jw-pkg/src/python/jw/pkg/lib/Uri.py
Jan Lindemann 104c6d9040
lib.Uri: Fix path join for empty authority
__new_with_path() joins base and path with exactly one '/', assuming a
trailing '/' in base is a path separator. That assumption breaks for
URIs with empty authority, where scheme_plus_authority ends in '://'
(e.g. 'file:///tmp/x'): new_replace_path('/etc/hosts') drops the
leading slash of the path and produces 'file://etc/hosts', which
parses with hostname 'etc' and path '/hosts' instead of path
'/etc/hosts'.

Treat a base ending in '://' separately and keep a leading slash in
the path, adding one if it is missing.

Assisted-by: unsloth/Qwen3.8-27B-GGUF:Q4_K_M with pi.dev v0.84.2
Signed-off-by: Jan Lindemann <jan@janware.com>
2026-08-22 10:07:58 +02:00

181 lines
5 KiB
Python

from __future__ import annotations
from functools import cached_property
from typing import TYPE_CHECKING, override
if TYPE_CHECKING:
import urllib.parse
from typing import Self
# Make sure URIs are interpreted indentically everywhere
class Uri:
def __assemble(
self, scheme: bool, credentials: bool, secure: bool, path: bool
) -> str:
ret = ''
if self.__p.path.startswith('~'):
return self.path
if scheme:
ret += f'{self.protocol}://'
if credentials and self.username:
ret += self.username
if self.password:
ret += ':' + '<hidden>' if secure else self.password
ret += '@'
if self.hostname:
ret += self.hostname
if self.port_str:
ret += ':' + self.port_str
if path:
ret += self.path
return ret
def __init__(self, string: str) -> None:
self.__string = string
self.__username: str | None = None
self.__password: str | None = None
@override
def __repr__(self) -> str:
return self.full
@override
def __str__(self) -> str:
return self.safe_full_with_username
@cached_property
def __p(self) -> urllib.parse.ParseResult:
from urllib.parse import urlparse
return urlparse(self.__string)
@classmethod
def pimp(cls, url: str | Self) -> Uri:
if isinstance(url, Uri):
return url
return Uri(url)
@property
def to_string(self) -> str:
return self.__string
@cached_property
def scheme(self) -> str:
ret = self.__p.scheme
if not ret and not self.__p.path.startswith('~'):
return 'file'
return ret
@cached_property
def protocol(self) -> str:
return self.scheme.replace('://', '')
@property
def username(self) -> str | None:
if self.__username is None:
return self.__p.username
return self.__username
def set_username(self, username: str) -> None:
self.__username = username
@property
def password(self) -> str | None:
if self.__password is None:
return self.__p.password
return self.__password
def set_password(self, password: str) -> None:
self.__password = password
@cached_property
def hostname(self) -> str | None:
return self.__p.hostname
@cached_property
def port(self) -> int | None:
return self.__p.port
@cached_property
def port_str(self) -> str | None:
if self.port is None:
return None
return str(self.port)
@cached_property
def path(self) -> str:
return self.__p.path
@cached_property
def basename(self) -> str:
return self.__p.path.rsplit('/')[-1]
@cached_property
def authority(self) -> str:
return self.__assemble(
scheme = False, credentials = True, secure = False, path = False
)
@cached_property
def origin(self) -> str:
return self.__assemble(
scheme = False, credentials = False, secure = True, path = False
)
@cached_property
def scheme_plus_authority(self) -> str:
return self.scheme + '://' + self.authority
@cached_property
def id(self) -> str:
return self.__assemble(
scheme = True, credentials = True, secure = True, path = False
)
@cached_property
def full(self) -> str:
return self.__assemble(
scheme = True, credentials = True, secure = False, path = True
)
@cached_property
def safe_full_with_username(self) -> str:
return self.__assemble(
scheme = True, credentials = True, secure = True, path = True
)
def __new_with_path(self, base: str, path: str) -> Self:
# -- Build a fresh instance rather than copying self: a copy would
# inherit computed cached_property values (e.g. __p, path, full),
# which would go stale as soon as __string is replaced below.
ret = object.__new__(type(self))
ret.__string = base
ret.__username = None
ret.__password = None
if not path:
return ret
if ret.__string.endswith('://'):
# -- Empty authority: the trailing '/' is part of the '://'
# separator, so a leading slash in path must be kept to stay
# an absolute path
ret.__string += path if path.startswith('/') else '/' + path
return ret
if ret.__string[-1] == '/':
if path[0] == '/':
ret.__string += path[1:]
else:
ret.__string += path
else:
if path[0] == '/':
ret.__string += path
else:
ret.__string += '/' + path
return ret
def new_add_path(self, path: str) -> Self:
return self.__new_with_path(self.__string, path)
def new_replace_path(self, path: str) -> Self:
return self.__new_with_path(self.scheme_plus_authority, path)