From 6a0ca027c12ffc9f4d0485faab840356e36b9696 Mon Sep 17 00:00:00 2001 From: greged93 <82421016+greged93@users.noreply.github.com> Date: Tue, 19 Sep 2023 16:55:14 +0700 Subject: [PATCH] feat: contributor guide (#96) * contributor guide * rename * rephrasing * fix comments * updates based on comments * fix comment typo --- docs/CONTRIBUTING.md | 64 ++++++++++++++++++++++++++++++++++++++++++++ 1 file changed, 64 insertions(+) create mode 100644 docs/CONTRIBUTING.md diff --git a/docs/CONTRIBUTING.md b/docs/CONTRIBUTING.md new file mode 100644 index 00000000..0835f416 --- /dev/null +++ b/docs/CONTRIBUTING.md @@ -0,0 +1,64 @@ +# Contributing + +Before contributing to this repository, always submit and discuss your proposed changes through an issue. Should you need further clarification, you can reach out to us on Telegram, or Discord. Please wait to be assigned to an issue before opening a pull request. + +## Roles + +Three roles are available for contributors and will lead to different tasks in the repository: + +- Integrator: adds a test suite, runs the suite, makes an issue for each failing test, marks failing tests as skipped. +- Debuggor: picks an issue raised by a Integrator, and dives in the EVM spec and the current +Cairo 0 implementation in order to pin point the exact issue. Ones the issue is correctly understood, checks if a similar issue hasn't been raised by another Debuggor yet. If it has, link the Integrator issue to the Debuggor issue for tracking and pick up another Integrator issue. If it hasn't, raises a detailed issue about the Cairo 0 implementation bug and links it to the Integrator issue. +- Implementor: picks up an detailed issue raised by a Debuggor and reimplements the test logic in Cairo 0 python tests. Once the error is replicated, fixes the test in Cairo 0. + +## How to + +### Integrator + +In order to match the current EVM implementation from Reth, we want to run the same ethereum/tests as them. A list of all applicable tests can be found [here](https://github.com/paradigmxyz/reth/blob/main/testing/ef-tests/tests/tests.rs#L17). To start: + +- Check for an unassigned `epic` issue for a test. +- Add it to `crates/ef-testing/tests.rs` by using the `blockchain_tests` macro. The first argument should be the name of your test, the second is the folder to find the test. Please use the snake case name of the folder for the name of the test (e.g. `blockchain_tests!(st_bad_opcode, stBadOpcode)` for adding stBadOpcode). +- Comment all the other lines except for the test you added and run `make ef-tests`. +After a while (count 10-15 minutes) the test will end and you should have in your terminal's output a list of all the failing tests. +- Start making issues. When raising issues, please try to match other `Integrator` issued on the style, and don't forget to add any raised issue to the `epic` issue's task list. + +Please find [here](https://github.com/kkrt-labs/ef-tests/issues/52) an example of an issue raised by an Integrator. + +### Debuggor + +Pick an issue from the available `Integrator` issues. Verify that the test fails (you can run a specific test by using `make target=your_test_name ef-test`). If it does, you can start debugging it. The following documentation can be used: + +- [Test fillers](https://github.com/ethereum/tests/tree/develop/src/GeneralStateTestsFiller): used to generated the actual test. This can be used to understand what code the transaction executes. +- [Execution specs](https://github.com/ethereum/execution-specs/tree/master): Ethereum execution specifications written in Python. This can be used to compare the expected behavior to the [Cairo 0 implementation](https://github.com/kkrt-labs/kakarot/tree/main/src). +- [Yellow paper](https://ethereum.github.io/yellowpaper/paper.pdf): Ethereum formal specification. This can be used to compare the expected behavior to the [Cairo 0 implementation](https://github.com/kkrt-labs/kakarot/tree/main/src). + +Please find [here](https://github.com/kkrt-labs/ef-tests/issues/57) an example of an issue raised by a Debuggor. + +#### Log opcodes + +The following section describes how to log the opcodes when running an ethereum/test: + +- Clone `https://github.com/kkrt-labs/kakarot-rpc.git` +- Update the `Cargo.toml` file at the root of the repository, by replacing the blockifier import by `blockifier = { git = "https://github.com/jobez/blockifier.git", rev = "7f00407" }`. +- In the file `./kakarot-rpc/lib/kakarot/src/kakarot/instructions.cairo`, add the following hint `%{print(ids.opcode, ids.ctx) %}` right after line 65. See the following [gist](https://gist.github.com/jobez/42941db9361d81778abd36309dfb60dc#file-instructions-cairo-L68-L70) for more details. +- From the root of the `kakarot-rpc` repository, run the following: `cd lib/kakarot && STARKNET_NETWORK=starknet-devnet make build && cd ../.. && make RUST_BACKTRACE=1 dump-katana` +- Now copy the `.katana` folder and paste it inside the ef-tests repository. +- Finally, update the `Cargo.toml` file at the root of the ef-tests repository, by replacing the blockifier import by `blockifier = { git = "https://github.com/jobez/blockifier.git", rev = "7f00407" }`. +- Run `make target=your_test_name ef-test`. You should see the executed opcodes being printed out. + +#### Reproducing in Python (to be updated once Python test flow is improved) + +In case the above doesn't help you with debugging, you can always reproduce the test case you are fixing in Python. This will give you unlimited hint usage, allowing for better tracing of the error. In order to set this up: + +- Clone `https://github.com/kkrt-labs/kakarot.git` +- Run `make setup && make build` +- Create a new file in `tests/integration/solidity_contracts/EFTests` if there is no file corresponding to the scope you want to test. Otherwise you can add your test to an existing file. +- The following fixtures can be used in order to set the test environment to the exact replica of the ethereum/test: `create_account_with_bytecode_and_storage`, `set_storage_at_evm_address`, `deploy_eoa`,... You can have a look at the current tests in the `EfTests` folder for help. +- Call `eth_send_transaction` with the exact same arguments as the transaction from your test. If the error you get is the same as the error on the ef-tests repository, you can now start debugging using hints in the Cairo 0 code. +- In order to run only your test, use `poetry run pytest -k your_test_name`. +- Debug the Cairo 0 code by adding `%{print(ids.x)%}` to print variables (where x is the variable you want to print). Be sure to recompile your Cairo 0 code EVERY TIME you make changes, by running `make build`. + +### Implementor + +Coming soon