Xer Shadow Tail ☄️🌺 @XerTheSquirrel@awoo.space

#SquirrelJME: Can run a Game Boy emulator and it actually runs pretty nicely. Only drawback is that the emulator uses purely numbers for control. I think I need alternative control scheme's now. https://t.co/WqD0OHAfDV

#SquirrelJME: I toyed building a blank executable for DOS, but it does not run on Windows 10. https://t.co/5w8YdGM9Ls

#SquirrelJME: Have to deal with these segmentation faults again since I am doing native stuff. https://t.co/5DL0SToQ4C

#SquirrelJME: This is another test, please do disregard this one! FYI I am redoing the publish system for updates on SquirrelJME for when I post something about my project!

#SquirrelJME: This is how we played #Doom in the mid 2000s! No fancy Android/iOS devices! Definitely shows that I need better screen size handling code.

#SquirrelJME running a basic block falling game! 💖

https://www.youtube.com/watch?v=-N0kbRYUoGs

#SquirrelJME: Opera is having a rough time connecting to the internet when there is no network code support.

#SquirrelJME: So the latest stack cached compiler for Java byte code has definitely been shaping up and I am at the point where I can export the result to compactified binaries. Of course there will need to be a multitude of tests just to make sure it works properly!

#SquirrelJME: I rewrote the stack handling code that I previously wrote and made this one immutable! Actually making it immutable makes it easier to use and more precise, which is nice!

#SquirrelJME: SquirrelJME has reached 11k commits! Yay!

#SquirrelJME: I have enforced HTTPS for the main SquirrelJME website, so anyone who uses HTTP will be automatically redirected to HTTPS. Also the site is fully available on IPv6 as well and it should work.

#SquirrelJME: Do not forget you can support this project on Patreon at https://www.patreon.com/SquirrelJME ! Your patronage is much supported!

#SquirrelJME: Refactoring my Java byte code to register code compiler and I decided that I need to map out all of the possible instructions along with their encodings. Currently there will be 16 groups/encodings of 16 instructions each.

#SquirrelJME: My compiler is making a bunch of reference uncounts and returns at the end, so I figured I can just order these so they all fall into each other!

#SquirrelJME: While working on this Java byte code to register compiler, even with basic caching I am noticing so far that the number of instructions is around half original size.

#SquirrelJME: So now useless jumps to the following instruction are removed, so this constructor which calls another constructor is now just a few instructions!

#SquirrelJME: Now duplicate reference count cleanups are not made and any exception handlers which point to these cleanups are removed. This makes the code a bit smaller with less duplicate areas which is nice.

#SquirrelJME: I implemented what I call QuickCaching, basically if a local variable that is never written to placed on the stack it becomes cached. So in the case of this method call it just uses the cached local instead. This makes the code faster and smaller!

#SquirrelJME: So I have this small constructor which calls another constructor translated to register format. I have noticed the duplicate UNCOUNT+RETURN for returns/exceptions. I figured there would be lots of duplicate code for various return points, so I am going to condense.

#SquirrelJME: Working on the translation of Java byte code to a format which uses registers and also has exception handlers which are part of the code itself. Making exceptions just part of the code simplifies the design of the upcoming virtual machine since now it is just code.