new(ish) games for the Nintendo Entertainment System! Joe and Austin introduce unsuspecting JDUBS patrons to the grandaddy of new NES Hombrews, Battle Kid!
Memblers is heavily working on a new NES cart. If you are a developer yourself, keep an eye out!
Quote:
#NESdev prototype teaser: A new GTROM board variation enters testing phase this week, with (optional) compatibility with crappy NES clones, and has a fun new feature unique to this cart. More info will be posted as development continues.
In this episode, we pair a golden sour blended with nectarines and blackberry and a game about the high stakes world of cowboy-taunting, light-gun weilding Russian roulette.
So, instead of working on collision today like I intended I ended up working on optimization of the code using Mesen’s awesome debugger and event viewer. I wasn’t feeling comfortable with the performance we were seeing in NMI and worked to get a better understanding of what was going on so that we can get our collision running in a later stream.
In this episode, I got the MMC3 banks swapped properly to show our game sprites, but it’s still not 100% clear why. I’ll play around with this and then explain it once I have a better understanding (hopefully Monday). After that, I talked about how I wanted to make world collision work and started to implement it, but it was very inefficient and we’ll have to revisit on Monday after I’ve had some time to think about it. Thanks for watching and have a great weekend!
Okay, attributes are working finally! After I got that taken care of, I spent some time looking at MMC3 problems we’re having as revealed by Mesen. MESEN IS AMAZING. We will be using that as our emulator moving forward because: holy crap, it’s so good.
So, more attribute programming… I know. But, we made some really good progress! We have a better handle on loading the attributes and where we have performance issues that are impacting our NMI code and giving us a problem. To be continued on Thursday!