How was encodings made?

[playX18]

Is it hand-made or somehow parsed from Intel manual? I am using instrs.txt in my project but looking to add RW info for operands and doing it by hand is quite painful.

[aengelke]

Is it hand-made or somehow parsed from Intel manual?

Hand-made and fuzz-tested against Intels XED.

add RW info for operands and doing it by hand is quite painful.

I do have plans to add rw-info and info on implicit operands at some point. It will be a larger effort, though, and I might want to add APX and AVX10.2 support first.

I am using instrs.txt in my project

Happy to hear this is helpful!

[playX18]

Is it hand-made or somehow parsed from Intel manual?

Hand-made and fuzz-tested against Intels XED.

add RW info for operands and doing it by hand is quite painful.

I do have plans to add rw-info and info on implicit operands at some point. It will be a larger effort, though, and I might want to add APX and AVX10.2 support first.

I am using instrs.txt in my project

Happy to hear this is helpful!

Glad to hear that you will work on RW info! I'll be waiting for it then. I tried to do it manually by adding [RrWwXx]: prefix to all operands and perhaps I can even make a PR then, but it would take me lots of time to make it complete...

My project is here if you're interested: https://github.com/playx18/uasm. I am also using your disarm project for ARM JIT but it's still WIP.

One thing I could not figure out from parseinstrs.py is how can I make encoder more easy to use: I tried to make it emit enum Encoding {...} which would contain all possible encodings and then match on opcode + operands passed to emit() to find out suitable encoding. So instead of doing add64rr() and add32rm() user could simply do add(RAX, RDI) and add(EAX, ptr(RDI, 16));