lz4ext 1.0__tar.gz

lz4ext-1.0/.gitattributes

@@ -0,0 +1,6 @@
+# Preserve the vendored upstream files byte-for-byte on every platform.
+src/lz4.c text eol=lf
+src/lz4.h text eol=lf
+src/lz4hc.c text eol=lf
+src/lz4hc.h text eol=lf
+src/LICENSE.lz4 text eol=lf

lz4ext-1.0/CHANGELOG.md

@@ -0,0 +1,17 @@
+# Changelog
+
+## 1.0
+
+- Update vendored LZ4 r119 to official v1.10.0.
+- Require CPython 3.11 or newer and contiguous binary input; reject text.
+- Check compression bounds and pass explicit output capacities.
+- Validate prefixed decoded lengths and empty framing; emit canonical empty blocks.
+- Return cleanly from raw decoder errors and check output capacity arithmetic.
+- Default decoder allocation limit to 64 MiB, with keyword-only `max_output_size`
+  overrides on every decoder alias.
+- Preserve the four-byte little-endian prefixed format and existing aliases.
+- Use PEP 517/621 packaging and include binding and vendored LZ4 licenses.
+- Expose package/library versions and default output limit.
+
+Historical malformed 20-byte empty records require trusted offline migration;
+see the README. Valid nonempty legacy blocks remain readable.

lz4ext-1.0/LICENSE

@@ -0,0 +1,28 @@
+Copyright (c) 2012-2013, Steeve Morin
+All rights reserved.
+
+Redistribution and use in source and binary forms, with or without
+modification, are permitted provided that the following conditions are met:
+
+1. Redistributions of source code must retain the above copyright notice,
+   this list of conditions and the following disclaimer.
+
+2. Redistributions in binary form must reproduce the above copyright notice,
+   this list of conditions and the following disclaimer in the documentation
+   and/or other materials provided with the distribution.
+
+3. Neither the name of Steeve Morin nor the names of its contributors may be
+   used to endorse or promote products derived from this software without
+   specific prior written permission.
+
+THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
+AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
+ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT HOLDER OR CONTRIBUTORS BE
+LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
+CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
+SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
+INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
+CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
+ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
+POSSIBILITY OF SUCH DAMAGE.

lz4ext-1.0/MANIFEST.in

@@ -0,0 +1,8 @@
+include README.rst
+include LICENSE CHANGELOG.md
+include RELEASING.md SECURITY.md
+include .gitattributes
+include tests/test.py tests/bench.py
+include src/*.h
+include src/LICENSE.lz4
+include src/README.lz4.md

lz4ext-1.0/PKG-INFO

@@ -0,0 +1,130 @@
+Metadata-Version: 2.4
+Name: lz4ext
+Version: 1.0
+Summary: Native LZ4 block compression bindings for CPython
+Author-email: Steeve Morin <steeve.morin@gmail.com>, Anton Stuk <sigman@ioupg.com>
+License-Expression: BSD-3-Clause AND BSD-2-Clause
+Project-URL: Homepage, https://github.com/sigman78/python-lz4
+Project-URL: Issues, https://github.com/sigman78/python-lz4/issues
+Classifier: Programming Language :: C
+Classifier: Programming Language :: Python :: 3 :: Only
+Classifier: Programming Language :: Python :: Implementation :: CPython
+Classifier: Programming Language :: Python :: 3.11
+Classifier: Programming Language :: Python :: 3.12
+Classifier: Programming Language :: Python :: 3.13
+Classifier: Programming Language :: Python :: 3.14
+Requires-Python: >=3.11
+Description-Content-Type: text/x-rst
+License-File: LICENSE
+License-File: src/LICENSE.lz4
+Dynamic: license-file
+
+=============
+python-lz4ext
+=============
+
+``lz4ext`` provides native LZ4 block compression for CPython 3.11–3.14,
+using vendored `LZ4 v1.10.0 <https://github.com/lz4/lz4/releases/tag/v1.10.0>`_.
+The distribution and import name are both ``lz4ext``.
+
+GitHub Actions tests/builds the supported Python matrix. See ``RELEASING.md``
+for release artifacts and the separate PyPI Trusted Publishing configuration,
+and ``SECURITY.md`` for vulnerability reporting.
+
+Install and develop
+===================
+
+Install the distribution::
+
+    python -m pip install lz4ext
+
+Build from this checkout (requires a C compiler, such as MSVC on Windows)::
+
+    python -m pip install .
+    python tests/test.py
+    python tests/bench.py --loops 1000
+
+Build distributable artifacts with an isolated PEP 517 backend::
+
+    python -m pip install build
+    python -m build
+
+The extension uses the CPython C API, retains the GIL, and does not declare
+free-threaded Python support. Builds use compiler defaults without
+``-march=native`` flags. No runtime Python dependencies are required.
+
+API
+===
+
+Input must be contiguous bytes-like data, including ``bytes``, ``bytearray``,
+and contiguous ``memoryview``. Text is rejected; encode it explicitly.
+All output is ``bytes``::
+
+    import lz4ext
+    data = b"hello" * 100
+    block = lz4ext.compress(data)
+    assert lz4ext.decompress(block) == data
+    assert lz4ext.loads(lz4ext.compressHC(data)) == data
+
+``compress(data)`` and ``compressHC(data)`` return a four-byte little-endian
+unsigned original size followed by an LZ4 block. This is the historical
+python-lz4ext format, not the LZ4 frame format.
+
+``decompress(data, *, max_output_size=67108864)`` validates framing and requires
+an exact match between the decoded length and header. ``max_output_size`` must
+be positive and at most ``INT_MAX`` (2,147,483,647). Declared output above the
+limit is rejected before allocation. To deliberately permit larger output,
+pass an explicit limit appropriate to the data and available memory.
+
+``compress_raw(data)`` returns a headerless block.
+``decompress_raw(data, output_size=0, *, max_output_size=67108864)`` allocates
+the supplied capacity, or twice the compressed input length when capacity is
+zero or omitted. Capacity must fit the output limit. High-ratio blocks usually
+need an explicit capacity::
+
+    raw = lz4ext.compress_raw(data)
+    assert lz4ext.decompress_raw(raw, len(data)) == data
+
+Malformed raw input and insufficient capacity both raise ``ValueError``;
+the decoder cannot distinguish them. There are no automatic allocation retries.
+Empty raw output is ``b'\x00'``; zero-length raw input is invalid.
+
+Compression rejects input above LZ4's supported maximum (2,113,929,216 bytes).
+Limits do not guarantee allocation success; ``MemoryError`` may still occur.
+Output caps bound the allocation for a call, not total process memory or
+cumulative work. This block format provides no checksum or authenticity check.
+
+Aliases and constants
+=====================
+
+``dumps`` and ``LZ4_compress`` behave like ``compress``.
+``loads``, ``uncompress``, and ``LZ4_uncompress`` behave like ``decompress``.
+``uncompress_raw`` behaves like ``decompress_raw``. All decoder aliases accept
+the keyword-only output limit.
+
+``VERSION`` and ``__version__`` contain the package version. ``LZ4_VERSION``
+reports the linked vendored library version. ``DEFAULT_MAX_OUTPUT_SIZE`` is
+64 MiB (67,108,864 bytes).
+
+Compatibility and migration
+===========================
+
+Version 1.0 requires Python 3.11 or newer and rejects implicit text input.
+Valid r119-era nonempty prefixed and raw blocks remain readable; compression
+may produce different bytes after the library upgrade. No frame API is added.
+Output above 64 MiB needs an explicit limit override. Incorrect size headers
+and detected malformed blocks now raise exceptions.
+
+Empty prefixed output is exactly five zero bytes (zero header plus canonical
+``00`` block). The exact four-byte zero-header legacy sentinel is accepted.
+Old releases could emit 20-byte empty records: a zero header followed by 16
+uninitialized bytes. These malformed records are rejected. If trusted records
+are known independently to represent empty data, migrate them offline to the
+canonical five-byte encoding. Arbitrary zero-header padding is not valid data.
+
+Licenses
+========
+
+The binding uses BSD 3-Clause in ``LICENSE``. The unmodified vendored LZ4
+sources use BSD 2-Clause; see ``src/LICENSE.lz4`` and ``src/README.lz4.md`` for
+upstream provenance. Both licenses accompany source and wheel distributions.

lz4ext-1.0/README.rst

@@ -0,0 +1,109 @@
+=============
+python-lz4ext
+=============
+
+``lz4ext`` provides native LZ4 block compression for CPython 3.11–3.14,
+using vendored `LZ4 v1.10.0 <https://github.com/lz4/lz4/releases/tag/v1.10.0>`_.
+The distribution and import name are both ``lz4ext``.
+
+GitHub Actions tests/builds the supported Python matrix. See ``RELEASING.md``
+for release artifacts and the separate PyPI Trusted Publishing configuration,
+and ``SECURITY.md`` for vulnerability reporting.
+
+Install and develop
+===================
+
+Install the distribution::
+
+    python -m pip install lz4ext
+
+Build from this checkout (requires a C compiler, such as MSVC on Windows)::
+
+    python -m pip install .
+    python tests/test.py
+    python tests/bench.py --loops 1000
+
+Build distributable artifacts with an isolated PEP 517 backend::
+
+    python -m pip install build
+    python -m build
+
+The extension uses the CPython C API, retains the GIL, and does not declare
+free-threaded Python support. Builds use compiler defaults without
+``-march=native`` flags. No runtime Python dependencies are required.
+
+API
+===
+
+Input must be contiguous bytes-like data, including ``bytes``, ``bytearray``,
+and contiguous ``memoryview``. Text is rejected; encode it explicitly.
+All output is ``bytes``::
+
+    import lz4ext
+    data = b"hello" * 100
+    block = lz4ext.compress(data)
+    assert lz4ext.decompress(block) == data
+    assert lz4ext.loads(lz4ext.compressHC(data)) == data
+
+``compress(data)`` and ``compressHC(data)`` return a four-byte little-endian
+unsigned original size followed by an LZ4 block. This is the historical
+python-lz4ext format, not the LZ4 frame format.
+
+``decompress(data, *, max_output_size=67108864)`` validates framing and requires
+an exact match between the decoded length and header. ``max_output_size`` must
+be positive and at most ``INT_MAX`` (2,147,483,647). Declared output above the
+limit is rejected before allocation. To deliberately permit larger output,
+pass an explicit limit appropriate to the data and available memory.
+
+``compress_raw(data)`` returns a headerless block.
+``decompress_raw(data, output_size=0, *, max_output_size=67108864)`` allocates
+the supplied capacity, or twice the compressed input length when capacity is
+zero or omitted. Capacity must fit the output limit. High-ratio blocks usually
+need an explicit capacity::
+
+    raw = lz4ext.compress_raw(data)
+    assert lz4ext.decompress_raw(raw, len(data)) == data
+
+Malformed raw input and insufficient capacity both raise ``ValueError``;
+the decoder cannot distinguish them. There are no automatic allocation retries.
+Empty raw output is ``b'\x00'``; zero-length raw input is invalid.
+
+Compression rejects input above LZ4's supported maximum (2,113,929,216 bytes).
+Limits do not guarantee allocation success; ``MemoryError`` may still occur.
+Output caps bound the allocation for a call, not total process memory or
+cumulative work. This block format provides no checksum or authenticity check.
+
+Aliases and constants
+=====================
+
+``dumps`` and ``LZ4_compress`` behave like ``compress``.
+``loads``, ``uncompress``, and ``LZ4_uncompress`` behave like ``decompress``.
+``uncompress_raw`` behaves like ``decompress_raw``. All decoder aliases accept
+the keyword-only output limit.
+
+``VERSION`` and ``__version__`` contain the package version. ``LZ4_VERSION``
+reports the linked vendored library version. ``DEFAULT_MAX_OUTPUT_SIZE`` is
+64 MiB (67,108,864 bytes).
+
+Compatibility and migration
+===========================
+
+Version 1.0 requires Python 3.11 or newer and rejects implicit text input.
+Valid r119-era nonempty prefixed and raw blocks remain readable; compression
+may produce different bytes after the library upgrade. No frame API is added.
+Output above 64 MiB needs an explicit limit override. Incorrect size headers
+and detected malformed blocks now raise exceptions.
+
+Empty prefixed output is exactly five zero bytes (zero header plus canonical
+``00`` block). The exact four-byte zero-header legacy sentinel is accepted.
+Old releases could emit 20-byte empty records: a zero header followed by 16
+uninitialized bytes. These malformed records are rejected. If trusted records
+are known independently to represent empty data, migrate them offline to the
+canonical five-byte encoding. Arbitrary zero-header padding is not valid data.
+
+Licenses
+========
+
+The binding uses BSD 3-Clause in ``LICENSE``. The unmodified vendored LZ4
+sources use BSD 2-Clause; see ``src/LICENSE.lz4`` and ``src/README.lz4.md`` for
+upstream provenance. Both licenses accompany source and wheel distributions.

lz4ext-1.0/RELEASING.md

@@ -0,0 +1,60 @@
+# Release configuration and procedure
+
+The workflow is `.github/workflows/ci.yml`. Pushes, pull requests, and manual
+runs test/build only. A **published GitHub release** starts a fresh build and
+can publish its tested artifacts to PyPI. Creating a tag or pushing a branch
+alone does not publish. The tag must exactly equal `v` plus the project version.
+
+The matrix tests CPython 3.11–3.14 on Linux, Windows, and macOS with
+`PYTHONMALLOC=debug`. cibuildwheel tests installed wheels on Linux x86_64
+(manylinux and musllinux), Windows AMD64, and macOS x86_64/arm64. It selects
+normal CPython builds, excluding free-threaded builds. The source distribution
+is installed and tested in a clean environment. Distribution artifacts from
+each workflow run are available for review; remote checks must actually run
+before platform compatibility is considered verified.
+
+## One-time external setup
+
+These settings must be configured by the repository/PyPI owner; adding this
+workflow does not create them or publish anything.
+
+1. Enable GitHub Actions in `sigman78/python-lz4` and create the GitHub
+   environment named exactly `pypi` under repository Settings → Environments.
+   Restrict deployment tags to release tags such as `v*`. Configure required
+   reviewers if the maintainers want an additional release approval gate.
+2. In the `lz4ext` project on PyPI, open Manage → Publishing and add a GitHub
+   Trusted Publisher with these exact values:
+   - Owner: `sigman78`
+   - Repository: `python-lz4`
+   - Workflow filename: `ci.yml` (filename only)
+   - Environment: `pypi`
+3. If the PyPI project does not exist or is not owned by the maintainer, resolve
+   ownership first or register a pending publisher using these same fields.
+   Do not substitute the unrelated `lz4` project. Forks must configure their
+   own owner/repository and PyPI project before enabling publication.
+
+See the [PyPI Trusted Publisher instructions](https://docs.pypi.org/trusted-publishers/adding-a-publisher/).
+No PyPI password or API token is needed. OIDC `id-token: write` is granted only
+to the isolated publish job, which does not check out or execute this source.
+It downloads only this run's `dist-*` artifacts after every required job passes.
+Build jobs have read-only repository permissions; pull requests do not publish.
+
+## Release
+
+1. Update `project.version` in `pyproject.toml`, the direct-build fallback in
+   `src/python-lz4.c`, the version test, and the changelog.
+2. Run `python -m build`, install the built wheel in a clean environment, run
+   `python tests/test.py`, and run `python -m twine check dist/*` (install `build`
+   and `twine` first). On PowerShell, enumerate artifacts for pip rather than
+   assuming shell expansion of `*.whl`.
+3. Push the reviewed commit and confirm the GitHub Actions test/build run passes.
+4. Create the matching tag, for example `v1.0` for version `1.0`, and publish
+   its GitHub release. Publishing the release is the action that enables PyPI
+   upload. The workflow rebuilds and tests artifacts for that tagged commit.
+5. Inspect the release run and PyPI files. Failed publishing should be diagnosed
+   before rerunning; existing files on PyPI cannot be replaced. The workflow
+   does not silently skip existing artifacts.
+
+Actions are pinned to verified upstream commit IDs with version comments.
+Review new action versions and update pins deliberately. No remote publication,
+tag creation, or external-account configuration is performed by local builds.

lz4ext-1.0/SECURITY.md

@@ -0,0 +1,12 @@
+# Reporting a vulnerability
+
+Avoid putting sensitive exploit details in a public issue. Check the repository's
+Security tab for a private vulnerability reporting option. Its availability
+depends on the repository owner's configuration; this file does not enable it.
+If unavailable, contact the maintainers privately using the contact details in
+`pyproject.toml` to arrange disclosure.
+
+Use the current release. Historical versions before this modernization contain
+known native-wrapper memory handling defects. The README describes decoder
+allocation limits and format compatibility. This block format does not provide
+integrity or authenticity; use application-level verification when needed.

lz4ext-1.0/lz4ext.egg-info/PKG-INFO

@@ -0,0 +1,130 @@
+Metadata-Version: 2.4
+Name: lz4ext
+Version: 1.0
+Summary: Native LZ4 block compression bindings for CPython
+Author-email: Steeve Morin <steeve.morin@gmail.com>, Anton Stuk <sigman@ioupg.com>
+License-Expression: BSD-3-Clause AND BSD-2-Clause
+Project-URL: Homepage, https://github.com/sigman78/python-lz4
+Project-URL: Issues, https://github.com/sigman78/python-lz4/issues
+Classifier: Programming Language :: C
+Classifier: Programming Language :: Python :: 3 :: Only
+Classifier: Programming Language :: Python :: Implementation :: CPython
+Classifier: Programming Language :: Python :: 3.11
+Classifier: Programming Language :: Python :: 3.12
+Classifier: Programming Language :: Python :: 3.13
+Classifier: Programming Language :: Python :: 3.14
+Requires-Python: >=3.11
+Description-Content-Type: text/x-rst
+License-File: LICENSE
+License-File: src/LICENSE.lz4
+Dynamic: license-file
+
+=============
+python-lz4ext
+=============
+
+``lz4ext`` provides native LZ4 block compression for CPython 3.11–3.14,
+using vendored `LZ4 v1.10.0 <https://github.com/lz4/lz4/releases/tag/v1.10.0>`_.
+The distribution and import name are both ``lz4ext``.
+
+GitHub Actions tests/builds the supported Python matrix. See ``RELEASING.md``
+for release artifacts and the separate PyPI Trusted Publishing configuration,
+and ``SECURITY.md`` for vulnerability reporting.
+
+Install and develop
+===================
+
+Install the distribution::
+
+    python -m pip install lz4ext
+
+Build from this checkout (requires a C compiler, such as MSVC on Windows)::
+
+    python -m pip install .
+    python tests/test.py
+    python tests/bench.py --loops 1000
+
+Build distributable artifacts with an isolated PEP 517 backend::
+
+    python -m pip install build
+    python -m build
+
+The extension uses the CPython C API, retains the GIL, and does not declare
+free-threaded Python support. Builds use compiler defaults without
+``-march=native`` flags. No runtime Python dependencies are required.
+
+API
+===
+
+Input must be contiguous bytes-like data, including ``bytes``, ``bytearray``,
+and contiguous ``memoryview``. Text is rejected; encode it explicitly.
+All output is ``bytes``::
+
+    import lz4ext
+    data = b"hello" * 100
+    block = lz4ext.compress(data)
+    assert lz4ext.decompress(block) == data
+    assert lz4ext.loads(lz4ext.compressHC(data)) == data
+
+``compress(data)`` and ``compressHC(data)`` return a four-byte little-endian
+unsigned original size followed by an LZ4 block. This is the historical
+python-lz4ext format, not the LZ4 frame format.
+
+``decompress(data, *, max_output_size=67108864)`` validates framing and requires
+an exact match between the decoded length and header. ``max_output_size`` must
+be positive and at most ``INT_MAX`` (2,147,483,647). Declared output above the
+limit is rejected before allocation. To deliberately permit larger output,
+pass an explicit limit appropriate to the data and available memory.
+
+``compress_raw(data)`` returns a headerless block.
+``decompress_raw(data, output_size=0, *, max_output_size=67108864)`` allocates
+the supplied capacity, or twice the compressed input length when capacity is
+zero or omitted. Capacity must fit the output limit. High-ratio blocks usually
+need an explicit capacity::
+
+    raw = lz4ext.compress_raw(data)
+    assert lz4ext.decompress_raw(raw, len(data)) == data
+
+Malformed raw input and insufficient capacity both raise ``ValueError``;
+the decoder cannot distinguish them. There are no automatic allocation retries.
+Empty raw output is ``b'\x00'``; zero-length raw input is invalid.
+
+Compression rejects input above LZ4's supported maximum (2,113,929,216 bytes).
+Limits do not guarantee allocation success; ``MemoryError`` may still occur.
+Output caps bound the allocation for a call, not total process memory or
+cumulative work. This block format provides no checksum or authenticity check.
+
+Aliases and constants
+=====================
+
+``dumps`` and ``LZ4_compress`` behave like ``compress``.
+``loads``, ``uncompress``, and ``LZ4_uncompress`` behave like ``decompress``.
+``uncompress_raw`` behaves like ``decompress_raw``. All decoder aliases accept
+the keyword-only output limit.
+
+``VERSION`` and ``__version__`` contain the package version. ``LZ4_VERSION``
+reports the linked vendored library version. ``DEFAULT_MAX_OUTPUT_SIZE`` is
+64 MiB (67,108,864 bytes).
+
+Compatibility and migration
+===========================
+
+Version 1.0 requires Python 3.11 or newer and rejects implicit text input.
+Valid r119-era nonempty prefixed and raw blocks remain readable; compression
+may produce different bytes after the library upgrade. No frame API is added.
+Output above 64 MiB needs an explicit limit override. Incorrect size headers
+and detected malformed blocks now raise exceptions.
+
+Empty prefixed output is exactly five zero bytes (zero header plus canonical
+``00`` block). The exact four-byte zero-header legacy sentinel is accepted.
+Old releases could emit 20-byte empty records: a zero header followed by 16
+uninitialized bytes. These malformed records are rejected. If trusted records
+are known independently to represent empty data, migrate them offline to the
+canonical five-byte encoding. Arbitrary zero-header padding is not valid data.
+
+Licenses
+========
+
+The binding uses BSD 3-Clause in ``LICENSE``. The unmodified vendored LZ4
+sources use BSD 2-Clause; see ``src/LICENSE.lz4`` and ``src/README.lz4.md`` for
+upstream provenance. Both licenses accompany source and wheel distributions.

lz4ext-1.0/lz4ext.egg-info/SOURCES.txt

@@ -0,0 +1,23 @@
+.gitattributes
+CHANGELOG.md
+LICENSE
+MANIFEST.in
+README.rst
+RELEASING.md
+SECURITY.md
+pyproject.toml
+setup.py
+lz4ext.egg-info/PKG-INFO
+lz4ext.egg-info/SOURCES.txt
+lz4ext.egg-info/dependency_links.txt
+lz4ext.egg-info/top_level.txt
+src/LICENSE.lz4
+src/README.lz4.md
+src/lz4.c
+src/lz4.h
+src/lz4hc.c
+src/lz4hc.h
+src/python-lz4.c
+src/python-lz4.h
+tests/bench.py
+tests/test.py

lz4ext-1.0/lz4ext.egg-info/dependency_links.txt

@@ -0,0 +1 @@
+

lz4ext-1.0/lz4ext.egg-info/top_level.txt

@@ -0,0 +1 @@
+lz4ext

lz4ext-1.0/pyproject.toml

@@ -0,0 +1,32 @@
+[build-system]
+requires = ["setuptools>=77.0.3"]
+build-backend = "setuptools.build_meta"
+
+[project]
+name = "lz4ext"
+version = "1.0"
+description = "Native LZ4 block compression bindings for CPython"
+readme = "README.rst"
+requires-python = ">=3.11"
+license = "BSD-3-Clause AND BSD-2-Clause"
+license-files = ["LICENSE", "src/LICENSE.lz4"]
+authors = [
+    {name = "Steeve Morin", email = "steeve.morin@gmail.com"},
+    {name = "Anton Stuk", email = "sigman@ioupg.com"},
+]
+classifiers = [
+    "Programming Language :: C",
+    "Programming Language :: Python :: 3 :: Only",
+    "Programming Language :: Python :: Implementation :: CPython",
+    "Programming Language :: Python :: 3.11",
+    "Programming Language :: Python :: 3.12",
+    "Programming Language :: Python :: 3.13",
+    "Programming Language :: Python :: 3.14",
+]
+
+[project.urls]
+Homepage = "https://github.com/sigman78/python-lz4"
+Issues = "https://github.com/sigman78/python-lz4/issues"
+
+[tool.setuptools]
+packages = []

lz4ext-1.0/setup.cfg

@@ -0,0 +1,4 @@
+[egg_info]
+tag_build = 
+tag_date = 0
+

lz4ext-1.0/setup.py

@@ -0,0 +1,9 @@
+from pathlib import Path
+import tomllib
+from setuptools import Extension, setup
+
+metadata = tomllib.loads(Path(__file__).with_name("pyproject.toml").read_text(encoding="utf-8"))
+setup(ext_modules=[Extension(
+    "lz4ext", ["src/python-lz4.c", "src/lz4.c", "src/lz4hc.c"],
+    define_macros=[("LZ4EXT_VERSION", '"%s"' % metadata["project"]["version"])],
+)])

lz4ext-1.0/src/LICENSE.lz4

@@ -0,0 +1,24 @@
+LZ4 Library
+Copyright (c) 2011-2020, Yann Collet
+All rights reserved.
+
+Redistribution and use in source and binary forms, with or without modification,
+are permitted provided that the following conditions are met:
+
+* Redistributions of source code must retain the above copyright notice, this
+  list of conditions and the following disclaimer.
+
+* Redistributions in binary form must reproduce the above copyright notice, this
+  list of conditions and the following disclaimer in the documentation and/or
+  other materials provided with the distribution.
+
+THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" AND
+ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED
+WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
+DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT HOLDER OR CONTRIBUTORS BE LIABLE FOR
+ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES
+(INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
+LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON
+ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+(INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
+SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.

lz4ext-1.0/src/README.lz4.md

@@ -0,0 +1,14 @@
+# Vendored LZ4
+
+`lz4.c`, `lz4.h`, `lz4hc.c`, and `lz4hc.h` are unmodified copies of
+the `lib/` files from the official LZ4 v1.10.0 release:
+https://github.com/lz4/lz4/releases/tag/v1.10.0
+
+Upstream commit: `ebb370ca83af193212df4dcbadcc5d87bc0de2f0`.
+
+Retrieved from https://raw.githubusercontent.com/lz4/lz4/v1.10.0/lib/
+on 2026-10-05. The upstream BSD-2-Clause license is in `LICENSE.lz4`.
+
+When updating these files, copy them verbatim from an official release,
+update this provenance and the library version in the README/changelog, and
+retain the license. The extension reports `LZ4_VERSION` from the linked library.