Tidying __init__

[pepoluan]

trafficstars

~~Depends on #256 , #259 , and #260 ; please merge them first.~~ All deps have been merged. We can go ahead with this one.

What do these changes do?

The __init__ method is very unwieldy. This PR makes it simpler by extracting some logic into properties.

This led to several improvements:

• event_handler can be replaced "on-the-fly"
• tls_context can be replaced "on-the-fly" and will automagically change require_tls accordingly
• User can selectively overload logic in the properties; no longer have to copy > 100 lines of init code just to override one behavior 😉

Are there changes in behavior for the user?

None.

Users have to be aware though that some properties are more 'noisy' in logging.

Related issue number

Inspired partially by #253

Checklist

• [x] I think the code is well written
• [x] Unit tests for the changes exist
• [x] tox testenvs have been executed in the following environments:
  • [x] Windows 10 (via PyCharm tox runner)
  • [x] Windows 10 (via PSCore 7.1.2)
  • [x] Windows 10 (via Cygwin)
  • [x] Ubuntu 18.04 on WSL 1.0
  • [x] FreeBSD 12.2 on VBox
  • [x] OpenSUSE Leap 15.2 on VBox
• [x] Documentation reflects the changes
• [x] Add a news fragment into the NEWS.rst file

[codecov[bot]]

Codecov Report

Merging #261 (4bd08bb) into master (4bd08bb) will not change coverage. The diff coverage is n/a.

:exclamation: Current head 4bd08bb differs from pull request most recent head 5f3557c. Consider uploading reports for the commit 5f3557c to get more accurate results

@@            Coverage Diff            @@
##            master      #261   +/-   ##
=========================================
  Coverage   100.00%   100.00%           
=========================================
  Files            7         7           
  Lines         1706      1706           
  Branches       310       310           
=========================================
  Hits          1706      1706           

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

[pepoluan]

Now that I think of it ... this should be merged ONLY after #260 has been merged.

That way, I can test for the behavior of the library when tls_context is being replaced totally. (Current tests are replacing to/from None, and replacing with a new SSLContext object but using the same cert-key pair.)

[waynew]

Nice. I'm digging at least the idea :)

[pepoluan]

As soon as test passes on my testing systems, I'm undrafting this and let this marinate.

I might not be able to do code changes for several days ahead, due to me transitioning from "being employed" to "being a free agent" and needing to acquire a new laptop. I reckon up to a week of radio silence due to the Easter holidays. Hopefully not as long as that, depending on how soon the store can fulfill my order.

[pepoluan]

Nice. I'm digging at least the idea :)

A side effect is that SMTP instatiation is now faster 😉

[waynew]

@pepoluan I'm going to plan on taking a closer look at this one this week, too. Looks like there are a couple of conflicts, if you're able to tackle those.

[pepoluan]

@pepoluan I'm going to plan on taking a closer look at this one this week, too. Looks like there are a couple of conflicts, if you're able to tackle those.

Fixed the conflicts, rebase, and let's see...

[pepoluan]

Oh gosh it blew up! LOL

[pepoluan]

Heh seriously at this point I'll just close this PR, but grab the ideas and make a new PR based on the latest master.

There are just too many incremental changes since this PR was created, the proposed changes in this PR makes things go awry.