pyproject.toml (3164cd7) - fakesshagent.git

pyproject.toml

[build-system]
requires = ["hatchling"]
build-backend = "hatchling.build"

[project]
name = "fakesshagent"
description = "A non-hardened implementation of the SSH agent protocol, for use in testing."
readme = "README.md"
version = "0.1"
requires-python = ">= 3.9"
license = "Zlib"
keywords = []
authors = [
  { name = "Marco Ricci", email = "software@the13thletter.info" },
]
classifiers = [
    "Development Status :: 4 - Beta",
    "Environment :: Console",
    "Operating System :: OS Independent",
    "Programming Language :: Python :: 3",
    "Programming Language :: Python :: 3.9",
    "Programming Language :: Python :: 3.10",
    "Programming Language :: Python :: 3.11",
    "Programming Language :: Python :: 3.12",
    "Programming Language :: Python :: 3.13",
    "Programming Language :: Python :: 3.14",
    "Programming Language :: Python :: 3.15",
    "Programming Language :: Python :: Implementation :: CPython",
    "Programming Language :: Python :: Implementation :: PyPy",
    "Topic :: Software Development :: Testing",
    "Typing :: Typed",
]
dependencies = [
    "derivepassphrase-sshagentsocketprovider >= 1.0",
    "typing_extensions",
]

[project.optional-dependencies]
dev = [
    # dev - developers' setup
    #
    # Development uses the hatch build system, to isolate all tools in
    # their own virtual environment...
    "hatch >= 1.10",
    # ...but it is still sensible to have most static analysis tools
    # available in the main virtual environment as well, so that your
    # editor/IDE setup can access them as well.
    "mypy >= 1.0",
    "ruff >= 0.16.0",
    # Test dependencies should be installed as well, to have static
    # analysis in the test suite working too.
    #
    # `packaging` is needed because the test suite compares version
    # numbers.
    #
    # `pytest-xdist` is used by `hatch test -p`, but for the `psutil`
    # extra, the `psutil` package provides no wheels for PyPy.  Users on
    # The Annoying OS are especially screwed if they lack the admin
    # priviledges necessary to install the developers' tools.  So skip
    # the extra, and accept that for PyPy environments, `hatch test -p`
    # will not be able to distinguish logical and physical CPU core
    # counts.
    #
    # We use `hypothesis.is_hypothesis_test`, which was introduced in
    # 6.131.0, and `exc_type` for `pytest.importorskip`, which was
    # introduced in 8.2.0.
    "coverage[toml] >= 7.4",
    "coverage-enable-subprocess >= 1.0",
    "hypothesis >= 6.131.0",
    "packaging",
    "pytest >= 8.2",
    "pytest-randomly >= 3.15",
    "pytest-xdist >= 3.6.0",
    'pytest-xdist[psutil] >= 3.6.0; platform_python_implementation != "PyPy"',
    # `pytest` 8.4 and higher, and `hypothesis` versions 6.130.13
    # through 6.135.1, use syntax only properly available on Python 3.10
    # and later: `with (x as y, a as b)`.  However, those versions are
    # not marked as incompatible with Python 3.9.
    'pytest < 8.4 ; python_version < "3.10"',
    'hypothesis != 6.130.13, != 6.131.*, != 6.132.*, != 6.133.*, != 6.134.*, != 6.135.0, != 6.135.1 ; python_version < "3.10"',
    # `hypothesis` requires a Rust toolchain since version 6.156, and
    # publishes wheels for PyPy 3.11 only. PyO3 version 0.27 also dropped
    # support for PyPy 3.9 and 3.10, and cannot be overridden within
    # pyproject.toml.
    'hypothesis < 6.156 ; platform_python_implementation == "PyPy" and python_version < "3.11"',
]

[project.urls]
Documentation = "https://the13thletter.info/fakesshagent/"
Issues = "https://the13thletter.info/fakesshagent/latest/wishlist/"
Source = "https://git.schokokeks.org/fakesshagent.git"

[project.entry-points."derivepassphrase.ssh_agent_socket_providers"]
stub_agent = "fakesshagent.machinery: STUB_AGENT_ENTRY_POINT"
stub_agent_with_address = "fakesshagent.machinery: STUB_AGENT_WITH_ADDRESS_ENTRY_POINT"
stub_agent_with_address_and_deterministic_dsa = "fakesshagent.machinery: STUB_AGENT_WITH_ADDRESS_AND_DETERMINISTIC_DSA_ENTRY_POINT"

[tool.coverage.html]
directory = "html/coverage"

[tool.coverage.paths]
src = ["src"]
tests = ["tests"]

[tool.coverage.report]
skip_covered = false
skip_empty = true
precision = 3
partial_branches = [
    'pragma: no branch',
]
exclude_also = [
    "if __name__ == .__main__.:",
    'if (?:typing\.)?TYPE_CHECKING:',
    "raise AssertionError",
    "raise NotImplementedError",
    'assert False',
    '(?:typing\.)?assert_never$',
    '@overload',
    'class .*\(Protocol$:',
    'pytest\.fail\(',
    '@(?:(?:pytest\.)?mark\.)?xfail\(',
    'class _Omitted:',
    'class _DummyModule:',
]

# We use a bunch of "codes", similar to the "type: ignore" comments, as
# shorthand notation for why we are excluding a certain piece of code
# manually from coverage, instead of writing a verbose explanation at each
# call site:
#
# "external-api" and "unused": The code is defined by an external API not
# under our control, or it implements an external interface, and some of the
# branches of that code we do not yet use ("unused") or don't ever intend to
# use ("external-api").  In particular, in production and in testing use, we
# do not expect to hit these code branches unless there are errors in our
# code.
#
# "external": Whether or not we hit this code branch is non-deterministic,
# and dependent on external factors that are difficult or impossible for us
# to control, predict, or simulate.  We therefore categorically exclude all
# branches from coverage.  (Typically though, one particular branch *will*
# actually run.  We just don't know ahead of time which one.)  (While this
# applies to "choice of operating system" and "installed Python/library
# version" as well, these cases are usually tagged with more specific codes,
# or explained in long-form commentary.)
#
# "failsafe": Akin to assertions, these are used to catch "impossible"
# situations and to still fail in a controlled manner.  We do not expect to
# hit these code branches during production or testing.
#
# "debug", "internal" and "interactive": These code branches are not
# intended to be called by the general public ("debug", "internal"), or they
# are only intended to be called interactively ("interactive").  We do not
# expect to hit these code branches during production or testing.

[tool.coverage.run]
source_pkgs = ["fakesshagent", "tests"]
branch = true
parallel = true
omit = [
  "__main__.py",
]
dynamic_context = 'test_function'
concurrency = ['thread', 'multiprocessing']
sigterm = true
relative_files = true

[tool.hatch.build.targets.sdist]
exclude = [
    'docs/changelog.d/*.md',
]

[tool.hatch.env]
requires = [
    "hatch-mkdocs",
]

[tool.hatch.env.collectors.mkdocs.docs]
path = "mkdocs.yml"

[tool.hatch.envs.docs]
extra-dependencies = [
    # Our documentation uses formatted function signatures (i.e. with
    # formatted type annotations), which requires `black`.
    "black",
    # We actually use functionality from MkDocs 1.6, so ensure the
    # version is at least this high.
    "mkdocs >= 1.6",
    # Our changelog is assembled from singular entries, orchestrated by
    # `scriv`.
    "scriv >= 1.4",
]
detached = false

[tool.hatch.envs.hatch-check-code]
config-path = "/dev/null"
dependencies = [
    # Override the default dependencies, lest the ruff requirements
    # conflict.
]
features = [
    "dev",
]
[tool.hatch.envs.hatch-check-fmt]
config-path = "/dev/null"
dependencies = [
    # Override the default dependencies, lest the ruff requirements
    # conflict.
]
features = [
    "dev",
]
[tool.hatch.envs.hatch-check-types]
config-path = "/dev/null"
dependencies = [
    # Override the default dependencies, lest the ruff requirements
    # conflict.
    "pyrefly >= 1.0.0",
]
features = [
    "dev",
]

[tool.hatch.envs.hatch-test]
default-args = ['src', 'tests']
dependencies = [
    "coverage[toml] >= 7.4",
    "coverage-enable-subprocess >= 1.0",
    "hypothesis >= 6.131.0",
    "packaging",
    "pytest >= 8.2",
    "pytest-randomly >= 3.15",
    "pytest-xdist >= 3.6.0",
    'pytest-xdist[psutil] >= 3.6.0; platform_python_implementation != "PyPy"',
    'pytest < 8.4 ; python_version < "3.10"',
    'hypothesis != 6.130.13, != 6.131.*, != 6.132.*, != 6.133.*, != 6.134.*, != 6.135.0, != 6.135.1 ; python_version < "3.10"',
    'hypothesis < 6.156 ; platform_python_implementation == "PyPy" and python_version < "3.11"',
]
matrix-name-format = '{variable}_{value}'

[[tool.hatch.envs.hatch-test.matrix]]
python = ["3.15", "3.14", "3.13", "3.12", "3.11", "3.10", "3.9", "pypy3.11", "pypy3.10", "pypy3.9"]
parser-version = ["PEG"]

[[tool.hatch.envs.hatch-test.matrix]]
python = ["3.9", "pypy3.9"]
parser-version = ["LL1"]

[tool.hatch.envs.hatch-test.overrides]
matrix.parser-version.env-vars = [
    { key = "PYTHONOLDPARSER", value = "1", if = ["LL1"] },
]

[tool.hatch.envs.hatch-test.scripts]
run = "pytest --hypothesis-profile={env:HYPOTHESIS_PROFILE:default}{env:HATCH_TEST_ARGS:} {args}"
run-cov = "coverage run -m pytest --hypothesis-profile={env:HYPOTHESIS_PROFILE:default}{env:HATCH_TEST_ARGS:} {args}"
cov-combine = "coverage combine"
cov-report = "coverage report"

[tool.hatch.envs.types]
extra-dependencies = [
  "mypy>=1.0.0",
]
[tool.hatch.envs.types.scripts]
check = "mypy --install-types --non-interactive {args:src/fakesshagent tests}"

[tool.mypy]
files = ['src/**/*.py', 'tests/**/*.py']
mypy_path = '$MYPY_CONFIG_FILE_DIR/src:$MYPY_CONFIG_FILE_DIR/other-stubs'
explicit_package_bases = true
implicit_reexport = false
sqlite_cache = true
enable_error_code = ['ignore-without-code']

[tool.pyrefly]
project-includes = [
    "src/**/*.py",
    "tests/**/*.py",
]
search-path = [
    "src",
    ".",
]
preset = "legacy"
infer-return-types = "never"
permissive-ignores = true

[tool.pyrefly.errors]
unnecessary-type-conversion = "ignore"

[tool.pytest.ini_options]
addopts = '--doctest-modules --dist=loadgroup --import-mode=importlib'
pythonpath = [
    'tests', # needed for re-importing test machinery with multiprocessing (spawn)
]
testpaths = [
    'src', # doctests in API documentation
    'tests', # other tests
]
xfail_strict = true

[tool.ruff]
line-length = 79
src = ["src"]

[tool.ruff.format]
docstring-code-format = true
docstring-code-line-length = "dynamic"
preview = true
quote-style = "double"

[tool.ruff.lint]
ignore = [
    # Suggested ignore by ruff when also using ruff to format.  We *do*
    # check for E501, because this usually only happens when there is
    # a text string that should be manually broken.  However, for
    # automated quality control, that specific check is turned off so
    # that the linting and formatting can proceed, and then *afterwards*
    # the files are re-linted.  We hope that any *true* E501 errors left
    # over from this *are* text strings that should be manually broken.
    'tab-indentation', 'indentation-with-invalid-multiple',
    'indentation-with-invalid-multiple-comment', 'over-indented',
    'docstring-tab-indentation', 'triple-single-quotes',
    'bad-quotes-inline-string', 'bad-quotes-multiline-string',
    'bad-quotes-docstring', 'avoidable-escaped-quote',
    'missing-trailing-comma', 'prohibited-trailing-comma',
    'single-line-implicit-string-concatenation',
    'multi-line-implicit-string-concatenation',
    # We use `assert` regularly to appease the type checker, and because
    # it is the right language tool for this job.
    'assert',
    # The formatter takes care of trailing commas and docstring code
    # automatically.
    'missing-trailing-comma', 'doc-line-too-long',
    # We document transitive exceptions as well (if we feel they would
    # be surprising to the user otherwise).
    'docstring-extraneous-exception',
    # We currently don't have issues for every TODO.  Forcing an issue
    # also goes against the philosophy of TODOs as low-overhead markers
    # for future work; see
    # https://gist.github.com/dmnd/ed5d8ef8de2e4cfea174bd5dafcda382 .
    'missing-todo-link',
    # We somewhat regularly use loops where each iteration needs
    # a separate try-except block.
    'try-except-in-loop',
    # We do not currently use pathlib.  The PTH rules are unselected,
    # but FURB includes several pathlib-related rules.
    'read-whole-file', 'write-whole-file',
    # We catch type-ignore comments without specific code via the mypy
    # configuration, not via ruff.
    'blanket-type-ignore',
    # Provisionally ignore the `__init__.py` rule, until we can adapt our
    # codebase.
    'non-empty-init-module',
    # We prefer noqa comments over ruff-specific ignores, for tool
    # compatibility reasons, even if the comments themselves become harder
    # to read. Using them in the ruff config section is fine thought,
    # because that *is* ruff-specific.
    'noqa-comments',
    # Still undecided about this one.
    'too-many-statements-in-try-clause',
]
preview = true
# We select here in the order of presentation on the ruff documentation
# website.  ruff default selection (v0.6.2) is merely E4, E7, E9 and F.
select = [
    'F', 'E', 'W', 'C90', 'I', 'N', 'D', 'UP', 'YTT',
    'ANN', 'ASYNC', 'S', 'BLE', 'FBT', 'B', 'A', 'COM',
    'CPY', 'C4', 'DTZ', 'T10', 'DJ', 'EM', 'EXE', 'FA',
    'ISC', 'ICN', 'LOG', 'G', 'INP', 'PIE', 'T20', 'PYI',
    'PT', 'Q', 'RET', 'SLF', 'SLOT', 'SIM', 'TID', 'TC',
    'INT', 'ARG', 'PTH', 'TD',
    # We use TODOs and FIXMEs as notes for later, and don't want the
    # linter to nag about every occurrence.  Disable 'FIX'.
    #
    # The "eradicate" rule is prone to a lot of false positives, and it
    # is unclear to me, and probably confusing to read, where to apply
    # a noqa marker.  Instead, disable 'ERA', and if necessary, specify
    # it on the command-line.
    'PD', 'PGH', 'PL', 'TRY', 'FLY', 'NPY', 'FAST',
    'AIR', 'PERF', 'FURB', 'DOC', 'RUF',
]

[tool.ruff.lint.per-file-ignores]
"**/scripts/*" = [
    # Suggested by hatch.
    'INP',
    # Suggested by hatch.
    'T20',
]
"**/tests/**/*" = [
    # Suggested by hatch, assumingly because it may be important to
    # verify that the value is exactly the empty string, and not just
    # any falsy value.
    'compare-to-empty-string',
    # Suggested by hatch, assumingly because tests may use "magic
    # values".
    'magic-value-comparison',
    # Suggested by hatch, because tests are typically organized as
    # classes and instance methods but may not really be using the
    # `self` argument.
    'no-self-use',
    # Suggested by hatch, because these warnings may be precisely what
    # the tests are supposed to test.
    'S',
    # Suggested by hatch, because pytest-style tests conventionally
    # import code from each other via relative imports.
    'relative-imports',
    # Our tests regularly use arguments named `input` to store an input
    # (text-/byte-)string.
    'builtin-argument-shadowing',
    # We regularly annotate pytest fixtures like monkeypatch as `Any`.
    'any-type',
    # Our tests generally don't contain docstrings.
    'D', 'DOC',
    # Our tests are regularly parametrized with booleans, for benign
    # purposes.
    'FBT',
    # One of our standard modules is called `derivepassphrase._types`.
    # Importing this from the tests directory would then automatically
    # trigger `PLC2701`.
    'import-private-name',
    # Too many public methods/arguments/returns/branches/locals doesn't
    # really apply here.
    'too-many-public-methods', 'too-many-return-statements',
    'too-many-branches', 'too-many-arguments', 'too-many-locals',
    'too-many-statements', 'too-many-boolean-expressions',
    'too-many-positional-arguments',
    # To fully test the `derivepassphrase.cli` module (and a couple
    # other things), we need to call and to mock several internal
    # functions, which would automatically trigger `SLF001`.
    'private-member-access',
    # pytest does not support sensible introspection of
    # `assert all(...)` expressions in tests the same way it supports
    # introspection in `asssert all([...])`.  So the extra list
    # comprehension actually improves debuggability in this case.
    'unnecessary-comprehension-in-call',
    # The tests sometimes include long strings (in non-Python formats)
    # that should be included verbatim, without artificial line
    # breaking, so they can be grepped for.
    'line-too-long',
]

[tool.ruff.lint.flake8-copyright]
# Include hatch-enforced SPDX-FileCopyrightText in check.
notice-rgx = '(?i)(?:Copyright\s+((?:$C$|©)\s+)?|SPDX-FileCopyrightText:\s+)\d{4}((-|,\s)\d{4})*'

[tool.ruff.lint.flake8-pytest-style]
fixture-parentheses = false
mark-parentheses = false
parametrize-names-type = 'list'

[tool.ruff.lint.flake8-tidy-imports]
ban-relative-imports = "all"

[tool.ruff.lint.isort]
known-first-party = [
    "derivepassphrase_sshagentsocketprovider",
    "fakesshagent",
]

[tool.ruff.lint.pycodestyle]
ignore-overlong-task-comments = true  # for E501
max-doc-length = 72  # for W505

[tool.ruff.lint.pydocstyle]
convention = 'google'

[tool.scriv]
version = "literal: pyproject.toml: project.version"
format = "md"
fragment_directory = "docs/changelog.d"
output_file = "docs/changelog.md"
insert_marker = ""
end_marker = ""
md_header_level = "2"
entry_title_template = "{% if version %}{{ version }} ({% endif %}{{ date.strftime('%Y-%m-%d') }}{% if version %}){% endif %}"

fakesshagent.git

Release 0.1