Unwrapping output iterators in range algorithms #5015
Merged
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
CaseyCarter
force-pushed
the
ranges-output-unwrapping
branch
from
cd75c40 to
fd21582
Compare
* Add `_Distance_unbounded` to represent the distance between an iterator and `unreachable_sentinel`. * `_Idl_distance` takes iterator/sentinel pairs in C++20 mode. * Add `_RANGES _Idl_distance(range)`, equivalent to `_STD _Idl_distance` but for `sized_range` instead of `sized_sentinel_for`. * Support integer-class counts in `_Get_unwrapped_n`.
Now that `_Get_unwrapped_n` supports integer-class counts, we don't need to apply `_Algorithm_int_t` to `_Count`.
* Refactor `_Idl_dist_add`: * Use `if constexpr` dispatch instead of overloading. * `_Distance_unbounded` plus anything is `_Distance_unbounded`. * Implement `_Idl_dist_min` to compute the smaller of two `_Idl_distance`s.
…_Output` This changes our convention. I think `_Output` is more appropriate than `_Result` for this purpose given that most algorithms produce more results than just the output sequence. This also makes `_Result` unambiguously available to refer to the return value of the algorithm which is consistent with our use of `_UResult` for intermediate unwrapped results.
These tests failed because an iterator claims to be random-access without supporting subtraction, which breaks with our pre-C++20 `_Idl_distance`. C++20-and-later `_Idl_distance` isn't fooled by such shenanigans. (We only test libc++ in `/std:c++latest` mode.)
CaseyCarter
force-pushed
the
ranges-output-unwrapping
branch
from
fd21582 to
0b2139b
Compare
StephanTLavavej
added
performance
StephanTLavavej
requested changes
Oct 14, 2024
StephanTLavavej
approved these changes
Oct 15, 2024
|
I'm mirroring this to the MSVC-internal repo - please notify me if any further changes are pushed. |
|
Thanks for this long-awaited improvement! 😻 🪄 🎉 |
CaseyCarter
added a commit
to CaseyCarter/STL
that referenced
this pull request
Oct 18, 2024
microsoft#5015 unskipped these, they now go through the same old codepath.
CaseyCarter
added a commit
to CaseyCarter/STL
that referenced
this pull request
Oct 27, 2024
…gative counts * `ranges::copy_n` takes a difference argument that it was blindly converting to `size_t` Drive-by: `_Copy_n_unchecked4` no longer needs to guard its `_STL_INTERNAL_STATIC_ASSERT(_Integer_like<_SizeTy>);` for `_HAS_CXX20` since microsoft#5015 defined `_Integer_like` in pre-20 modes.
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
This completes the last bits of the iterator unwrapping design for Ranges algorithms.
Detailed commit-wise description:
Expand
_Idl_distancedesign for Ranges_Distance_unboundedto represent the distance between an iterator andunreachable_sentinel._Idl_distancetakes iterator/sentinel pairs in C++20 mode._RANGES _Idl_distance(range), equivalent to_STD _Idl_distancebut forsized_rangeinstead ofsized_sentinel_for._Get_unwrapped_n.Simplify
ranges::destroy_n. Now that_Get_unwrapped_nsupports integer-class counts, we don't need to apply_Algorithm_int_tto_Count.Update operations on
_Idl_distances_Idl_dist_add:if constexprdispatch instead of overloading._Distance_unboundedplus anything is_Distance_unbounded._Idl_dist_minto compute the smaller of two_Idl_distances.Rename Range algorithm output iterator parameters from
_Resultto_Output. This changes our convention. I think_Outputis more appropriate than_Resultfor this purpose given that most algorithms produce more results than just the output sequence. This also makes_Resultunambiguously available to refer to the return value of the algorithm which is consistent with our use of_UResultfor intermediate unwrapped results.Unwrap outputs in
ranges::copyUnwrap outputs in
ranges::copy_backwardUnwrap outputs in
ranges::copy_ifUnwrap outputs in
ranges::copy_nUnwrap outputs in
ranges::mergeUnwrap outputs in
ranges::moveUnwrap outputs in
ranges::move_backwardUnwrap outputs in
ranges::replace_copyUnwrap outputs in
ranges::replace_copy_ifUnwrap outputs in
ranges::reverse_copyUnwrap outputs in
ranges::rotate_copyUnwrap outputs in
ranges::sampleUnwrap outputs in
ranges::transform(The warning suppression is necessary to prevent the static analyzer diagnosing the computation of_Min_size_determinableas redundant when_Rng1and_Rng2are the same type.)Fixes #893.