feat(fw): allow adding verbatim bytes to Bytecode

[chfast]

🗒️ Description

This provides a basic requested feature to allow concatenating Bytecode with verbatim bytes.
This implementation overload the __add__ and make the basic use case work.
However, this is not very consistent. E.g.

• code + b"\x00" means something different than code + Bytecode(b"\x00")
• Bytecode() + b"\x00" works in places where Bytecode() is expected but b"\x00" don't.

The reason this is not a complete solution is that Bytecode() already has a constructor from bytes but its purpose is different and unknown to me. We should consider changing the semantic of conversion from bytes to Bytecode() by just copying the bytes.

There is one more option for handling truncated opcode immediate bytes: take verbatim bytes from [], e.g. Op.PUSH3[b"\x00"]. This currently compiles, but Bytecode also tries to be smart here and extends the provided bytes to the desired length. I'm not 100% sure what it does but I couldn't built a code with truncated immediates this way. However, there is one example that seems to be working:

execution-spec-tests/tests/osaka/eip7692_eof_v1/eip4200_relative_jumps/test_rjumpv.py

Line 273 in ca91240

data=Container.Code(code=Op.PUSH1(1) + Op.RJUMPV[rjumpv_bytes]),

🔗 Related Issues

✅ Checklist

• All: Set appropriate labels for the changes.
• All: Considered squashing commits to improve commit history.
• All: Added an entry to CHANGELOG.md.
• All: Considered updating the online docs in the ./docs/ directory.
• Tests: All converted JSON/YML tests from ethereum/tests have been added to converted-ethereum-tests.txt.
• Tests: A PR with removal of converted JSON/YML tests from ethereum/tests have been opened.
• Tests: Included the type and version of evm t8n tool used to locally execute test cases: e.g., ref with commit hash or geth 1.13.1-stable-3f40e65.
• Tests: Ran mkdocs serve locally and verified the auto-generated docs for new tests in the Test Case Reference are correctly formatted.

[marioevz]

LGTM. I pushed some minor fixes and will merge once the CI passes.

[marioevz]

Regarding the inconsistencies, this feature should be handled with care anyway, and I think we should only use it when absolutely necessary, so I will keep an eye on usages during reviews to make sure it doesn't get out of control :P

[pdobacz]

I had some late comments

[pdobacz]

to address @marioevz 's concerns, maybe this should be Op.RJUMP + verbatim(b"\x00"), and verbatim produces some type which can be added to Bytecode but isn't bytes explicitly, possibly Bytecode would be fine? This would make spotting abuse much easier.

Basically verbatim could be what @chfast envisioned for the constructor Bytecode.