Remove chia-blockchain dependency for most of chia_rs #887
Conversation
matt-o-how
changed the title
There was a problem hiding this comment.
this is a really nice change. However, I think it would be good to check the correctness of this against the existing python implementation. One way to do that would be to break this off into a separate PR that still relies on the chia-blockchain dependency to compare the result between the rust and python implementation. Can you think of another way to be confident in the correctness of this port?
Pull Request Test Coverage Report for Build 12992847296Warning: This coverage report may be inaccurate.This pull request's base commit is no longer the HEAD commit of its target branch. This means it includes changes from outside the original pull request, including, potentially, unrelated coverage changes.
Details
💛 - Coveralls |
| @@ -4,7 +4,9 @@ | |||
| pub const UI_ACTUAL_SPACE_CONSTANT_FACTOR: f32 = 0.78; | |||
|
|
|||
| // TODO: Update this when new plot format releases | |||
| pub fn expected_plot_size(k: u32) -> u64 { | |||
| #[cfg(feature = "py-bindings")] | |||
| #[pyo3::pyfunction] | |||
There was a problem hiding this comment.
I think you have to do #[cfg_attr(feature = "py-bindings", pyo3::pyfunction)] instead?
| fn sp_total_iters_impl(&self, py: Python<'_>, constants: &Bound<'_, PyAny>) -> PyResult<u128> { | ||
| self.sp_sub_slot_total_iters_impl(py, constants)? | ||
| .checked_add(self.sp_iters_impl(py, constants)? as u128) | ||
| fn sp_iters_impl(&self, constants: &Bound<'_, PyAny>) -> PyResult<u64> { |
There was a problem hiding this comment.
now that these are implemented in rust, I don't think they should be limited to just the python bindings anymore. I imagine you put them here to simplify error handling, so make them return PyResult<>. If you make them part of the rust type and make them return chia_error::Result<> you could map those errors with map_err()
|
|
||
| #[cfg(feature = "py-bindings")] | ||
| #[pyo3::pyfunction] | ||
| pub fn calculate_ip_iters( |
There was a problem hiding this comment.
do you feel like the unit tests below are sufficient to be confident that this is correct and identical to the python implementation?
I'm especially worried about integer width overflow
| num_sps_sub_slot: int, | ||
| num_sp_intervals_extra: int, | ||
| sub_slot_iters: int, | ||
| signage_point_index: int, | ||
| required_iters: int, | ||
| ) -> int: ... |
There was a problem hiding this comment.
are all of these plain int in the python implementation? it seems like it might make sense to make them sized ints
| ) | ||
| assert sp_iters > ip_iters | ||
|
|
||
| # def test_win_percentage(self): |
There was a problem hiding this comment.
I think this commented-out code can be removed, right?
|
|
||
| br = get_block_record(rng, ssi=ssi, spi=13, ri=1) | ||
| res = br.ip_iters_impl(DEFAULT_CONSTANTS) | ||
| assert res is not None |
There was a problem hiding this comment.
not None, it seems like the wrong value would also pass here
This PR removes imports from
test_blscache.pyand deletestest_program_fidelity.py.It also removes the python running from
block_record.rsby implementingpot_iterations.pyinside the class