replace udis with iced-x86

[LunNova]

#36

Not properly tested/needs more work.

[Hpmason]

Thanks for working on this. Any other thoughts about the safety are welcome

[LunNova]

I feel like using region as an extra dependency internally might be going too far but can go for that if you want.

I've edited the doc for Builder::build to note that target..target+margin+15 must be valid to read as u8 to avoid UB, and applied the changes you requested.

[LunNova]

CI failure for i686-pc-windows-gnu seems to be unrelated, egor-tensin/setup-mingw@v2 is failing.

[Hpmason]

I feel like using region as an extra dependency internally might be going too far but can go for that if you want.

I've edited the doc for Builder::build to note that target..target+margin+15 must be valid to read as u8 to avoid UB, and applied the changes you requested.

Looking at Detour::new, it already does some region checking so the docs for Builder::build should be enough.

CI failure for i686-pc-windows-gnu seems to be unrelated, egor-tensin/setup-mingw@v2 is failing.

Yeah, it's not the first time it's broken. I need to go in and fix what's breaking it, possibly look into finding a better GH Action to rely on. But it should be fine for this PR.

Thanks again for working on this replacing udis, I really appreciate the help!

[Hpmason]

closes #36