urllib.request.pathname2url(): generate RFC 1738 URLs where possible

[barneygale]

Feature or enhancement

Proposal:

urllib.request.pathname2url currently generates RFC 1738-compliant file: URIs in the following cases:

• DOS drive paths (including drive-relative paths) on Windows.
• UNC paths on Windows.
• POSIX paths beginning // (since pathname2url() does not work if path starts with // #127217)

This function cannot generate RFC 1738-compliant URLs for:

• True relative paths, because of differing requirements for the first character of the path.

That leaves one case where the function could generate RFC 1738-compatible URLs, but doesn't:

• Paths beginning with exactly one slash (both POSIX and Windows)

>>> from urllib.request import pathname2url
>>> pathname2url('/etc/hosts')
'/etc/hosts'  # expected: '///etc/hosts'

For consistency with pathname2url()'s handling of other paths, and consistency with pathlib.Path.as_uri(), I propose we prepend two slashes to any path beginning with precisely one slash to produce a URL authority section with a zero-length authority.

Has this already been discussed elsewhere?

This is a minor feature, which does not need previous discussion elsewhere

Links to previous discussion of this feature:

#125866 (comment)

Linked PRs

• GH-127236: pathname2url(): generate RFC 1738 URL for absolute POSIX path #127194