chore: cli benchmark

[timofei-iatsenko]

Description

This PR brings a CLI benchmark

Read readme.md from this PR to know more about it. Also CLAUDE.md has information about how the fixtures are generated.

══════════════════════════════════════════════════════════════
  Lingui Benchmark — Preset: medium
  1000 files · 10000 messages · 5 locales
  Node v24.13.1 · darwin arm64
  Apple M3 Max (16 cores)
══════════════════════════════════════════════════════════════

Running: Extract (full pipeline)...

  Extract (scan → extract → merge → write):
    Babel · 1 worker   █████████████████████████  2.75s  364 files/s
    Babel · 2 workers  ██████████████████░░░░░░░  1.95s  514 files/s
    SWC · 1 worker     █████████████░░░░░░░░░░░░  1.44s  694 files/s
    SWC · 2 workers    ██████████░░░░░░░░░░░░░░░  1.06s  942 files/s ⚡

  Summary:
    SWC · 2 workers is 2.6x faster than Babel · 1 worker
    SWC · 2 workers is 1.8x faster than Babel · 2 workers
    SWC · 2 workers is 1.4x faster than SWC · 1 worker

Running: Extract Template (no merge)...

  Extract Template (scan → extract → write, no merge):
    Babel · 1 worker   █████████████████████████  2.49s  401 files/s ±6.8%
    Babel · 2 workers  █████████████████░░░░░░░░  1.72s  582 files/s ±6.9%
    SWC · 1 worker     ███████████░░░░░░░░░░░░░░  1.09s  921 files/s ±5.8%
    SWC · 2 workers    ████████░░░░░░░░░░░░░░░░░  812ms  1232 files/s ⚡

  Summary:
    SWC · 2 workers is 3.1x faster than Babel · 1 worker
    SWC · 2 workers is 2.1x faster than Babel · 2 workers
    SWC · 2 workers is 1.3x faster than SWC · 1 worker

Running: Compile...

  Compile (read PO → compile ICU → write JS):
    1 worker   █████████████████████████  596ms  83.9k msgs/s
    2 workers  █████████████████████████  588ms  85.0k msgs/s ⚡

  Summary:
    2 workers is 1.0x faster than 1 worker

Running: Pure Macro Transform...

  Pure Macro Transform (no I/O, no catalogs):
    Babel             █████████████████████████  1.64s  611 files/s ±13.0%
    Babel - no macro  ██████████████░░░░░░░░░░░  932ms  1073 files/s ±5.9%
    SWC               ███████████░░░░░░░░░░░░░░  695ms  1438 files/s ±9.3%
    SWC - no macro    █░░░░░░░░░░░░░░░░░░░░░░░░   28ms  36.0k files/s ⚡

  Summary:
    SWC - no macro is 58.9x faster than Babel
    SWC - no macro is 33.6x faster than Babel - no macro
    SWC - no macro is 25.0x faster than SWC

Types of changes

• Bugfix (non-breaking change which fixes an issue)
• New feature (non-breaking change which adds functionality)
• Breaking change (fix or feature that would cause existing functionality to not work as expected)
• Documentation update
• Examples update

Fixes # (issue)

Checklist

• I have read the CONTRIBUTING and CODE_OF_CONDUCT docs
• I have added tests that prove my fix is effective or that my feature works
• I have added the necessary documentation (if appropriate)

[vercel]

The latest updates on your projects. Learn more about Vercel for GitHub.

Project	Deployment	Actions	Updated (UTC)
js-lingui	Ready	Preview	Jun 10, 2026 11:19am

[github-actions]

size-limit report 📦

Path	Size
packages/core/dist/index.mjs	1.69 KB (0%)
packages/detect-locale/dist/index.mjs	633 B (0%)
packages/react/dist/index.mjs	1.22 KB (0%)

[Copilot]

Pull request overview

Adds a new @lingui/benchmarks workspace providing a CLI-focused benchmark suite (fixture generation + scenario runners) intended to measure extraction/compile/macro-transform performance across different configurations.

Changes:

• Adds a new benchmarks/ workspace with fixture generation, benchmark scenarios, and console/JSON reporting.
• Updates the root workspace configuration and lockfile to include benchmark dependencies (SWC, tinybench, lingui-swc, etc.).
• Documents benchmark usage and fixture generation in benchmarks/README.md and benchmarks/CLAUDE.md.

Reviewed changes

Copilot reviewed 23 out of 24 changed files in this pull request and generated 22 comments.

Show a summary per file

File	Description
yarn.lock	Adds workspace + benchmark dependency resolutions (SWC, tinybench, lingui-swc, etc.).
package.json	Registers benchmarks/ as a workspace.
benchmarks/package.json	Defines @lingui/benchmarks scripts and dependencies.
benchmarks/tsconfig.json	Adds a TS config for the benchmarks workspace.
benchmarks/README.md	Documents running the benchmark suite and scenarios.
benchmarks/CLAUDE.md	Developer guide for architecture/fixtures/scenarios.
benchmarks/.gitignore	Ignores generated fixtures/results and SWC artifacts.
benchmarks/src/presets.ts	Defines benchmark presets (small/medium/large).
benchmarks/src/generate-fixtures.ts	Generates synthetic fixtures and catalogs for benchmarking.
benchmarks/src/run-benchmarks.ts	Runs benchmark scenarios and writes aggregated results.
benchmarks/src/reporters/console-reporter.ts	Renders console output (bars, throughput, summary).
benchmarks/src/reporters/json-reporter.ts	Writes machine-readable benchmark results to JSON.
benchmarks/src/scenarios/extract.bench.ts	Benchmarks lingui extract across Babel/SWC and worker counts.
benchmarks/src/scenarios/extract-template.bench.ts	Benchmarks lingui extract-template across Babel/SWC and worker counts.
benchmarks/src/scenarios/compile.bench.ts	Benchmarks lingui compile across worker counts.
benchmarks/src/scenarios/macro-transform.bench.ts	Benchmarks pure Babel vs SWC macro transforms (no CLI).
benchmarks/src/utils/run-cli.ts	Runs Lingui CLI in a subprocess with error surfacing.
benchmarks/src/utils/config-builder.ts	Writes temporary Lingui config files for fixtures.
benchmarks/src/utils/silence.ts	Utility to silence console/stdout during iterations.
benchmarks/src/utils/deterministic-random.ts	Seeded RNG utilities (currently unused).
benchmarks/src/generators/message-pool.ts	Message pool + variability rules for fixture generation.
benchmarks/src/generators/jsx-file-generator.ts	Generates TSX files using Lingui React macros.
benchmarks/src/generators/js-file-generator.ts	Generates TS files using Lingui core macros.
benchmarks/src/generators/po-catalog-generator.ts	Generates PO catalogs by extracting a real template then filling translations.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[andrii-bodnar]

I would suggest not shipping these unused things. We can always add them separately if needed

[andrii-bodnar]

This is not yet an official Lingui extractor, so better to hold off on using it for benchmarks until it goes live.

[andrii-bodnar]

the example imports buildConfig and silenceConsole, but neither exists in the current code. Scenarios now use writeConfigs() + runLingui()