Implement runner.addresses property

[asvetlov]

Fix #3036

[codecov-io]

Codecov Report

Merging #3037 into master will decrease coverage by 0.03%.
The diff coverage is 100%.

@@            Coverage Diff             @@
##           master    #3037      +/-   ##
==========================================
- Coverage   98.13%   98.09%   -0.04%     
==========================================
  Files          42       42              
  Lines        7721     7723       +2     
  Branches     1345     1346       +1     
==========================================
- Hits         7577     7576       -1     
- Misses         49       51       +2     
- Partials       95       96       +1

Impacted Files	Coverage Δ
aiohttp/web_runner.py	100% <100%> (ø)	⬆️
aiohttp/tcp_helpers.py	89.28% <0%> (-7.15%)	⬇️
aiohttp/client_reqrep.py	97.41% <0%> (-0.18%)	⬇️

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update dfe7582...1891b82. Read the comment docs.

[hynek]

How about returning the addresses as ipaddress? It might be worthwhile to go the extra mile and define IPEndpoint(ip, port) and UNIXEndpoint(path) to make it more elegant to detect which is which?

[asvetlov]

IPEndpoint(ipaddress, int) or UNIXEndpoint(str)?

[hynek]

👍

I don’t know how Endpoint fits into aiothttp’s naming but yeah. That looks good. Gonna be really easy to implement with attrs. ;)

[asvetlov]

Hmm. IPv6 has additional flow_info and scope_id.
I have not enough information how people use it but suspect keeping everything is essential.
My motivation is: let's support the same naming structures as well-known socket module do.
Later we can invent a better API by adding .endpoints property :)

[webknjaz]

@asvetlov what is the motivation behind this?

[asvetlov]

To keep a stable order of returned addresses.
Registering a new site in a runner is not a bottleneck, we can live with O(n) for it.

[webknjaz]

Ah, okay. I just didn't know why you'd do this without any comments or commit description :)

[lock]

This thread has been automatically locked since there has not been any recent activity after it was closed. Please open a [new issue] for related bugs.
If you feel like there's important points made in this discussion, please include those exceprts into that [new issue].
[new issue]: https://github.com/aio-libs/aiohttp/issues/new