Allow setting the user which we drop permissions to using suid.

[alexmv]

What do these changes do?

We may want a different user than nobody when dropping permissions -- for example, if we intend to connect to a PostgreSQL database which does ident authentication.

Are there changes in behavior for the user?

No behaviour change.

Checklist

• I think the code is well written
• Unit tests for the changes exist
• tox testenvs have been executed in the following environments:
  • Linux (Ubuntu 18.04, Ubuntu 20.04, Arch): {py36,py37,py38,py39}-{nocov,cov,diffcov}, qa, docs
  • Windows (7, 10): {py36,py37,py38,py39}-{nocov,cov,diffcov}
  • WSL 1.0 (Ubuntu 18.04): {py36,py37,py38,py39}-{nocov,cov,diffcov}, pypy3-{nocov,cov}, qa, docs
  • FreeBSD (12.2, 12.1, 11.4): {py36,pypy3}-{nocov,cov,diffcov}, qa
  • Cygwin: py36-{nocov,cov,diffcov}, qa, docs
• Documentation reflects the changes
• Add a news fragment into the NEWS.rst file

[codecov]

Codecov Report

Merging #370 (bedda90) into master (83168cd) will not change coverage.
The diff coverage is 100.00%.

❗ Current head bedda90 differs from pull request most recent head 9182c42. Consider uploading reports for the commit 9182c42 to get more accurate results

@@            Coverage Diff            @@
##            master      #370   +/-   ##
=========================================
  Coverage   100.00%   100.00%           
=========================================
  Files            7         7           
  Lines         1706      1707    +1     
  Branches       310       310           
=========================================
+ Hits          1706      1707    +1     

Impacted Files	Coverage Δ
aiosmtpd/main.py	100.00% <100.00%> (ø)

📣 We’re building smart automated test selection to slash your CI/CD build times. Learn more

[waynew]

Just a small change in regards to the documentation. And another optional change.

Otherwise, I'm +1 on this. I went ahead and checked it out locally and verified that the tests fail when the code is as-is on the main branch.

I do have a slight concern that there might be a better way to have the other connections made and then drop privileges to nobody... but I don't think that's necessarily a blocker here.

(FYI I used to run test clinics for the Salt project - I'm running my own, and I went through this one at https://twitch.tv/wayneswonderarium tonight if anyone wants to catch the rerun. Also noticed a failure on mac but I don't think that was related to this PR -- was probably just a flaky issue with mac)

[waynew]

suggestion (non-blocking) Since we're going ahead and exiting with status 1 we could go ahead and simply sys.exit(f"Cannot ...") -- and use the (documented) functionality of sys.exit. However one of our tests does check that the exception .value.code == 1, and it would have to equal the error message instead.

Not a big deal, but that would simplify things here just a bit.

[alexmv]

This feels very tangential to this PR. Personally, I feel like the explicit print and exit code is more readable -- I've never seen the sys.exit("...") formulation, though it's by no means new!

So I'm not going to include this change in the PR, because I'm not entirely sold on it.

[alexmv]

Thanks for the review! I've repushed with your documentation change.

I do have a slight concern that there might be a better way to have the other connections made and then drop privileges to nobody... but I don't think that's necessarily a blocker here.

I don't think there is -- neither root nor nobody can auth, so both making the connections before or after calling suid(nobody) isn't feasible.

[alexmv]

Gentle nudge on this -- anything I can do to help move this forward?