pybun-cli 0.1.25__tar.gz → 0.1.27__tar.gz

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
@@ -1,6 +1,6 @@
1
1
  Metadata-Version: 2.4
2
2
  Name: pybun-cli
3
- Version: 0.1.25
3
+ Version: 0.1.27
4
4
  Summary: PyBun shim bootstrapper for signed releases
5
5
  Author: PyBun Authors
6
6
  License-Expression: MIT
@@ -26,7 +26,7 @@ Dynamic: license-file
26
26
 
27
27
  <p align="center">
28
28
  <em>Safe, structured, deterministic Python execution for AI coding agents —<br>
29
- powered by uv/pytest underneath, with JSON-first output and a built-in MCP server.</em>
29
+ orchestrating Python, pytest, and command-specific uv backends with JSON-first output and a built-in MCP server.</em>
30
30
  </p>
31
31
 
32
32
  <p align="center">
@@ -90,7 +90,7 @@ pybun run -c "import requests; print('Hello, PyBun!')"
90
90
 
91
91
  Existing Python tools are built for **humans**. PyBun is designed for **AI agents** — and humans who work alongside them.
92
92
 
93
- **PyBun is not a speed competitor to uv.** Tools like uv and pip are excellent at dependency resolution and installation — PyBun doesn't try to out-run them, and where uv is available PyBun delegates to it directly. What those tools lack is an **agent-facing control layer**: structured output, MCP integration, sandboxed execution, and audit/provenance that AI systems can rely on without fragile text scraping or unverifiable side effects.
93
+ **PyBun is not a speed competitor to uv.** Tools like uv and pip are excellent at dependency resolution and installation. PyBun adds an **agent-facing control layer**: structured output, MCP integration, sandboxed execution, and audit/provenance that AI systems can rely on without fragile text scraping or unverifiable side effects. Delegation is currently command-specific: `pybun x` and eligible PEP 723 script runs use uv when it is available, while ordinary dependency commands still use PyBun's native path.
94
94
 
95
95
  The value PyBun adds isn't "faster pip." It's removing the ambiguity, non-determinism, and risk an agent faces when it operates a Python environment on its own.
96
96
 
@@ -110,7 +110,7 @@ The value PyBun adds isn't "faster pip." It's removing the ambiguity, non-determ
110
110
  uv pytest Python
111
111
  ```
112
112
 
113
- PyBun stays a thin, structured control layer on top of proven execution backends rather than re-implementing the Python packaging/import ecosystem from scratch. See [`docs/SPECS.md`](docs/SPECS.md#03-外部レビューによる方向性提言-2026-08-29) for the reasoning behind this scope decision.
113
+ PyBun is moving toward a thin, structured control layer on top of proven execution backends rather than expanding its reimplementation of the Python packaging/import ecosystem. The v0.2.0 target architecture broadens uv delegation; it does not describe every current command path. See [`docs/SPECS.md`](docs/SPECS.md#03-外部レビューによる方向性提言-2026-08-29) for the reasoning behind this scope decision.
114
114
 
115
115
  ### ✨ What PyBun adds that other tools don't
116
116
 
@@ -137,11 +137,24 @@ The AI receives structured JSON — no parsing required, no ambiguity.
137
137
 
138
138
  ## Status
139
139
 
140
- - **Current:** M1 (Fast Installer), M2 (Runtime Optimization), M3 (Tester), and M4 (MCP/JSON) are stable or near-stable.
141
- - `pybun install` / `pybun x` (with uv backend) / `pybun run` / `pybun test` (default pytest/unittest wrapper backend) are **Stable** for the common case. The native wheel installer used by `pybun install` does not yet implement the full PEP 427 `.data` installation semantics (entry points/scripts/headers beyond the common paths) — see [Issue #402](https://github.com/VOID-TECHNOLOGY-INC/PyBun/issues/402). Packages relying on those payloads may need `pybun x uv -- pip install ...` as a fallback.
142
- - `pybun test --backend=pybun` (native executor, integrated per PR-A4) and `pybun watch` (native monitoring on macOS/Linux, polling fallback on standard builds) are **Preview** — the native test backend still surfaces `W_TEST_BACKEND_COMPAT_*` diagnostics for known pytest-plugin/fixture gaps.
143
- - The Rust Module Finder (`pybun module-find`) is **Experimental / not wired into CPython's runtime import path** ([Issue #403](https://github.com/VOID-TECHNOLOGY-INC/PyBun/issues/403)). Treat it as an opportunistic optimization under ROI evaluation, not a core dependency.
144
- - Windows support is **Preview**.
140
+ Maturity is assigned per execution path rather than to an entire milestone:
141
+
142
+ | Surface | Maturity | Current execution path |
143
+ | --- | --- | --- |
144
+ | `pybun install` | **Preview** | PyBun's native resolver, downloader, and wheel installer. Path containment is hardened by [Issue #431](https://github.com/VOID-TECHNOLOGY-INC/PyBun/issues/431), but this is not a claim of full Python packaging compatibility. |
145
+ | `pybun x` | **Stable for common cases** | Creates a temporary environment and uses `uv pip install` when uv is available; otherwise it falls back to pip. |
146
+ | `pybun run` | **Stable for common cases** | Runs ordinary scripts with Python. Eligible PEP 723 scripts use `uv run --script` automatically when uv is available. Other PEP 723 paths are orchestrated by PyBun and may still use `uv pip install` for dependencies. |
147
+ | `pybun test` | **Stable for common cases** | The default backend wraps pytest/unittest. `--backend=pybun` is **Preview** and reports `W_TEST_BACKEND_COMPAT_*` diagnostics for known plugin/fixture gaps. |
148
+ | `pybun watch` | **Preview** | Native monitoring on macOS/Linux, with polling fallback on standard builds. |
149
+ | `pybun module-find` | **Experimental** | Standalone Rust Module Finder; it is not wired into CPython's runtime import path ([Issue #403](https://github.com/VOID-TECHNOLOGY-INC/PyBun/issues/403)). |
150
+ | Windows support | **Preview** | CI compile coverage and selected cross-platform behavior; continue to validate command-specific behavior. |
151
+
152
+ `pybun install`, `pybun lock`, and `pybun upgrade` use PyBun's native dependency path today; installing uv does not switch those commands to an uv backend. Broader dependency-operation delegation is a v0.2.0 target.
153
+
154
+ PyBun-managed PEP 723 environments may still use `uv pip install` for dependency installation when uv is available, including sandboxed, no-cache, and PyBun-lockfile paths. Without uv, those paths fall back to native resolution/installation or pip depending on the environment mode. This is separate from direct `uv run --script` delegation.
155
+
156
+ The native installer's `.data/{purelib,platlib,scripts,headers,data}` routing is implemented and tested by the work that closed [Issue #402](https://github.com/VOID-TECHNOLOGY-INC/PyBun/issues/402). That tested subset includes relocation, script permissions/shebang handling, and headers placement; it does not promise complete wheel compatibility such as console-entry-point generation, every platform scheme, source distributions, or arbitrary install hooks.
157
+
145
158
  - **Platforms:** macOS/Linux (arm64/amd64), Windows (preview)
146
159
 
147
160
  > PyBun's priority is the agent-facing control layer (JSON/MCP/diagnostics/sandbox/audit/drift), not re-implementing Python packaging/import internals. See [`docs/SPECS.md`](docs/SPECS.md#03-外部レビューによる方向性提言-2026-08-29) for feature maturity, phased rollout policy, and the current scope decisions.
@@ -232,12 +245,12 @@ PEP 723 inline metadata is also supported:
232
245
  # ///
233
246
  import requests
234
247
  ```
235
- ※ Metadata parsing, automatic dependency installation, and isolated-environment execution are all implemented and stable (cached per script/dependency/Python-version key; see `docs/PLAN.md` for details).
248
+ ※ Metadata parsing and isolated-environment execution are supported. Direct `uv run --script` is Stable for common cases; PyBun-managed fallback paths inherit the native installer's Preview limitations. Environments are cached per script/dependency/Python-version key; see `docs/PLAN.md` for details.
236
249
 
237
250
  ### Ad-hoc Execution (`pybun x`)
238
251
 
239
252
  Install a package in a temporary environment and execute it (Python version of `npx`).
240
- If `uv` is available, it is used for faster environment creation.
253
+ The environment is created with `python -m venv`. If `uv` is available, PyBun uses `uv pip install` for package installation; otherwise it uses pip.
241
254
 
242
255
  ```bash
243
256
  # Temporarily install and run cowsay
@@ -619,9 +632,9 @@ cargo test mcp
619
632
 
620
633
  ## Benchmarks
621
634
 
622
- PyBun is not a speed competitor to uv — it's an interface layer. PyBun uses uv as an optional execution backend for some operations (e.g. PEP 723 script runs). Where uv is available, PyBun delegates to it transparently — so warm-cache script execution is at parity with running uv directly.
635
+ PyBun is not a speed competitor to uv — it is an interface layer. Current uv delegation is limited to command paths such as `pybun x` package installation and eligible PEP 723 script runs. Ordinary `pybun install`, `pybun lock`, and `pybun upgrade` still use native dependency code; the v0.2.0 architecture aims to delegate more of that work.
623
636
 
624
- The areas where PyBun intentionally differs from uv (JSON output, MCP, sandbox) are not speed-sensitive. For raw dependency resolution speed, uv's PubGrub solver is significantly faster than PyBun's current greedy resolver — this is a known roadmap item tracked in [Issue #117](https://github.com/VOID-TECHNOLOGY-INC/PyBun/issues/117).
637
+ The areas where PyBun intentionally differs from uv (JSON output, MCP, sandbox) are evaluated for structured correctness and safety first. Native resolver performance work and measurements are tracked in [Issue #239](https://github.com/VOID-TECHNOLOGY-INC/PyBun/issues/239); they should not be read as a product goal to out-run uv.
625
638
 
626
639
  Full numbers: [docs/BENCHMARK_UV_COMPARISON.md](docs/BENCHMARK_UV_COMPARISON.md)
627
640
 
@@ -630,7 +643,7 @@ Full numbers: [docs/BENCHMARK_UV_COMPARISON.md](docs/BENCHMARK_UV_COMPARISON.md)
630
643
  ## Roadmap
631
644
 
632
645
  - [x] M0: Repository & CI scaffold
633
- - [x] M1: Fast installer (lockfile, resolver, PEP 723)
646
+ - [x] M1: Installer foundation (lockfile, native resolver, PEP 723); native install remains Preview while the v0.2.0 backend transition is pending
634
647
  - [x] M2: Runtime optimization (module finder, lazy import, hot reload) — *opportunistic, ROI under evaluation; see [Issue #403](https://github.com/VOID-TECHNOLOGY-INC/PyBun/issues/403)*
635
648
  - [x] M3: Test runner (discovery, parallel execution, snapshots)
636
649
  - [x] M4: JSON/MCP & diagnostics
@@ -6,7 +6,7 @@
6
6
 
7
7
  <p align="center">
8
8
  <em>Safe, structured, deterministic Python execution for AI coding agents —<br>
9
- powered by uv/pytest underneath, with JSON-first output and a built-in MCP server.</em>
9
+ orchestrating Python, pytest, and command-specific uv backends with JSON-first output and a built-in MCP server.</em>
10
10
  </p>
11
11
 
12
12
  <p align="center">
@@ -70,7 +70,7 @@ pybun run -c "import requests; print('Hello, PyBun!')"
70
70
 
71
71
  Existing Python tools are built for **humans**. PyBun is designed for **AI agents** — and humans who work alongside them.
72
72
 
73
- **PyBun is not a speed competitor to uv.** Tools like uv and pip are excellent at dependency resolution and installation — PyBun doesn't try to out-run them, and where uv is available PyBun delegates to it directly. What those tools lack is an **agent-facing control layer**: structured output, MCP integration, sandboxed execution, and audit/provenance that AI systems can rely on without fragile text scraping or unverifiable side effects.
73
+ **PyBun is not a speed competitor to uv.** Tools like uv and pip are excellent at dependency resolution and installation. PyBun adds an **agent-facing control layer**: structured output, MCP integration, sandboxed execution, and audit/provenance that AI systems can rely on without fragile text scraping or unverifiable side effects. Delegation is currently command-specific: `pybun x` and eligible PEP 723 script runs use uv when it is available, while ordinary dependency commands still use PyBun's native path.
74
74
 
75
75
  The value PyBun adds isn't "faster pip." It's removing the ambiguity, non-determinism, and risk an agent faces when it operates a Python environment on its own.
76
76
 
@@ -90,7 +90,7 @@ The value PyBun adds isn't "faster pip." It's removing the ambiguity, non-determ
90
90
  uv pytest Python
91
91
  ```
92
92
 
93
- PyBun stays a thin, structured control layer on top of proven execution backends rather than re-implementing the Python packaging/import ecosystem from scratch. See [`docs/SPECS.md`](docs/SPECS.md#03-外部レビューによる方向性提言-2026-08-29) for the reasoning behind this scope decision.
93
+ PyBun is moving toward a thin, structured control layer on top of proven execution backends rather than expanding its reimplementation of the Python packaging/import ecosystem. The v0.2.0 target architecture broadens uv delegation; it does not describe every current command path. See [`docs/SPECS.md`](docs/SPECS.md#03-外部レビューによる方向性提言-2026-08-29) for the reasoning behind this scope decision.
94
94
 
95
95
  ### ✨ What PyBun adds that other tools don't
96
96
 
@@ -117,11 +117,24 @@ The AI receives structured JSON — no parsing required, no ambiguity.
117
117
 
118
118
  ## Status
119
119
 
120
- - **Current:** M1 (Fast Installer), M2 (Runtime Optimization), M3 (Tester), and M4 (MCP/JSON) are stable or near-stable.
121
- - `pybun install` / `pybun x` (with uv backend) / `pybun run` / `pybun test` (default pytest/unittest wrapper backend) are **Stable** for the common case. The native wheel installer used by `pybun install` does not yet implement the full PEP 427 `.data` installation semantics (entry points/scripts/headers beyond the common paths) — see [Issue #402](https://github.com/VOID-TECHNOLOGY-INC/PyBun/issues/402). Packages relying on those payloads may need `pybun x uv -- pip install ...` as a fallback.
122
- - `pybun test --backend=pybun` (native executor, integrated per PR-A4) and `pybun watch` (native monitoring on macOS/Linux, polling fallback on standard builds) are **Preview** — the native test backend still surfaces `W_TEST_BACKEND_COMPAT_*` diagnostics for known pytest-plugin/fixture gaps.
123
- - The Rust Module Finder (`pybun module-find`) is **Experimental / not wired into CPython's runtime import path** ([Issue #403](https://github.com/VOID-TECHNOLOGY-INC/PyBun/issues/403)). Treat it as an opportunistic optimization under ROI evaluation, not a core dependency.
124
- - Windows support is **Preview**.
120
+ Maturity is assigned per execution path rather than to an entire milestone:
121
+
122
+ | Surface | Maturity | Current execution path |
123
+ | --- | --- | --- |
124
+ | `pybun install` | **Preview** | PyBun's native resolver, downloader, and wheel installer. Path containment is hardened by [Issue #431](https://github.com/VOID-TECHNOLOGY-INC/PyBun/issues/431), but this is not a claim of full Python packaging compatibility. |
125
+ | `pybun x` | **Stable for common cases** | Creates a temporary environment and uses `uv pip install` when uv is available; otherwise it falls back to pip. |
126
+ | `pybun run` | **Stable for common cases** | Runs ordinary scripts with Python. Eligible PEP 723 scripts use `uv run --script` automatically when uv is available. Other PEP 723 paths are orchestrated by PyBun and may still use `uv pip install` for dependencies. |
127
+ | `pybun test` | **Stable for common cases** | The default backend wraps pytest/unittest. `--backend=pybun` is **Preview** and reports `W_TEST_BACKEND_COMPAT_*` diagnostics for known plugin/fixture gaps. |
128
+ | `pybun watch` | **Preview** | Native monitoring on macOS/Linux, with polling fallback on standard builds. |
129
+ | `pybun module-find` | **Experimental** | Standalone Rust Module Finder; it is not wired into CPython's runtime import path ([Issue #403](https://github.com/VOID-TECHNOLOGY-INC/PyBun/issues/403)). |
130
+ | Windows support | **Preview** | CI compile coverage and selected cross-platform behavior; continue to validate command-specific behavior. |
131
+
132
+ `pybun install`, `pybun lock`, and `pybun upgrade` use PyBun's native dependency path today; installing uv does not switch those commands to an uv backend. Broader dependency-operation delegation is a v0.2.0 target.
133
+
134
+ PyBun-managed PEP 723 environments may still use `uv pip install` for dependency installation when uv is available, including sandboxed, no-cache, and PyBun-lockfile paths. Without uv, those paths fall back to native resolution/installation or pip depending on the environment mode. This is separate from direct `uv run --script` delegation.
135
+
136
+ The native installer's `.data/{purelib,platlib,scripts,headers,data}` routing is implemented and tested by the work that closed [Issue #402](https://github.com/VOID-TECHNOLOGY-INC/PyBun/issues/402). That tested subset includes relocation, script permissions/shebang handling, and headers placement; it does not promise complete wheel compatibility such as console-entry-point generation, every platform scheme, source distributions, or arbitrary install hooks.
137
+
125
138
  - **Platforms:** macOS/Linux (arm64/amd64), Windows (preview)
126
139
 
127
140
  > PyBun's priority is the agent-facing control layer (JSON/MCP/diagnostics/sandbox/audit/drift), not re-implementing Python packaging/import internals. See [`docs/SPECS.md`](docs/SPECS.md#03-外部レビューによる方向性提言-2026-08-29) for feature maturity, phased rollout policy, and the current scope decisions.
@@ -212,12 +225,12 @@ PEP 723 inline metadata is also supported:
212
225
  # ///
213
226
  import requests
214
227
  ```
215
- ※ Metadata parsing, automatic dependency installation, and isolated-environment execution are all implemented and stable (cached per script/dependency/Python-version key; see `docs/PLAN.md` for details).
228
+ ※ Metadata parsing and isolated-environment execution are supported. Direct `uv run --script` is Stable for common cases; PyBun-managed fallback paths inherit the native installer's Preview limitations. Environments are cached per script/dependency/Python-version key; see `docs/PLAN.md` for details.
216
229
 
217
230
  ### Ad-hoc Execution (`pybun x`)
218
231
 
219
232
  Install a package in a temporary environment and execute it (Python version of `npx`).
220
- If `uv` is available, it is used for faster environment creation.
233
+ The environment is created with `python -m venv`. If `uv` is available, PyBun uses `uv pip install` for package installation; otherwise it uses pip.
221
234
 
222
235
  ```bash
223
236
  # Temporarily install and run cowsay
@@ -599,9 +612,9 @@ cargo test mcp
599
612
 
600
613
  ## Benchmarks
601
614
 
602
- PyBun is not a speed competitor to uv — it's an interface layer. PyBun uses uv as an optional execution backend for some operations (e.g. PEP 723 script runs). Where uv is available, PyBun delegates to it transparently — so warm-cache script execution is at parity with running uv directly.
615
+ PyBun is not a speed competitor to uv — it is an interface layer. Current uv delegation is limited to command paths such as `pybun x` package installation and eligible PEP 723 script runs. Ordinary `pybun install`, `pybun lock`, and `pybun upgrade` still use native dependency code; the v0.2.0 architecture aims to delegate more of that work.
603
616
 
604
- The areas where PyBun intentionally differs from uv (JSON output, MCP, sandbox) are not speed-sensitive. For raw dependency resolution speed, uv's PubGrub solver is significantly faster than PyBun's current greedy resolver — this is a known roadmap item tracked in [Issue #117](https://github.com/VOID-TECHNOLOGY-INC/PyBun/issues/117).
617
+ The areas where PyBun intentionally differs from uv (JSON output, MCP, sandbox) are evaluated for structured correctness and safety first. Native resolver performance work and measurements are tracked in [Issue #239](https://github.com/VOID-TECHNOLOGY-INC/PyBun/issues/239); they should not be read as a product goal to out-run uv.
605
618
 
606
619
  Full numbers: [docs/BENCHMARK_UV_COMPARISON.md](docs/BENCHMARK_UV_COMPARISON.md)
607
620
 
@@ -610,7 +623,7 @@ Full numbers: [docs/BENCHMARK_UV_COMPARISON.md](docs/BENCHMARK_UV_COMPARISON.md)
610
623
  ## Roadmap
611
624
 
612
625
  - [x] M0: Repository & CI scaffold
613
- - [x] M1: Fast installer (lockfile, resolver, PEP 723)
626
+ - [x] M1: Installer foundation (lockfile, native resolver, PEP 723); native install remains Preview while the v0.2.0 backend transition is pending
614
627
  - [x] M2: Runtime optimization (module finder, lazy import, hot reload) — *opportunistic, ROI under evaluation; see [Issue #403](https://github.com/VOID-TECHNOLOGY-INC/PyBun/issues/403)*
615
628
  - [x] M3: Test runner (discovery, parallel execution, snapshots)
616
629
  - [x] M4: JSON/MCP & diagnostics
@@ -1,6 +1,6 @@
1
1
  Metadata-Version: 2.4
2
2
  Name: pybun-cli
3
- Version: 0.1.25
3
+ Version: 0.1.27
4
4
  Summary: PyBun shim bootstrapper for signed releases
5
5
  Author: PyBun Authors
6
6
  License-Expression: MIT
@@ -26,7 +26,7 @@ Dynamic: license-file
26
26
 
27
27
  <p align="center">
28
28
  <em>Safe, structured, deterministic Python execution for AI coding agents —<br>
29
- powered by uv/pytest underneath, with JSON-first output and a built-in MCP server.</em>
29
+ orchestrating Python, pytest, and command-specific uv backends with JSON-first output and a built-in MCP server.</em>
30
30
  </p>
31
31
 
32
32
  <p align="center">
@@ -90,7 +90,7 @@ pybun run -c "import requests; print('Hello, PyBun!')"
90
90
 
91
91
  Existing Python tools are built for **humans**. PyBun is designed for **AI agents** — and humans who work alongside them.
92
92
 
93
- **PyBun is not a speed competitor to uv.** Tools like uv and pip are excellent at dependency resolution and installation — PyBun doesn't try to out-run them, and where uv is available PyBun delegates to it directly. What those tools lack is an **agent-facing control layer**: structured output, MCP integration, sandboxed execution, and audit/provenance that AI systems can rely on without fragile text scraping or unverifiable side effects.
93
+ **PyBun is not a speed competitor to uv.** Tools like uv and pip are excellent at dependency resolution and installation. PyBun adds an **agent-facing control layer**: structured output, MCP integration, sandboxed execution, and audit/provenance that AI systems can rely on without fragile text scraping or unverifiable side effects. Delegation is currently command-specific: `pybun x` and eligible PEP 723 script runs use uv when it is available, while ordinary dependency commands still use PyBun's native path.
94
94
 
95
95
  The value PyBun adds isn't "faster pip." It's removing the ambiguity, non-determinism, and risk an agent faces when it operates a Python environment on its own.
96
96
 
@@ -110,7 +110,7 @@ The value PyBun adds isn't "faster pip." It's removing the ambiguity, non-determ
110
110
  uv pytest Python
111
111
  ```
112
112
 
113
- PyBun stays a thin, structured control layer on top of proven execution backends rather than re-implementing the Python packaging/import ecosystem from scratch. See [`docs/SPECS.md`](docs/SPECS.md#03-外部レビューによる方向性提言-2026-08-29) for the reasoning behind this scope decision.
113
+ PyBun is moving toward a thin, structured control layer on top of proven execution backends rather than expanding its reimplementation of the Python packaging/import ecosystem. The v0.2.0 target architecture broadens uv delegation; it does not describe every current command path. See [`docs/SPECS.md`](docs/SPECS.md#03-外部レビューによる方向性提言-2026-08-29) for the reasoning behind this scope decision.
114
114
 
115
115
  ### ✨ What PyBun adds that other tools don't
116
116
 
@@ -137,11 +137,24 @@ The AI receives structured JSON — no parsing required, no ambiguity.
137
137
 
138
138
  ## Status
139
139
 
140
- - **Current:** M1 (Fast Installer), M2 (Runtime Optimization), M3 (Tester), and M4 (MCP/JSON) are stable or near-stable.
141
- - `pybun install` / `pybun x` (with uv backend) / `pybun run` / `pybun test` (default pytest/unittest wrapper backend) are **Stable** for the common case. The native wheel installer used by `pybun install` does not yet implement the full PEP 427 `.data` installation semantics (entry points/scripts/headers beyond the common paths) — see [Issue #402](https://github.com/VOID-TECHNOLOGY-INC/PyBun/issues/402). Packages relying on those payloads may need `pybun x uv -- pip install ...` as a fallback.
142
- - `pybun test --backend=pybun` (native executor, integrated per PR-A4) and `pybun watch` (native monitoring on macOS/Linux, polling fallback on standard builds) are **Preview** — the native test backend still surfaces `W_TEST_BACKEND_COMPAT_*` diagnostics for known pytest-plugin/fixture gaps.
143
- - The Rust Module Finder (`pybun module-find`) is **Experimental / not wired into CPython's runtime import path** ([Issue #403](https://github.com/VOID-TECHNOLOGY-INC/PyBun/issues/403)). Treat it as an opportunistic optimization under ROI evaluation, not a core dependency.
144
- - Windows support is **Preview**.
140
+ Maturity is assigned per execution path rather than to an entire milestone:
141
+
142
+ | Surface | Maturity | Current execution path |
143
+ | --- | --- | --- |
144
+ | `pybun install` | **Preview** | PyBun's native resolver, downloader, and wheel installer. Path containment is hardened by [Issue #431](https://github.com/VOID-TECHNOLOGY-INC/PyBun/issues/431), but this is not a claim of full Python packaging compatibility. |
145
+ | `pybun x` | **Stable for common cases** | Creates a temporary environment and uses `uv pip install` when uv is available; otherwise it falls back to pip. |
146
+ | `pybun run` | **Stable for common cases** | Runs ordinary scripts with Python. Eligible PEP 723 scripts use `uv run --script` automatically when uv is available. Other PEP 723 paths are orchestrated by PyBun and may still use `uv pip install` for dependencies. |
147
+ | `pybun test` | **Stable for common cases** | The default backend wraps pytest/unittest. `--backend=pybun` is **Preview** and reports `W_TEST_BACKEND_COMPAT_*` diagnostics for known plugin/fixture gaps. |
148
+ | `pybun watch` | **Preview** | Native monitoring on macOS/Linux, with polling fallback on standard builds. |
149
+ | `pybun module-find` | **Experimental** | Standalone Rust Module Finder; it is not wired into CPython's runtime import path ([Issue #403](https://github.com/VOID-TECHNOLOGY-INC/PyBun/issues/403)). |
150
+ | Windows support | **Preview** | CI compile coverage and selected cross-platform behavior; continue to validate command-specific behavior. |
151
+
152
+ `pybun install`, `pybun lock`, and `pybun upgrade` use PyBun's native dependency path today; installing uv does not switch those commands to an uv backend. Broader dependency-operation delegation is a v0.2.0 target.
153
+
154
+ PyBun-managed PEP 723 environments may still use `uv pip install` for dependency installation when uv is available, including sandboxed, no-cache, and PyBun-lockfile paths. Without uv, those paths fall back to native resolution/installation or pip depending on the environment mode. This is separate from direct `uv run --script` delegation.
155
+
156
+ The native installer's `.data/{purelib,platlib,scripts,headers,data}` routing is implemented and tested by the work that closed [Issue #402](https://github.com/VOID-TECHNOLOGY-INC/PyBun/issues/402). That tested subset includes relocation, script permissions/shebang handling, and headers placement; it does not promise complete wheel compatibility such as console-entry-point generation, every platform scheme, source distributions, or arbitrary install hooks.
157
+
145
158
  - **Platforms:** macOS/Linux (arm64/amd64), Windows (preview)
146
159
 
147
160
  > PyBun's priority is the agent-facing control layer (JSON/MCP/diagnostics/sandbox/audit/drift), not re-implementing Python packaging/import internals. See [`docs/SPECS.md`](docs/SPECS.md#03-外部レビューによる方向性提言-2026-08-29) for feature maturity, phased rollout policy, and the current scope decisions.
@@ -232,12 +245,12 @@ PEP 723 inline metadata is also supported:
232
245
  # ///
233
246
  import requests
234
247
  ```
235
- ※ Metadata parsing, automatic dependency installation, and isolated-environment execution are all implemented and stable (cached per script/dependency/Python-version key; see `docs/PLAN.md` for details).
248
+ ※ Metadata parsing and isolated-environment execution are supported. Direct `uv run --script` is Stable for common cases; PyBun-managed fallback paths inherit the native installer's Preview limitations. Environments are cached per script/dependency/Python-version key; see `docs/PLAN.md` for details.
236
249
 
237
250
  ### Ad-hoc Execution (`pybun x`)
238
251
 
239
252
  Install a package in a temporary environment and execute it (Python version of `npx`).
240
- If `uv` is available, it is used for faster environment creation.
253
+ The environment is created with `python -m venv`. If `uv` is available, PyBun uses `uv pip install` for package installation; otherwise it uses pip.
241
254
 
242
255
  ```bash
243
256
  # Temporarily install and run cowsay
@@ -619,9 +632,9 @@ cargo test mcp
619
632
 
620
633
  ## Benchmarks
621
634
 
622
- PyBun is not a speed competitor to uv — it's an interface layer. PyBun uses uv as an optional execution backend for some operations (e.g. PEP 723 script runs). Where uv is available, PyBun delegates to it transparently — so warm-cache script execution is at parity with running uv directly.
635
+ PyBun is not a speed competitor to uv — it is an interface layer. Current uv delegation is limited to command paths such as `pybun x` package installation and eligible PEP 723 script runs. Ordinary `pybun install`, `pybun lock`, and `pybun upgrade` still use native dependency code; the v0.2.0 architecture aims to delegate more of that work.
623
636
 
624
- The areas where PyBun intentionally differs from uv (JSON output, MCP, sandbox) are not speed-sensitive. For raw dependency resolution speed, uv's PubGrub solver is significantly faster than PyBun's current greedy resolver — this is a known roadmap item tracked in [Issue #117](https://github.com/VOID-TECHNOLOGY-INC/PyBun/issues/117).
637
+ The areas where PyBun intentionally differs from uv (JSON output, MCP, sandbox) are evaluated for structured correctness and safety first. Native resolver performance work and measurements are tracked in [Issue #239](https://github.com/VOID-TECHNOLOGY-INC/PyBun/issues/239); they should not be read as a product goal to out-run uv.
625
638
 
626
639
  Full numbers: [docs/BENCHMARK_UV_COMPARISON.md](docs/BENCHMARK_UV_COMPARISON.md)
627
640
 
@@ -630,7 +643,7 @@ Full numbers: [docs/BENCHMARK_UV_COMPARISON.md](docs/BENCHMARK_UV_COMPARISON.md)
630
643
  ## Roadmap
631
644
 
632
645
  - [x] M0: Repository & CI scaffold
633
- - [x] M1: Fast installer (lockfile, resolver, PEP 723)
646
+ - [x] M1: Installer foundation (lockfile, native resolver, PEP 723); native install remains Preview while the v0.2.0 backend transition is pending
634
647
  - [x] M2: Runtime optimization (module finder, lazy import, hot reload) — *opportunistic, ROI under evaluation; see [Issue #403](https://github.com/VOID-TECHNOLOGY-INC/PyBun/issues/403)*
635
648
  - [x] M3: Test runner (discovery, parallel execution, snapshots)
636
649
  - [x] M4: JSON/MCP & diagnostics
@@ -22,7 +22,7 @@ readme = { file = "README.md", content-type = "text/markdown" }
22
22
  requires-python = ">=3.8"
23
23
  license = "MIT"
24
24
  license-files = ["LICENSE"]
25
- version = "0.1.25"
25
+ version = "0.1.27"
26
26
 
27
27
  [[project.authors]]
28
28
  name = "PyBun Authors"
File without changes
File without changes
File without changes
File without changes
File without changes
File without changes