Implement nth, count, and last for EscapeUnicode
#31049
Conversation
|
(rust_highfive has picked a reviewer for you, use r? to override) |
|
AFAICT this should allow the last tick missing in #30520 to be set. Some care has been taken in order to avoid code duplication (between WARNING: this PR marks Additional tests are missing; I will take care of this ASAP. |
| Done, | ||
| RightBrace, |
There was a problem hiding this comment.
This is just a micro-optimization (so that state_len is a single load instead of load+add).
I can make it more explicit and/or reference the comment in state_len or I can remove it if it is deemed undesirable.
The `offset` value was computed both in `next` and in `size_hint`; computing it in a single place ensures consistency and makes it easier to apply improvements. The value is now computed as soon as the iterator is constructed. This means that the time to compute it is spent immediately and cannot be avoided, but it also guarantees that it is only spent once.
Instead of iteratively scanning the bits, use `leading_zeros`.
In rust-lang#28662, `size_hint` was made exact for `EscapeUnicode` and `EscapeDefault`, but neither was marked as `ExactSizeIterator`.
Trivial implementation, as both are `ExactSizeIterator`s. Part of rust-lang#24214.
and reuse it in `size_hint`.
Extract a function that updates the iterator state and returns the result of an arbitrary step of iteration. This implements the same logic as `next`, but it can be shared with `nth`.
as a step from the appropriate state. Part of rust-lang#24214.
This makes it easier to have a unique path for handling all of them.
by extracting a shared `step` function.
ranma42
force-pushed
the
escape-iters-take3
branch
from
4cc6093 to
266da8e
Compare
|
This seems pretty complicated for a ... pretty minor iterator, out of curiosity was this motivated beyond ensuring that the implementations are specialized for all iterators? |
|
It is basically a continuation of #30624, mainly to finish #24214. There are several changes because the PR includes some improvements to the current state of Single commits should be easier to understand, but I can also split it into multiple PRs if it is more convenient. EDIT: wrong ref |
|
Abandoned in favour of (multiple) simpler PRs. |
Improve computation of offset in `EscapeUnicode` Unify the computation of `offset` and use `leading_zeros` instead of manually scanning the bits. This PR removes some duplicated code and makes it a little simpler . The computation of `offset` is also faster, but it is unlikely to have an impact on actual code. (split from #31049)
Implement `last` for `EscapeUnicode`
The implementation is quite trivial as the last character is always `'{'`.
As a side-effect it also improves the implementation of `last` for `EscapeUnicode`.
Part of #24214, split from #31049.
Maybe this (and the other changes that I will split from #31049) should wait for a test like `ed_iterator_specializations` to be added. Would it be sufficient to do the same for each possible escape length?
…richton Implement `count` for `EscapeUnicode` and cleanup the code for `count` for `EscapeDefault` (instead of repeating the `match` for `size_hint` and `count`). This PR marks EscapeUnicode and EscapeDefault as ExactSizeIterator. The constraints for the trait implementations held even before this PR, but I am not sure if this is something we want to guarantee/expose (I would love feedback on this, especially on what would be the appropriate way to handle stabilisation, if needed). Part of rust-lang#24214, split from rust-lang#31049. The test for `count` was added in rust-lang#33103.
and align the implementations in
EscapeUnicodeandEscapeDefaultPart of #24214.
EDIT: corrected reference to issue.