Thanks for the report, it seems I didn't have the proper fix to the sprite issue. I fixed it correctly and backported to 2.3.2, so the current download of 2.3.2 should be good now.
I fixed RTC emulation and now pass all of the tests by CasualPokePlayer. I think a lot of things are still wrong, but this now runs a lot closer to how the actual chip works so is improvable in the future if more tests are made.
Link to video
I console verified Rockman Zero 2. With a few other NES games I recently verified, this brings us too 490 current verified runs, only 10 to go to get to 500!
I still want to eventually get 4 player linking working, but since the performance will be ~30 fps I decided to try doing some multi-threading on the 2-player linked core to see what, if any, improvements I might expect.
It turns out instead of an improvement I got big performance penalty, from 60 fps to 20 fps. I guess not enough work is done in a single cycle to offset the synchronization overhead. I would have to do some kind of bulk processing of cycles to make it worth while, but it's not worth breaking my one cycle at a time rule for.
So based off that for 4 players it might just break even with single threaded, not really worth it. I think it was a worth while experiment though. I'm not exactly a multi-threading mastermind but at least now I have some idea of when I can expect it to be useful.
I've been rewriting GBHawk as a C++ core so that all my cores have the same architecture. I've managed to get linking working in a much more compact way in the new core. I'll take what I learned in this process to redo the GBA linking for 4 players eventually.
I've also moved my NES core here as experimental so that everything is in the same place, no point having things spread out to multiple projects.
I've also added blank cores for SNES and N64. No clue when I might actually develop them, but I've found that when I'm too tired to do core development I can still make progress on the infrastructure stuff that doesn't take a lot of brain power largely copy paste from the other cores.
I've also removed a LOT of stuff from the front end code as I try to shrink and simplify it. Eventually I will move to the same lua structure as current BizHawk which will be another simplification.
Now that the GBHawk rewrite is done I can get back to GBA linking improvements and work on a new release.
I've been reworking my linking code to closer match how hardware works. This results in much more bookkeeping on the core side but much simpler linking logic. Now I match the current mGBA serial tests to my GBP, as well as have code that is much easier to improve when I start to implement the other linking modes.
As I was working on this I came across this cool video of a demo that streams video from one GBA to another:
Link to video
This isn't quite working yet on GBAHawk. I'm not sure if I will try to fix it for this release or save it for the next one.
Flevoriux has also recently released a new version of the painful sprite horizontal mosaic test that I currently don't pass. I haven't looked into the details of this one yet, hopefully this doesn't break my increasingly strained pixel pipeline.
I fixed RTC emulation and now pass all of the tests by CasualPokePlayer. I think a lot of things are still wrong, but this now runs a lot closer to how the actual chip works so is improvable in the future if more tests are made.
I've been rewriting GBHawk as a C++ core so that all my cores have the same architecture. I've managed to get linking working in a much more compact way in the new core. I'll take what I learned in this process to redo the GBA linking for 4 players eventually.
I've also moved my NES core here as experimental so that everything is in the same place, no point having things spread out to multiple projects.
I've also added blank cores for SNES and N64. No clue when I might actually develop them, but I've found that when I'm too tired to do core development I can still make progress on the infrastructure stuff that doesn't take a lot of brain power largely copy paste from the other cores.
I've also removed a LOT of stuff from the front end code as I try to shrink and simplify it. Eventually I will move to the same lua structure as current BizHawk which will be another simplification.
Now that the GBHawk rewrite is done I can get back to GBA linking improvements and work on a new release.