Add reversed iterator type.#2900
Conversation
DimitrisJim
force-pushed
the
reversed_iter_type
branch
from
aa3833b to
ca6ab79
Compare
DimitrisJim
force-pushed
the
reversed_iter_type
branch
from
00445b5 to
fe4c3a5
Compare
| use crate::iterator; | ||
| use crate::slots::PyIter; | ||
| use crate::vm::VirtualMachine; | ||
| use crate::{iterator, ItemProtocol, TypeProtocol}; |
There was a problem hiding this comment.
same level as the next line (Line 13)
|
|
||
| #[pyclass(module = false, name = "reversed")] | ||
| #[derive(Debug)] | ||
| pub struct PyReverseSequenceIterator { |
There was a problem hiding this comment.
is this belong to enuemerate rather than sequence? I feel like this iterator generate (int, object) rather than object due to the placed module.
There was a problem hiding this comment.
I know. I don't know why CPython has them in the same file but it does. So I kept the same structure.
| #[pyclass(module = false, name = "reversed")] | ||
| #[derive(Debug)] | ||
| pub struct PyReverseSequenceIterator { | ||
| pub position: AtomicCell<isize>, |
There was a problem hiding this comment.
is this isize rather than a usize?
There was a problem hiding this comment.
Usually the reverse iterators keep an isize in order to check that its > 0 while decreasing it and indexing the underlying object.
|
|
||
| impl PyIter for PyReverseSequenceIterator { | ||
| fn next(zelf: &PyRef<Self>, vm: &VirtualMachine) -> PyResult { | ||
| let pos = zelf.position.fetch_sub(1); |
There was a problem hiding this comment.
what happens if this is called forever until negatively overflow?
There was a problem hiding this comment.
I'm fixing all these methods in another pr. Currently I've just split these appart for now to make the changes a little smaller in size.
There was a problem hiding this comment.
Will you make another huge PR with the whole diff for a reference? I will also check the final result.
There was a problem hiding this comment.
I could push the changes on this branch if it would make it easier for you.
There was a problem hiding this comment.
I think seperated small PR still have benefits. So another PR would be more convenient. It will be reusable after merging this one.
For this PR, I thought I can compare other parts also, but after checking your answers about it, it doesn't look that's nessessary for now.
2 participants
PySequenceIteratorhandles forward iteration whilePyReverseSequenceIterator, located inenumerate.rs(not sure why) as in CPython, handles reverse iteration.