omlish 0.0.0.dev122__py3-none-any.whl → 0.0.0.dev124__py3-none-any.whl

omlish/lite/http/parsing.py

@@ -0,0 +1,376 @@
+# ruff: noqa: UP006 UP007
+import abc
+import http.client
+import http.server
+import io
+import typing as ta
+
+from .versions import HttpProtocolVersion
+from .versions import HttpProtocolVersions
+
+
+T = ta.TypeVar('T')
+
+
+HttpHeaders = http.client.HTTPMessage  # ta.TypeAlias
+
+
+##
+
+
+class ParseHttpRequestResult(abc.ABC):  # noqa
+    __slots__ = (
+        'server_version',
+        'request_line',
+        'request_version',
+        'version',
+        'headers',
+        'close_connection',
+    )
+
+    def __init__(
+            self,
+            *,
+            server_version: HttpProtocolVersion,
+            request_line: str,
+            request_version: HttpProtocolVersion,
+            version: HttpProtocolVersion,
+            headers: ta.Optional[HttpHeaders],
+            close_connection: bool,
+    ) -> None:
+        super().__init__()
+
+        self.server_version = server_version
+        self.request_line = request_line
+        self.request_version = request_version
+        self.version = version
+        self.headers = headers
+        self.close_connection = close_connection
+
+    def __repr__(self) -> str:
+        return f'{self.__class__.__name__}({", ".join(f"{a}={getattr(self, a)!r}" for a in self.__slots__)})'
+
+
+class EmptyParsedHttpResult(ParseHttpRequestResult):
+    pass
+
+
+class ParseHttpRequestError(ParseHttpRequestResult):
+    __slots__ = (
+        'code',
+        'message',
+        *ParseHttpRequestResult.__slots__,
+    )
+
+    def __init__(
+            self,
+            *,
+            code: http.HTTPStatus,
+            message: ta.Union[str, ta.Tuple[str, str]],
+
+            **kwargs: ta.Any,
+    ) -> None:
+        super().__init__(**kwargs)
+
+        self.code = code
+        self.message = message
+
+
+class ParsedHttpRequest(ParseHttpRequestResult):
+    __slots__ = (
+        'method',
+        'path',
+        'headers',
+        'expects_continue',
+        *[a for a in ParseHttpRequestResult.__slots__ if a != 'headers'],
+    )
+
+    def __init__(
+            self,
+            *,
+            method: str,
+            path: str,
+            headers: HttpHeaders,
+            expects_continue: bool,
+
+            **kwargs: ta.Any,
+    ) -> None:
+        super().__init__(
+            headers=headers,
+            **kwargs,
+        )
+
+        self.method = method
+        self.path = path
+        self.expects_continue = expects_continue
+
+    headers: HttpHeaders
+
+
+#
+
+
+class HttpRequestParser:
+    DEFAULT_SERVER_VERSION = HttpProtocolVersions.HTTP_1_0
+
+    # The default request version. This only affects responses up until the point where the request line is parsed, so
+    # it mainly decides what the client gets back when sending a malformed request line.
+    # Most web servers default to HTTP 0.9, i.e. don't send a status line.
+    DEFAULT_REQUEST_VERSION = HttpProtocolVersions.HTTP_0_9
+
+    #
+
+    DEFAULT_MAX_LINE: int = 0x10000
+    DEFAULT_MAX_HEADERS: int = 100
+
+    #
+
+    def __init__(
+            self,
+            *,
+            server_version: HttpProtocolVersion = DEFAULT_SERVER_VERSION,
+
+            max_line: int = DEFAULT_MAX_LINE,
+            max_headers: int = DEFAULT_MAX_HEADERS,
+    ) -> None:
+        super().__init__()
+
+        if server_version >= HttpProtocolVersions.HTTP_2_0:
+            raise ValueError(f'Unsupported protocol version: {server_version}')
+        self._server_version = server_version
+
+        self._max_line = max_line
+        self._max_headers = max_headers
+
+    #
+
+    @property
+    def server_version(self) -> HttpProtocolVersion:
+        return self._server_version
+
+    #
+
+    def _run_read_line_coro(
+            self,
+            gen: ta.Generator[int, bytes, T],
+            read_line: ta.Callable[[int], bytes],
+    ) -> T:
+        sz = next(gen)
+        while True:
+            try:
+                sz = gen.send(read_line(sz))
+            except StopIteration as e:
+                return e.value
+
+    #
+
+    def parse_request_version(self, version_str: str) -> HttpProtocolVersion:
+        if not version_str.startswith('HTTP/'):
+            raise ValueError(version_str)  # noqa
+
+        base_version_number = version_str.split('/', 1)[1]
+        version_number_parts = base_version_number.split('.')
+
+        # RFC 2145 section 3.1 says there can be only one "." and
+        #   - major and minor numbers MUST be treated as separate integers;
+        #   - HTTP/2.4 is a lower version than HTTP/2.13, which in turn is lower than HTTP/12.3;
+        #   - Leading zeros MUST be ignored by recipients.
+        if len(version_number_parts) != 2:
+            raise ValueError(version_number_parts)  # noqa
+        if any(not component.isdigit() for component in version_number_parts):
+            raise ValueError('non digit in http version')  # noqa
+        if any(len(component) > 10 for component in version_number_parts):
+            raise ValueError('unreasonable length http version')  # noqa
+
+        return HttpProtocolVersion(
+            int(version_number_parts[0]),
+            int(version_number_parts[1]),
+        )
+
+    #
+
+    def coro_read_raw_headers(self) -> ta.Generator[int, bytes, ta.List[bytes]]:
+        raw_headers: ta.List[bytes] = []
+        while True:
+            line = yield self._max_line + 1
+            if len(line) > self._max_line:
+                raise http.client.LineTooLong('header line')
+            raw_headers.append(line)
+            if len(raw_headers) > self._max_headers:
+                raise http.client.HTTPException(f'got more than {self._max_headers} headers')
+            if line in (b'\r\n', b'\n', b''):
+                break
+        return raw_headers
+
+    def read_raw_headers(self, read_line: ta.Callable[[int], bytes]) -> ta.List[bytes]:
+        return self._run_read_line_coro(self.coro_read_raw_headers(), read_line)
+
+    def parse_raw_headers(self, raw_headers: ta.Sequence[bytes]) -> HttpHeaders:
+        return http.client.parse_headers(io.BytesIO(b''.join(raw_headers)))
+
+    #
+
+    def coro_parse(self) -> ta.Generator[int, bytes, ParseHttpRequestResult]:
+        raw_request_line = yield self._max_line + 1
+
+        # Common result kwargs
+
+        request_line = '-'
+        request_version = self.DEFAULT_REQUEST_VERSION
+
+        # Set to min(server, request) when it gets that far, but if it fails before that the server authoritatively
+        # responds with its own version.
+        version = self._server_version
+
+        headers: HttpHeaders | None = None
+
+        close_connection = True
+
+        def result_kwargs():
+            return dict(
+                server_version=self._server_version,
+                request_line=request_line,
+                request_version=request_version,
+                version=version,
+                headers=headers,
+                close_connection=close_connection,
+            )
+
+        # Decode line
+
+        if len(raw_request_line) > self._max_line:
+            return ParseHttpRequestError(
+                code=http.HTTPStatus.REQUEST_URI_TOO_LONG,
+                message='Request line too long',
+                **result_kwargs(),
+            )
+
+        if not raw_request_line:
+            return EmptyParsedHttpResult(**result_kwargs())
+
+        request_line = raw_request_line.decode('iso-8859-1').rstrip('\r\n')
+
+        # Split words
+
+        words = request_line.split()
+        if len(words) == 0:
+            return EmptyParsedHttpResult(**result_kwargs())
+
+        # Parse and set version
+
+        if len(words) >= 3:  # Enough to determine protocol version
+            version_str = words[-1]
+            try:
+                request_version = self.parse_request_version(version_str)
+
+            except (ValueError, IndexError):
+                return ParseHttpRequestError(
+                    code=http.HTTPStatus.BAD_REQUEST,
+                    message=f'Bad request version ({version_str!r})',
+                    **result_kwargs(),
+                )
+
+            if (
+                    request_version < HttpProtocolVersions.HTTP_0_9 or
+                    request_version >= HttpProtocolVersions.HTTP_2_0
+            ):
+                return ParseHttpRequestError(
+                    code=http.HTTPStatus.HTTP_VERSION_NOT_SUPPORTED,
+                    message=f'Invalid HTTP version ({version_str})',
+                    **result_kwargs(),
+                )
+
+            version = min([self._server_version, request_version])
+
+            if version >= HttpProtocolVersions.HTTP_1_1:
+                close_connection = False
+
+        # Verify word count
+
+        if not 2 <= len(words) <= 3:
+            return ParseHttpRequestError(
+                code=http.HTTPStatus.BAD_REQUEST,
+                message=f'Bad request syntax ({request_line!r})',
+                **result_kwargs(),
+            )
+
+        # Parse method and path
+
+        method, path = words[:2]
+        if len(words) == 2:
+            close_connection = True
+            if method != 'GET':
+                return ParseHttpRequestError(
+                    code=http.HTTPStatus.BAD_REQUEST,
+                    message=f'Bad HTTP/0.9 request type ({method!r})',
+                    **result_kwargs(),
+                )
+
+        # gh-87389: The purpose of replacing '//' with '/' is to protect against open redirect attacks possibly
+        # triggered if the path starts with '//' because http clients treat //path as an absolute URI without scheme
+        # (similar to http://path) rather than a path.
+        if path.startswith('//'):
+            path = '/' + path.lstrip('/')  # Reduce to a single /
+
+        # Parse headers
+
+        try:
+            raw_gen = self.coro_read_raw_headers()
+            raw_sz = next(raw_gen)
+            while True:
+                buf = yield raw_sz
+                try:
+                    raw_sz = raw_gen.send(buf)
+                except StopIteration as e:
+                    raw_headers = e.value
+                    break
+
+            headers = self.parse_raw_headers(raw_headers)
+
+        except http.client.LineTooLong as err:
+            return ParseHttpRequestError(
+                code=http.HTTPStatus.REQUEST_HEADER_FIELDS_TOO_LARGE,
+                message=('Line too long', str(err)),
+                **result_kwargs(),
+            )
+
+        except http.client.HTTPException as err:
+            return ParseHttpRequestError(
+                code=http.HTTPStatus.REQUEST_HEADER_FIELDS_TOO_LARGE,
+                message=('Too many headers', str(err)),
+                **result_kwargs(),
+            )
+
+        # Check for connection directive
+
+        conn_type = headers.get('Connection', '')
+        if conn_type.lower() == 'close':
+            close_connection = True
+        elif (
+                conn_type.lower() == 'keep-alive' and
+                version >= HttpProtocolVersions.HTTP_1_1
+        ):
+            close_connection = False
+
+        # Check for expect directive
+
+        expect = headers.get('Expect', '')
+        if (
+                expect.lower() == '100-continue' and
+                version >= HttpProtocolVersions.HTTP_1_1
+        ):
+            expects_continue = True
+        else:
+            expects_continue = False
+
+        # Return
+
+        return ParsedHttpRequest(
+            method=method,
+            path=path,
+            expects_continue=expects_continue,
+            **result_kwargs(),
+        )
+
+    def parse(self, read_line: ta.Callable[[int], bytes]) -> ParseHttpRequestResult:
+        return self._run_read_line_coro(self.coro_parse(), read_line)

omlish/lite/http/versions.py

@@ -0,0 +1,17 @@
+# ruff: noqa: UP006 UP007
+import typing as ta
+
+
+class HttpProtocolVersion(ta.NamedTuple):
+    major: int
+    minor: int
+
+    def __str__(self) -> str:
+        return f'HTTP/{self.major}.{self.minor}'
+
+
+class HttpProtocolVersions:
+    HTTP_0_9 = HttpProtocolVersion(0, 9)
+    HTTP_1_0 = HttpProtocolVersion(1, 0)
+    HTTP_1_1 = HttpProtocolVersion(1, 1)
+    HTTP_2_0 = HttpProtocolVersion(2, 0)

omlish/lite/inject.py

@@ -30,12 +30,22 @@ InjectorBindingOrBindings = ta.Union['InjectorBinding', 'InjectorBindings']
 
 
 @dc.dataclass(frozen=True)
-class InjectorKey:
+class InjectorKey(ta.Generic[T]):
     cls: InjectorKeyCls
     tag: ta.Any = None
     array: bool = False
 
 
+def is_valid_injector_key_cls(cls: ta.Any) -> bool:
+    return isinstance(cls, type) or is_new_type(cls)
+
+
+def check_valid_injector_key_cls(cls: T) -> T:
+    if not is_valid_injector_key_cls(cls):
+        raise TypeError(cls)
+    return cls
+
+
 ##
 
 
@@ -79,6 +89,12 @@ class Injector(abc.ABC):
     def inject(self, obj: ta.Any) -> ta.Any:
         raise NotImplementedError
 
+    def __getitem__(
+            self,
+            target: ta.Union[InjectorKey[T], ta.Type[T]],
+    ) -> T:
+        return self.provide(target)
+
 
 ###
 # exceptions
@@ -111,7 +127,7 @@ def as_injector_key(o: ta.Any) -> InjectorKey:
         raise TypeError(o)
     if isinstance(o, InjectorKey):
         return o
-    if isinstance(o, type) or is_new_type(o):
+    if is_valid_injector_key_cls(o):
         return InjectorKey(o)
     raise TypeError(o)
 
@@ -443,8 +459,8 @@ class InjectorBinder:
             to_fn = obj
             if key is None:
                 sig = _injection_signature(obj)
-                ty = check_isinstance(sig.return_annotation, type)
-                key = InjectorKey(ty)
+                key_cls = check_valid_injector_key_cls(sig.return_annotation)
+                key = InjectorKey(key_cls)
         else:
             if to_const is not None:
                 raise TypeError('Cannot bind instance with to_const')
@@ -498,7 +514,7 @@ class InjectorBinder:
 # injector
 
 
-_INJECTOR_INJECTOR_KEY = InjectorKey(Injector)
+_INJECTOR_INJECTOR_KEY: InjectorKey[Injector] = InjectorKey(Injector)
 
 
 class _Injector(Injector):

omlish/lite/journald.py

@@ -29,7 +29,6 @@ sd_iovec._fields_ = [
 def sd_libsystemd() -> ta.Any:
     lib = ct.CDLL('libsystemd.so.0')
 
-    lib.sd_journal_sendv = lib['sd_journal_sendv']  # type: ignore
     lib.sd_journal_sendv.restype = ct.c_int
     lib.sd_journal_sendv.argtypes = [ct.POINTER(sd_iovec), ct.c_int]
 

omlish/lite/runtime.py

@@ -14,5 +14,4 @@ REQUIRED_PYTHON_VERSION = (3, 8)
 
 def check_runtime_version() -> None:
     if sys.version_info < REQUIRED_PYTHON_VERSION:
-        raise OSError(
-            f'Requires python {REQUIRED_PYTHON_VERSION}, got {sys.version_info} from {sys.executable}')  # noqa
+        raise OSError(f'Requires python {REQUIRED_PYTHON_VERSION}, got {sys.version_info} from {sys.executable}')  # noqa

omlish/lite/socket.py

@@ -0,0 +1,77 @@
+# ruff: noqa: UP006 UP007
+"""
+TODO:
+ - SocketClientAddress family / tuple pairs
+  + codification of https://docs.python.org/3/library/socket.html#socket-families
+"""
+import abc
+import dataclasses as dc
+import socket
+import typing as ta
+
+
+SocketAddress = ta.Any
+
+
+SocketHandlerFactory = ta.Callable[[SocketAddress, ta.BinaryIO, ta.BinaryIO], 'SocketHandler']
+
+
+##
+
+
+@dc.dataclass(frozen=True)
+class SocketAddressInfoArgs:
+    host: ta.Optional[str]
+    port: ta.Union[str, int, None]
+    family: socket.AddressFamily = socket.AddressFamily.AF_UNSPEC
+    type: int = 0
+    proto: int = 0
+    flags: socket.AddressInfo = socket.AddressInfo(0)
+
+
+@dc.dataclass(frozen=True)
+class SocketAddressInfo:
+    family: socket.AddressFamily
+    type: int
+    proto: int
+    canonname: ta.Optional[str]
+    sockaddr: SocketAddress
+
+
+def get_best_socket_family(
+        host: ta.Optional[str],
+        port: ta.Union[str, int, None],
+        family: ta.Union[int, socket.AddressFamily] = socket.AddressFamily.AF_UNSPEC,
+) -> ta.Tuple[socket.AddressFamily, SocketAddress]:
+    """https://github.com/python/cpython/commit/f289084c83190cc72db4a70c58f007ec62e75247"""
+
+    infos = socket.getaddrinfo(
+        host,
+        port,
+        family,
+        type=socket.SOCK_STREAM,
+        flags=socket.AI_PASSIVE,
+    )
+    ai = SocketAddressInfo(*next(iter(infos)))
+    return ai.family, ai.sockaddr
+
+
+##
+
+
+class SocketHandler(abc.ABC):
+    def __init__(
+            self,
+            client_address: SocketAddress,
+            rfile: ta.BinaryIO,
+            wfile: ta.BinaryIO,
+    ) -> None:
+        super().__init__()
+
+        self._client_address = client_address
+        self._rfile = rfile
+        self._wfile = wfile
+
+    @abc.abstractmethod
+    def handle(self) -> None:
+        raise NotImplementedError

omlish/lite/socketserver.py

@@ -0,0 +1,66 @@
+# ruff: noqa: UP006 UP007
+import socket
+import socketserver
+import typing as ta
+
+from omlish.lite.check import check_not_none
+
+from .socket import SocketAddress
+from .socket import SocketHandlerFactory
+
+
+##
+
+
+class SocketServerBaseRequestHandler_:  # noqa
+    request: socket.socket
+    client_address: SocketAddress
+    server: socketserver.TCPServer
+
+
+class SocketServerStreamRequestHandler_(SocketServerBaseRequestHandler_):  # noqa
+    rbufsize: int
+    wbufsize: int
+
+    timeout: ta.Optional[float]
+
+    disable_nagle_algorithm: bool
+
+    connection: socket.socket
+    rfile: ta.BinaryIO
+    wfile: ta.BinaryIO
+
+
+##
+
+
+class SocketHandlerSocketServerStreamRequestHandler(  # type: ignore[misc]
+    socketserver.StreamRequestHandler,
+    SocketServerStreamRequestHandler_,
+):
+    socket_handler_factory: ta.Optional[SocketHandlerFactory] = None
+
+    def __init__(
+            self,
+            request: socket.socket,
+            client_address: SocketAddress,
+            server: socketserver.TCPServer,
+            *,
+            socket_handler_factory: ta.Optional[SocketHandlerFactory] = None,
+    ) -> None:
+        if socket_handler_factory is not None:
+            self.socket_handler_factory = socket_handler_factory
+
+        super().__init__(
+            request,
+            client_address,
+            server,
+        )
+
+    def handle(self) -> None:
+        target = check_not_none(self.socket_handler_factory)(
+            self.client_address,
+            self.rfile,  # type: ignore[arg-type]
+            self.wfile,  # type: ignore[arg-type]
+        )
+        target.handle()