Write path normalization without array allocations #60812
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
typescript-bot
added
Author: Team
For Uncommitted Bug
|
@typescript-bot perf test this |
| @@ -624,27 +624,103 @@ export function getNormalizedPathComponents(path: string, currentDirectory: stri | |||
| } | |||
|
|
|||
| /** @internal */ | |||
| export function getNormalizedAbsolutePath(fileName: string, currentDirectory: string | undefined): string { | |||
| return getPathFromPathComponents(getNormalizedPathComponents(fileName, currentDirectory)); | |||
| export function getNormalizedAbsolutePath(path: string, currentDirectory: string | undefined): string { | |||
There was a problem hiding this comment.
I wonder if currentDirectory is almost always normalized.
|
@andrewbranch Here they are:
tscComparison Report - baseline..pr
System info unknown
Hosts
Scenarios
tsserverComparison Report - baseline..pr
System info unknown
Hosts
Scenarios
startupComparison Report - baseline..pr
System info unknown
Hosts
Scenarios
Developer Information: |
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
src/compiler/path.ts
Outdated
Comment on lines
633
to
636
| const root = path.substring(0, rootLength); | ||
| const normalizedRoot = root && normalizeSlashes(root); | ||
| // `normalized` is only initialized once `path` is determined to be non-normalized | ||
| let normalized = normalizedRoot === root ? undefined : normalizedRoot; |
There was a problem hiding this comment.
You should just use lastIndexOf to see if altDirectorySeparator occurs. That way you can avoid creating a substring for root at all.
|
I’m puzzled by the perf results |
|
Changing the scenarios to have long paths that are already normalized shows the old |
andrewbranch
changed the title
|
@typescript-bot perf test |
src/compiler/path.ts
Outdated
Comment on lines
661
to
662
| const sepIndex = path.indexOf(directorySeparator, index + 1); | ||
| const altSepIndex = path.indexOf(altDirectorySeparator, index + 1); |
There was a problem hiding this comment.
This is quadratic because you have to find at least one of the other kinds of slashes over and over. You should just tight-loop on !isAnyDirectorySeparator.
There was a problem hiding this comment.
or use findIndex(path, isAnyDirectorySeparator, index + 1)
There was a problem hiding this comment.
My only concern with findIndex is that it might not optimize well it passed in both arrays and strings. Maybe I'm superstitious.
src/compiler/path.ts
Outdated
| // At beginning of segment | ||
| segmentStart = index; | ||
| let ch = path.charCodeAt(index); | ||
| while (isAnyDirectorySeparator(ch) && index + 1 < path.length) { |
There was a problem hiding this comment.
Took me a bit to see why this works, but it's because index starts immediately after the root, or after the first separator following the prior segment. Figuring out if one of those is a backslash is handled below.
| normalizedUpTo = index + 2; | ||
| } | ||
| } | ||
| else if (normalized === undefined) { |
There was a problem hiding this comment.
I think that in this branch it is technically possible for you to avoid extra substrings by adjusting normalizedUpTo instead of initializing normalized.
|
@andrewbranch Here they are:
tscComparison Report - baseline..pr
System info unknown
Hosts
Scenarios
tsserverComparison Report - baseline..pr
System info unknown
Hosts
Scenarios
startupComparison Report - baseline..pr
System info unknown
Hosts
Scenarios
Developer Information: |
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| // the root is `file://Users/` | ||
| assert.strictEqual(ts.getNormalizedAbsolutePath("file://Users/matb/projects/san/../../../../../../typings/@epic/Core.d.ts", ""), "file://Users/typings/@epic/Core.d.ts"); | ||
| // this is real | ||
| assert.strictEqual(ts.getNormalizedAbsolutePath("file:///Users/matb/projects/san/../../../../../../typings/@epic/Core.d.ts", ""), "file:///typings/@epic/Core.d.ts"); |
There was a problem hiding this comment.
idk if we want to preserve specific examples in these, but that might be fine.
Maybe we should have tests on untitled:?
src/compiler/path.ts
Outdated
| return path; | ||
| } | ||
| // Some paths only require cleanup of `/./` or leading `./` | ||
| const simplified = path.replace(/\/\.\//g, "/").replace(/^\.\//, ""); |
There was a problem hiding this comment.
If we're already here, I'd also recommend
Suggested change
| const simplified = path.replace(/\/\.\//g, "/").replace(/^\.\//, ""); | |
| let simplified = path.replace(/\/\.\//g, "/"); | |
| if (simplified.startsWith("./") { | |
| simplified = simplified.slice(2); | |
| } |
There was a problem hiding this comment.
This is indeed a small but measurable improvement. Nice!
rbuckton
requested changes
Dec 19, 2024
| normalized = path.substring(0, normalizedUpTo); | ||
| } | ||
| } | ||
| else if (segmentLength === 2 && path.charCodeAt(index) === CharacterCodes.dot && path.charCodeAt(index + 1) === CharacterCodes.dot) { |
There was a problem hiding this comment.
Minor nit: If the previous segmentLength was 1, but the segment was not ., this performs an unnecessary recheck of segmentLength === 2.
src/compiler/path.ts
Outdated
Comment on lines
661
to
662
| const sepIndex = path.indexOf(directorySeparator, index + 1); | ||
| const altSepIndex = path.indexOf(altDirectorySeparator, index + 1); |
There was a problem hiding this comment.
or use findIndex(path, isAnyDirectorySeparator, index + 1)
|
From instrumenting the compilation of VS Code, I learned two things:
Optimizing I'm going to repeat the instrumentation for a TS Server program update and see if the ratios are similar. |
|
During a TS Server update, literally every call is a no-op. |
|
@typescript-bot pack this |
|
Hey @DanielRosenwasser, something went wrong when looking for the build artifact. (You can check the log here). |
|
@typescript-bot pack this |
|
Hey @DanielRosenwasser, I've packed this into an installable tgz. You can install it for testing by referencing it in your and then running There is also a playground for this build and an npm module you can use via |
|
|
After latest commit: and just comparing the last commit to the one before it: |
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.
Related: #60633
Alternative to #60755
Despite the good benchmarks, I didn’t see any measurable impact in real-world
updateGraphWorkerduration on my M2 Mac Mini. However, since that was a GC hot spot for @dbaeumer and #60755 improved things some, this may do even better for him.bench.js