perf: optimize compiler analysis phase#17823
Merged
Rich-Harris merged 2 commits intosveltejs:mainfrom Feb 27, 2026
Merged
Conversation
Cache ignore_stack snapshots to avoid structuredClone on every AST node, and restructure CSS pruning to walk the stylesheet once instead of once per element. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
🦋 Changeset detectedLatest commit: 41dbcc2 The changes in this PR will be included in the next version bump. This PR includes changesets to release 1 package
Not sure what this means? Click here to learn what changesets are. Click here if you're a maintainer who wants to add another changeset to this PR |
Optimize the compiler analysis phase for better performance.
Merged
Rich-Harris
pushed a commit
that referenced
this pull request
Feb 27, 2026
This PR was opened by the [Changesets release](https://github.com/changesets/action) GitHub action. When you're ready to do a release, you can merge this and the packages will be published to npm automatically. If you're not ready to do a release yet, that's fine, whenever you add more changesets to main, this PR will be updated. # Releases ## svelte@5.53.6 ### Patch Changes - perf: optimize parser hot paths for faster compilation ([#17811](#17811)) - fix: `SvelteMap` incorrectly handles keys with `undefined` values ([#17826](#17826)) - fix: SvelteURL `search` setter now returns the normalized value, matching native URL behavior ([#17828](#17828)) - fix: visit synthetic value node during ssr ([#17824](#17824)) - fix: always case insensitive event handlers during ssr ([#17822](#17822)) - chore: more efficient effect scheduling ([#17808](#17808)) - perf: optimize compiler analysis phase ([#17823](#17823)) - fix: skip redundant batch.apply ([#17816](#17816)) - chore: null out current_batch before committing branches ([#17809](#17809)) Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com>
This file contains hidden or 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
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.
Summary
Two optimizations to the compiler's analysis phase:
Cache
ignore_stacksnapshots instead ofstructuredCloneon every node. The universal_visitor in the analysis walk runs on every AST node and callsstructuredClone(ignore_stack)each time. In practice,svelte-ignorecomments are rare (0–5 per component), so 99%+ of nodes deep-clone an unchanged stack. This adds a copy-on-write cache that only re-creates the snapshot whenpush_ignore/pop_ignoreactually change the stack.Walk the CSS stylesheet once instead of once per element.
prune()was called in a loop for each element, each time doing a fullwalk()of the stylesheet AST. This restructures the loop so the stylesheet is walked once, and the element iteration happens inside theComplexSelectorvisitor.Benchmarks
Compiled each component 500 times (after 50 warmup iterations), measuring average time per
compile()call:has(80+ CSS selectors, 12 elements)siblings-combinator-each-nested(65 CSS rules, 15 elements)The CSS pruning optimization scales with
elements × CSS rules— the more elements a component has, the bigger the win since we go from N stylesheet walks down to 1. ThestructuredClonefix helps every component regardless of CSS, eliminating ~500–2000 deep clones per compile (one per AST node) and replacing them with 0–5 (one persvelte-ignorecomment).For typical real-world components with 10–20 elements and some CSS, expect roughly 20–30% faster compilation in the analysis phase.
Test plan
svelte-ignorebehavior unchanged (snapshot is consumed read-only via.has()/.some())🤖 Generated with Claude Code