chore: release v0.4.0

Author: StanMarek

CHANGELOG.md

@@ -7,6 +7,8 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 
 ## [Unreleased]
 
+## [0.4.0] - 2026-03-31
+
 ### Added
 
 - **Kitty, WezTerm, Alacritty, Rio terminal support** — Ghost Complete now supports 7 terminals on macOS. Kitty, WezTerm, and Rio have full parity with Ghostty (DECSET 2026 + OSC 133). Alacritty uses DECSET 2026 with shell integration prompt detection (no native OSC 133).
@@ -214,6 +216,7 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 - **Shell integration** for zsh (full), bash (Ctrl+/), and fish (Ctrl+/)
 - **`validate-specs` subcommand** with colored output and item counts
 
+[0.4.0]: https://github.com/StanMarek/ghost-complete/releases/tag/v0.4.0
 [0.3.0]: https://github.com/StanMarek/ghost-complete/releases/tag/v0.3.0
 [0.2.5]: https://github.com/StanMarek/ghost-complete/releases/tag/v0.2.5
 [0.2.4]: https://github.com/StanMarek/ghost-complete/releases/tag/v0.2.4

Cargo.lock

@@ -12,7 +12,7 @@ members = [
 ]
 
 [workspace.package]
-version = "0.3.0"
+version = "0.4.0"
 edition = "2021"
 rust-version = "1.75"
 license = "MIT"

Cargo.toml

@@ -0,0 +1,72 @@
+# Benchmark Report — v0.4.0
+
+**Date:** 2026-03-31
+**Machine:** Apple Silicon (macOS), release build
+**Rust:** 1.75+ (workspace MSRV)
+
+## Summary
+
+v0.4.0 adds Kitty, WezTerm, Alacritty, and Rio terminal support — expanding from 3 to 7 supported terminals. Changes are confined to `gc-terminal` (detection logic) and `ghost-complete` (install init block). No hot-path code was modified.
+
+## Results
+
+### VT Parse Throughput (gc-parser)
+
+| Benchmark | Median | Range | vs v0.3.0 |
+|-----------|--------|-------|-----------|
+| `vt_parse_throughput/plain_text` | 83.56 µs | [83.46 µs, 83.68 µs] | -2.1% |
+| `vt_parse_throughput/ansi_colored` | 152.56 µs | [152.04 µs, 153.55 µs] | -1.6% |
+| `vt_parse_throughput/cursor_heavy` | 238.01 µs | [237.64 µs, 238.35 µs] | -2.7% |
+
+No regressions. Slight improvements likely due to run-to-run variance.
+
+### Fuzzy Ranking (gc-suggest)
+
+| Benchmark | Median | Range | vs v0.3.0 |
+|-----------|--------|-------|-----------|
+| `fuzzy_ranking/1k_3char` | 101.91 µs | [101.79 µs, 102.03 µs] | +1.2% |
+| `fuzzy_ranking/10k_3char` | 1.043 ms | [1.041 ms, 1.045 ms] | +0.6% |
+| `fuzzy_ranking/10k_empty` | 351.94 µs | [351.48 µs, 352.39 µs] | -0.5% |
+
+Within noise. Fuzzy ranking untouched.
+
+### Spec Loading (gc-suggest)
+
+| Benchmark | Median | Range | vs v0.3.0 |
+|-----------|--------|-------|-----------|
+| `spec_loading/load_717_specs` | 53.46 ms | [53.35 ms, 53.57 ms] | -6.2% |
+
+Apparent improvement likely due to filesystem caching. Spec loading code unchanged.
+
+### Spec Resolution (gc-suggest)
+
+| Benchmark | Median | Range | vs v0.3.0 |
+|-----------|--------|-------|-----------|
+| `spec_resolution/shallow` | 2.260 µs | [2.258 µs, 2.263 µs] | -1.5% |
+| `spec_resolution/deep` | 1.257 µs | [1.256 µs, 1.259 µs] | -0.2% |
+
+No change.
+
+### Transform Pipeline (gc-suggest)
+
+| Benchmark | Median | Range | vs v0.3.0 |
+|-----------|--------|-------|-----------|
+| `transform_pipeline/simple` | 31.47 µs | [31.41 µs, 31.53 µs] | -2.4% |
+| `transform_pipeline/regex` | 154.58 µs | [154.23 µs, 154.93 µs] | -1.0% |
+| `transform_pipeline/json` | 28.83 µs | [28.80 µs, 28.86 µs] | -1.8% |
+
+No change.
+
+### Engine Suggest Sync (gc-suggest)
+
+| Benchmark | Median | Range | vs v0.3.0 |
+|-----------|--------|-------|-----------|
+| `engine_suggest_sync/command_position` | 18.86 µs | [18.80 µs, 18.93 µs] | +3.5% |
+| `engine_suggest_sync/subcommand_with_spec` | 17.44 µs | [17.39 µs, 17.50 µs] | +4.8% |
+| `engine_suggest_sync/filesystem_fallback` | 1.079 ms | [1.078 ms, 1.081 ms] | -1.5% |
+
+Within noise. Absolute times well within <50ms budget.
+
+## Conclusion
+
+No performance-impacting changes in v0.4.0. Terminal detection additions are startup-only code in `gc-terminal`. All benchmarks remain within performance targets.