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Outdated suggestions cannot be applied.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.
#92 introduced quite a serious performance regression by refactoring inlinable code in to a helper function which was only usable from inlined code, but not itself inlinable. The first commit corrects this.
Upon further inspection, I noticed a pattern - that all internal functions seemed to only be
@usableFromInline
and not@inlinable
, whereas their public siblings were all@inlinable
. Inlining is critical for exactly this kind of code - lazy wrappers and collections and so on, and especially when they are generic. For instance, how is the client's instance of the compiler supposed to be able to optimiseChain
, when the implementation ofoffsetForward
is hidden from it?So the second commit inlines basically everything that was previously
@usableFromInline
.Additionally, I would strongly recommend setting up a benchmark suite, so that small refactorings don't cause huge performance regressions in future. I have some concerns about adding
swift-algorithms
as a dependency for my performance-sensitive project, and a benchmark suite which ensures at least stable baseline performance would help convince me that it's safe to rely on this project.