Skip to content

perf(start): construct compiler-owned server function AST directly - #8524

Open
Sheraff wants to merge 1 commit into
codex/fuse-clone-provenancefrom
codex/direct-compiler-ast
Open

Sheraff wants to merge 1 commit into
codex/fuse-clone-provenancefrom
codex/direct-compiler-ast

Conversation

@Sheraff

@Sheraff Sheraff commented Sep 27, 2026 •

Copy link
Copy Markdown
Collaborator

🎯 Changes

Stacks on #8523; base branch: codex/fuse-clone-provenance.

Server-function compilation currently reparses compiler-owned JavaScript snippets for RPC calls, provider wrappers, imports, directives, and HMR guards. Construct those nodes directly with typed AST builders, preserving generated-reference ownership, source maps, and emitted code.

Compound builders live in named functions with readable inline snapshots of their generated JavaScript. Additional execution tests cover client/SSR/provider output across React, Solid, and Vue, escaping, binding cleanup, HMR, and invalidation.

Measured performance

Complete generated React Start production builds, with 10 server functions per module and seven alternating fresh-process pairs per size:

Server functions #8523 median This change median Reduction Peak RSS, before → after
100 333.07 ms 326.33 ms Within noise 329.8 → 324.2 MiB
1,000 886.94 ms 813.09 ms 8.3% 519.9 → 483.2 MiB
5,000 6,682.98 ms 6,048.85 ms 9.5% 883.8 → 849.5 MiB

The candidate wins all seven pairs at 1,000 and 5,000 functions. The 5,000-function fixture is a stress case. Timings include Vite builder creation, transformation, bundling, minification, and writes; they exclude fixture generation, plugin-module loading, and output validation. All RPCs and handlers are retained; client output is byte-identical and server output matches after normalizing build paths and chunk filenames.

Focused compiler measurements show 15.0% faster client transforms, 19.9% faster SSR callers, 27.7% faster providers, and 29.3% faster development providers. The existing smaller e2e app's whole-build result was inconclusive (950.26 → 962.54 ms), so these gains should not be generalized to every app.

Measured on Node 24.8.0 / Apple M3 Max against #8523 at 6c228b37ad, with the extracted candidate on 9148909cc5. The current stack base 3dfc160aee only adds Solid CSS test changes. Full methodology, raw results, and attribution.

Previously completed validation

  • 608 compiler tests and 165 browser tests; TypeScript 5.6–7, ESLint, and package build/export checks passed.
  • The latest inline-snapshot-only edit passed all eight builder tests.
  • All 18 bundle scenarios have zero size/chunk-count changes; all 38 emitted client JS files are byte-identical.
  • No checks were rerun when opening this PR; CI will validate the published branch.

✅ Checklist

  • I have followed the steps in the Contributing guide.
  • I have tested code changes locally with the relevant test commands, or tests do not apply to this pull request.
  • I fully understand the code in this pull request, including any code generated with AI assistance.

🚀 Release Impact

  • This change affects published code, and I have generated a changeset.
  • This change is docs/CI/dev-only (no release).

Summary by CodeRabbit

  • Performance
    • Server-function compilation and production builds now run faster in several benchmarked workloads, while preserving equivalent generated output.
  • Reliability
    • Added checks for generated server-function output across React, Solid, and Vue, including client, server-rendering, provider, and development scenarios.
  • Documentation
    • Added guidance and results for compiler and production-build benchmarks.

@coderabbitai

coderabbitai Bot commented Sep 27, 2026 •

Copy link
Copy Markdown
Contributor

Review in Change Stack →

Navigate logical layers of code changes, visualize relationships, and explore their blast radius.

No actionable comments were generated in the recent review. 🎉

ℹ️ Recent review info
⚙️ Run configuration

Configuration used: Repository: TanStack/router/.coderabbit.yaml

Review profile: CHILL

Plan: Advanced

Run ID: 605f17db-53a0-4dec-b85e-2831bde04fcc

📥 Commits

Reviewing files that changed from the base of the PR and between 080a540 and c6c33e2.

📒 Files selected for processing (43)
  • .changeset/fuzzy-berries-behave.md
  • benchmarks/compiler/README.md
  • benchmarks/compiler/direct-ast.md
  • benchmarks/compiler/results/direct-ast-app-output.json
  • benchmarks/compiler/results/direct-ast-builders-digests.json
  • benchmarks/compiler/results/direct-ast-builders-timings.json
  • benchmarks/compiler/results/direct-ast-builds.json
  • benchmarks/compiler/results/direct-ast-bundle-baseline-hashes.json
  • benchmarks/compiler/results/direct-ast-bundle-baseline.json
  • benchmarks/compiler/results/direct-ast-bundle-builders-final-hashes.json
  • benchmarks/compiler/results/direct-ast-bundle-builders-final.json
  • benchmarks/compiler/results/direct-ast-bundle-final-hashes.json
  • benchmarks/compiler/results/direct-ast-bundle-final.json
  • benchmarks/compiler/results/direct-ast-bundle-pre-attribution.json
  • benchmarks/compiler/results/direct-ast-caller-digests.json
  • benchmarks/compiler/results/direct-ast-caller-timings.json
  • benchmarks/compiler/results/direct-ast-dense-refined-timings.json
  • benchmarks/compiler/results/direct-ast-digests.json
  • benchmarks/compiler/results/direct-ast-directives-digests.json
  • benchmarks/compiler/results/direct-ast-directives-refined-timings.json
  • benchmarks/compiler/results/direct-ast-directives-timings.json
  • benchmarks/compiler/results/direct-ast-final-digests.json
  • benchmarks/compiler/results/direct-ast-final-timings.json
  • benchmarks/compiler/results/direct-ast-final-verification.json
  • benchmarks/compiler/results/direct-ast-heavy-builds-100.json
  • benchmarks/compiler/results/direct-ast-heavy-builds-1000.json
  • benchmarks/compiler/results/direct-ast-heavy-builds-5000.json
  • benchmarks/compiler/results/direct-ast-hmr-digests.json
  • benchmarks/compiler/results/direct-ast-hmr-refined-timings.json
  • benchmarks/compiler/results/direct-ast-hmr-timings.json
  • benchmarks/compiler/results/direct-ast-hunk-bundles.json
  • benchmarks/compiler/results/direct-ast-imports-digests.json
  • benchmarks/compiler/results/direct-ast-imports-refined-timings.json
  • benchmarks/compiler/results/direct-ast-imports-timings.json
  • benchmarks/compiler/results/direct-ast-provider-digests.json
  • benchmarks/compiler/results/direct-ast-provider-timings.json
  • benchmarks/compiler/results/direct-ast-timings.json
  • benchmarks/compiler/server-fn-ast.mjs
  • benchmarks/compiler/server-fn-build.mjs
  • packages/start-plugin-core/src/start-compiler/handleCreateServerFn.ts
  • packages/start-plugin-core/src/start-compiler/serverFnAst.ts
  • packages/start-plugin-core/tests/createServerFn/generated-ast.test.ts
  • packages/start-plugin-core/tests/createServerFn/serverFnAst.test.ts

Included review availability: This review used your included allowance. Your plan provides up to 10 included reviews per hour; 5 remain after this review.


📝 Walkthrough

Walkthrough

The server-function compiler now builds generated RPC calls, provider wrappers, imports, directives, exports, and HMR statements with AST helpers. New compiler and production-build benchmark harnesses record timing, memory, output comparisons, and bundle hashes. Tests cover generated output across frameworks and runtimes.

Changes

Server-function AST generation and benchmarking

Layer / File(s) Summary
Build and integrate server-function AST nodes
packages/start-plugin-core/src/start-compiler/serverFnAst.ts, packages/start-plugin-core/src/start-compiler/handleCreateServerFn.ts, packages/start-plugin-core/tests/createServerFn/*ast.test.ts, .changeset/fuzzy-berries-behave.md
AST helpers create server-function callers, providers, exports, HMR checks, imports, references, and string literals. The compiler uses the helpers instead of parsing generated source strings. Tests cover output and recompilation across frameworks and runtimes.
Measure compiler workloads
benchmarks/compiler/server-fn-ast.mjs, benchmarks/compiler/README.md, benchmarks/compiler/direct-ast.md, benchmarks/compiler/results/direct-ast-*.json
The compiler benchmark runs selected workloads in baseline and candidate processes, checks source hashes and optional output digests, and records timing and memory data. Documentation and results describe the workloads, measurement protocols, and comparisons.
Measure generated production builds
benchmarks/compiler/server-fn-build.mjs, benchmarks/compiler/direct-ast.md, benchmarks/compiler/results/direct-ast-app-output.json, benchmarks/compiler/results/direct-ast-builds.json, benchmarks/compiler/results/direct-ast-heavy-builds-*.json, benchmarks/compiler/results/direct-ast-bundle-*-hashes.json, benchmarks/compiler/results/direct-ast-hunk-bundles.json
The build benchmark generates a temporary React Start app, checks client and server outputs, compares normalized hashes, and records build and memory metrics. The report and result files include full-build measurements and bundle comparisons.

Priority: ➖ Normal

Estimated code review effort: 4 (Complex) | ~45 minutes

Merge Risk: ⚪ Minimal · up to c6c33

The server-function AST change has no identified issue that needs resolution before merging; normal checks remain appropriate.

Security Architecture Review

Security architecture risk: 🔵 Low · up to c6c33

The change affects how server-function code is generated, not an identified authorization rule or deployment boundary. Tests and recorded build comparisons support equivalent output, but they do not establish equivalence for every application.

Retained concerns
No architecture-level concerns identified.

Security review details

Security Blast Radius

  • inferred — A generation regression could propagate to applications using server functions across the supported runtimes. The changed test ranges do not themselves add reachable production entrypoints, and the benchmark fanout does not establish production exposure.

Trust Boundaries and Controls

  • inferred — The generated client RPC call and server provider meet at a trust boundary, but the reviewed change supplies no evidence of a new attacker-controlled input path, changed authorization control, or increased server authority. Runtime authorization behavior was not independently verified.
🚥 Pre-merge checks | ✅ 4 | ❌ 1

❌ Failed checks (1 warning)

Check name Status Explanation Resolution
Docstring Coverage ⚠️ Warning Docstring coverage is 9.09% which is insufficient. The required threshold is 80.00%. Docstring coverage is scoped to functions touched by this diff. Analyzed 11 functions across 6 files. (30 skipped: … Write docstrings for the functions missing them to satisfy the coverage threshold.
✅ Passed checks (4 passed)
Check name Status Explanation
Linked Issues check ✅ Passed Check skipped because no linked issues were found for this pull request.
Out of Scope Changes check ✅ Passed Check skipped because no linked issues were found for this pull request.
Title check ✅ Passed The title clearly and concisely describes the main change: constructing compiler-owned server-function AST directly for performance.
Description check ✅ Passed The description follows the required template, explains the motivation and implementation, documents validation and performance results, and completes the checklist and release-impact sections.
Full details: Docstring Coverage

Explanation

Docstring coverage is 9.09% which is insufficient. The required threshold is 80.00%. Docstring coverage is scoped to functions touched by this diff. Analyzed 11 functions across 6 files. (30 skipped: 30 unsupported.)

  • Fix all pre-merge checks with AI
✨ Finishing Touches 💡 1
📝 Generate docstrings 💡
  • Commit to this branch
  • Create a new PR
🧪 Generate unit tests (beta)
  • Commit to this branch
  • Create a new PR

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

Comment @coderabbitai help to get the list of available commands.

@github-actions

github-actions Bot commented Sep 27, 2026 •

Copy link
Copy Markdown
Contributor

🚀 Changeset Version Preview

6 package(s) bumped directly, 7 bumped as dependents.

🟥 Major bumps

Package Version Reason
@tanstack/router-plugin 1.168.41 → 2.0.0 Changeset
@tanstack/router-utils 1.162.3 → 2.0.0 Changeset
@tanstack/start-plugin-core 1.171.48 → 2.0.0 Changeset

🟩 Patch bumps

Package Version Reason
@tanstack/react-router 1.170.40 → 1.170.41 Changeset
@tanstack/react-start-rsc 0.1.58 → 0.1.59 Changeset
@tanstack/router-generator 1.167.39 → 1.167.40 Changeset
@tanstack/react-start 1.168.59 → 1.168.60 Dependent
@tanstack/react-start-client 1.168.38 → 1.168.39 Dependent
@tanstack/react-start-server 1.167.45 → 1.167.46 Dependent
@tanstack/router-cli 1.167.39 → 1.167.40 Dependent
@tanstack/router-vite-plugin 1.167.41 → 1.167.42 Dependent
@tanstack/solid-start 1.168.56 → 1.168.57 Dependent
@tanstack/vue-start 1.168.55 → 1.168.56 Dependent

@nx-cloud

nx-cloud Bot commented Sep 27, 2026 •

Copy link
Copy Markdown
Contributor

View your CI Pipeline Execution ↗ for commit c6c33e2

Command Status Duration Result
nx affected --targets=test:eslint,test:unit,tes... ✅ Succeeded 25m 52s View ↗
nx run-many --target=build --exclude=examples/*... ✅ Succeeded 2m 7s View ↗

☁️ Nx Cloud last updated this comment at 2026-09-27 20:42:27 UTC

@github-actions

Copy link
Copy Markdown
Contributor

Bundle Size Benchmarks

This pull request does not affect bundle size in any measured scenario.

@pkg-pr-new

pkg-pr-new Bot commented Sep 27, 2026 •

Copy link
Copy Markdown
More templates

@tanstack/arktype-adapter

npm i https://pkg.pr.new/@tanstack/arktype-adapter@8524

@tanstack/eslint-plugin-router

npm i https://pkg.pr.new/@tanstack/eslint-plugin-router@8524

@tanstack/eslint-plugin-start

npm i https://pkg.pr.new/@tanstack/eslint-plugin-start@8524

@tanstack/history

npm i https://pkg.pr.new/@tanstack/history@8524

@tanstack/nitro-v2-vite-plugin

npm i https://pkg.pr.new/@tanstack/nitro-v2-vite-plugin@8524

@tanstack/react-router

npm i https://pkg.pr.new/@tanstack/react-router@8524

@tanstack/react-router-devtools

npm i https://pkg.pr.new/@tanstack/react-router-devtools@8524

@tanstack/react-router-ssr-query

npm i https://pkg.pr.new/@tanstack/react-router-ssr-query@8524

@tanstack/react-start

npm i https://pkg.pr.new/@tanstack/react-start@8524

@tanstack/react-start-client

npm i https://pkg.pr.new/@tanstack/react-start-client@8524

@tanstack/react-start-rsc

npm i https://pkg.pr.new/@tanstack/react-start-rsc@8524

@tanstack/react-start-server

npm i https://pkg.pr.new/@tanstack/react-start-server@8524

@tanstack/router-cli

npm i https://pkg.pr.new/@tanstack/router-cli@8524

@tanstack/router-core

npm i https://pkg.pr.new/@tanstack/router-core@8524

@tanstack/router-devtools

npm i https://pkg.pr.new/@tanstack/router-devtools@8524

@tanstack/router-devtools-core

npm i https://pkg.pr.new/@tanstack/router-devtools-core@8524

@tanstack/router-generator

npm i https://pkg.pr.new/@tanstack/router-generator@8524

@tanstack/router-plugin

npm i https://pkg.pr.new/@tanstack/router-plugin@8524

@tanstack/router-ssr-query-core

npm i https://pkg.pr.new/@tanstack/router-ssr-query-core@8524

@tanstack/router-utils

npm i https://pkg.pr.new/@tanstack/router-utils@8524

@tanstack/router-vite-plugin

npm i https://pkg.pr.new/@tanstack/router-vite-plugin@8524

@tanstack/solid-router

npm i https://pkg.pr.new/@tanstack/solid-router@8524

@tanstack/solid-router-devtools

npm i https://pkg.pr.new/@tanstack/solid-router-devtools@8524

@tanstack/solid-router-ssr-query

npm i https://pkg.pr.new/@tanstack/solid-router-ssr-query@8524

@tanstack/solid-start

npm i https://pkg.pr.new/@tanstack/solid-start@8524

@tanstack/solid-start-client

npm i https://pkg.pr.new/@tanstack/solid-start-client@8524

@tanstack/solid-start-server

npm i https://pkg.pr.new/@tanstack/solid-start-server@8524

@tanstack/start-client-core

npm i https://pkg.pr.new/@tanstack/start-client-core@8524

@tanstack/start-fn-stubs

npm i https://pkg.pr.new/@tanstack/start-fn-stubs@8524

@tanstack/start-plugin-core

npm i https://pkg.pr.new/@tanstack/start-plugin-core@8524

@tanstack/start-server-core

npm i https://pkg.pr.new/@tanstack/start-server-core@8524

@tanstack/start-static-server-functions

npm i https://pkg.pr.new/@tanstack/start-static-server-functions@8524

@tanstack/start-storage-context

npm i https://pkg.pr.new/@tanstack/start-storage-context@8524

@tanstack/valibot-adapter

npm i https://pkg.pr.new/@tanstack/valibot-adapter@8524

@tanstack/virtual-file-routes

npm i https://pkg.pr.new/@tanstack/virtual-file-routes@8524

@tanstack/vue-router

npm i https://pkg.pr.new/@tanstack/vue-router@8524

@tanstack/vue-router-devtools

npm i https://pkg.pr.new/@tanstack/vue-router-devtools@8524

@tanstack/vue-router-ssr-query

npm i https://pkg.pr.new/@tanstack/vue-router-ssr-query@8524

@tanstack/vue-start

npm i https://pkg.pr.new/@tanstack/vue-start@8524

@tanstack/vue-start-client

npm i https://pkg.pr.new/@tanstack/vue-start-client@8524

@tanstack/vue-start-server

npm i https://pkg.pr.new/@tanstack/vue-start-server@8524

@tanstack/zod-adapter

npm i https://pkg.pr.new/@tanstack/zod-adapter@8524

commit: c6c33e2

@codspeed

codspeed Bot commented Sep 27, 2026 •

Copy link
Copy Markdown

Merging this PR will regress 1 benchmark

⚠️ Different runtime environments detected

Some benchmarks with significant performance changes were compared across different runtime environments,
which may affect the accuracy of the results.

Open the report in CodSpeed to investigate

⚡ 1 improved benchmark
❌ 1 regressed benchmark
✅ 189 untouched benchmarks
⏩ 3 skipped benchmarks1

Warning

Please fix the performance issues or acknowledge them on CodSpeed.

Performance Changes

Mode Benchmark BASE HEAD Efficiency
❌ Simulation hydrate restored route state and mixed Links (react) 213.2 ms 248.4 ms -14.18%
⚡ Memory mem server streaming-peak chunked (react) 642.3 KB 349.8 KB +83.63%

Tip

Investigate this regression by commenting @codspeedbot fix this regression on this PR, or directly use the CodSpeed MCP with your agent.


Comparing codex/direct-compiler-ast (c6c33e2) with codex/fuse-clone-provenance (080a540)

Open in CodSpeed

Footnotes

  1. 3 benchmarks were skipped, so the baseline results were used instead. If they were deleted from the codebase, click here and archive them to remove them from the performance reports. ↩

@Sheraff
Sheraff added this pull request to stack #8532 September 27, 2026 20:04
@Sheraff

Sheraff commented Oct 11, 2026

Copy link
Copy Markdown
Collaborator Author

Notes for when this is recut on top of #8504 (now on Yuku 0.19.0), from a review of Yuku's APIs:

  • Other compiler-owned snippets that fit builders, beyond server functions:
    • react-start-rsc/src/plugin/rscCssTransform.ts: the { __tanstackStartRscCss: null } options object and the <>{null}{null}</> fragment, which are parsed then patched behind runtime guards.
    • router-plugin/src/core/code-splitter/compilers.ts: the console.warn(...) statement. console is a global, so it needs linkGeneratedReference(node, 'console').
    • start-plugin-core/src/start-compiler/handleCreateServerFn.ts: the parseExpression(functionName) identifier.
    • start-plugin-core/src/hydrate-when-transform.ts: parseExpression(preloadName) and the <Component {...props} /> child.
  • Keep template parsing where builders would be 3–10× longer: HMR blocks, the provider's extracted handler, import-protection rewrite.ts, Hydrate lazy imports.
  • Every builder identifier that refers to a source or import binding must go through linkGeneratedReference, or removeUnusedBindings prunes its import.
  • Yuku 0.19's generate() prints any single node, so the builder snapshot tests no longer need to wrap nodes in a Program. generateModule now accepts a Node.
  • b.Property returns ObjectProperty | BindingProperty and b.Literal the whole Literal union, which is why this PR writes ObjectProperty and Directive records by hand. We may ask upstream for precise builders.
  • parseStatements('') now returns [] without parsing, so the provider-directives call no longer costs an analyze when nothing is missing.

Written with Claude Code (AI-assisted).

This branch has not been deployed

No deployments
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant