Initialize headers from config in reset_base_headers()

[glensc]

WIP: #706

[glensc]

OK, here's some idea that could work. Don't mind the PR title and body, will update them once the PR comes useful.

So, WDYT, is this acceptable?

[glensc]

Maybe it's about time to drop support for the 'header.platorm' typo?

[glensc]

it was fixed in 2018, matching 3.2.0 version:

• fix typo in header.platorm config param #291

[glensc]

How about this approach? If needed, I could extract to smaller changes.

[Hellowlol]

I still don’t see the point. IMO you need to add a example on why this is better. Fix your flake errors, run the test suit locally, including the tests for the client and squash the commits.

I don’t like the name reset_default_headers as isn’t anymore, is the headers is in config file (it was a bad name in the first place) Why is the headers a local variable the class? Why not use it directly?

[Hellowlol]

Don’t in-line std libs.

[Hellowlol]

Fix the slicing.

[glensc]

Explain fix how? fix what? what is wrong with the current code?

[glensc]

why is the current slicing wrong. how is the correct way?

[glensc]

Fix your flake errors, run the test suit locally, including the tests for the client and squash the commits.

I don't see point fixing tests until it's certain this is worth finishing! That's why I asked about the changes, not the style! Also, do note the PR is in a Draft state.

[glensc]

I don’t like the name reset_default_headers as isn’t anymore, is the headers is in config file (it was a bad name in the first place)

I used an existing name, please suggest a better name not only complain it's bad.

• get_base_headers?
• get_plex_headers?

[glensc]

Why is the headers a local variable the class? Why not use it directly?

I think I explained it in an issue this PR is referring to (#706). to make the headers instance-based, not global. so the self._headers is now copied into the instance. is that headers you are referring to?

since Config object may be modified runtime (after __init__) using copy initialized in __init__.py is the original problem.

ps: since Github has no threading in comment mode, submit a review as code comments, so can reply in a thread and resolve disucssions.