feat: add support for EXTERNAL auth without uid

[mvn23]

As a follow-up to #188 , this PR allows EXTERNAL auth without a uid as per RFC4422. If we pass -1 as the uid, it will trigger the following auth sequence (source: https://gitlab.freedesktop.org/dbus/dbus/-/issues/195):

C: AUTH EXTERNAL            # I want to use EXTERNAL, but I do not
                            # have an initial response
S: DATA                     # Who do you claim you are?
C: DATA                     # Whoever the kernel says I am
S: OK                       # Can't argue with that!

The upside of this is that when running in a user namespace (e.g. rootless docker), we don't need to know our uid mapping outside of the namespace to connect to the system dbus on the host.

[codecov]

Codecov Report

Base: 82.28% // Head: 82.38% // Increases project coverage by +0.09% 🎉

Coverage data is based on head (8a34008) compared to base (830183e).
Patch coverage: 83.33% of modified lines in pull request are covered.

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #193      +/-   ##
==========================================
+ Coverage   82.28%   82.38%   +0.09%     
==========================================
  Files          27       27              
  Lines        3162     3167       +5     
  Branches      652      654       +2     
==========================================
+ Hits         2602     2609       +7     
+ Misses        343      342       -1     
+ Partials      217      216       -1     

Impacted Files	Coverage Δ
src/dbus_fast/__version__.py	0.00% <0.00%> (ø)
src/dbus_fast/auth.py	87.93% <100.00%> (+4.91%)	⬆️

Help us with your feedback. Take ten seconds to tell us how you rate us. Have a feature suggestion? Share it here.

☔ View full report at Codecov.
📢 Do you have feedback about the report comment? Let us know in this issue.

[bdraco]

Can you add coverage for these lines?

[mvn23]

All done. Should we try to fall back to this method if the default method with the uid from os.getuid() fails?

[bdraco]

All done. Should we try to fall back to this method if the default method with the uid from os.getuid() fails?

I think its probably fine without it os.getuid() is unlikely to fail

[bdraco]

This looks good.

Would you update the doc string below as well?

    """An authenticator class for the external auth protocol for use with the
    :class:`MessageBus <dbus_fast.message_bus.BaseMessageBus>`.
    :sealso: https://dbus.freedesktop.org/doc/dbus-specification.html#auth-protocol
    """

[mvn23]

I think its probably fine without it os.getuid() is unlikely to fail

It is currently the only reason I'm patching the Home Assistant docker image locally...
Granted, it's quite a specific scenario (running containerized in a user namespace with mapped uids), but if we only fall back to it after the initial attempt with the uid has failed, what's the harm?

[bdraco]

I think its probably fine without it os.getuid() is unlikely to fail

It is currently the only reason I'm patching the Home Assistant docker image locally... Granted, it's quite a specific scenario (running containerized in a user namespace with mapped uids), but if we only fall back to it after the initial attempt with the uid has failed, what's the harm?

Isn't that silently hiding an issue, it might not be the behavior the caller wants?

[bdraco]

Tested 👍