remove 1to2 and depreciations

[andrewgsavage]

Here I've removed:

• the 1to2 code,
• the compatability code that handled obselete function calls,
• the depreciations in these calls,
• the tests that tested obselete behaviour

[codecov]

Codecov Report

Attention: Patch coverage is 57.14286% with 3 lines in your changes are missing coverage. Please review.

Project coverage is 95.75%. Comparing base (e9b82f0) to head (d963b9b).

❗ Current head d963b9b differs from pull request most recent head 53bf95a. Consider uploading reports for the commit 53bf95a to get more accurate results

Files	Patch %	Lines
uncertainties/unumpy/core.py	0.00%	3 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##           master     #214      +/-   ##
==========================================
+ Coverage   95.53%   95.75%   +0.21%     
==========================================
  Files          17       11       -6     
  Lines        2085     1910     -175     
==========================================
- Hits         1992     1829     -163     
+ Misses         93       81      -12     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[wshanks]

This looks good to me. Could you remove the 1to2 section from doc/index.rst as well?

Also, perhaps a separate discussion, but should the codecov checks be optional? It is nice to keep track of the coverage, but when it is not 100% it is hard to say whether it is good or bad that the number goes down slightly as code is added and removed. I think the ideal case is to require 100% coverage and mark code that doesn't need to be covered as "no cover".

[jagerber48]

Mmm I'm not exactly sure how the codecov CI works. the github action that I set up submits the code to codecov.io, but then it seems like codecov.io sends data back as a github action that can pass or fail.

I'm not sure how we would "make codecov checks optional". Right now I don't think they're part of the "final results" github action, so codecov not passing doesn't block merging a PR. This seems fine to me.

I think the ideal case is to require 100% coverage and mark code that doesn't need to be covered as "no cover".

I haven't looked at the code coverage on uncertainties yet. This PR will help clean things up and then we'll be able to get a clear picture of if uncertainties is hitting 100% coverage or not. But yes, once we have that picture we might mark some lines as no cover, but I prefer to be pretty conservative when it comes to marking lines as no cover. We may need to write more tests.

[andrewgsavage]

remove the 1to2 section from doc/index.rst

done

I think codecov is set up fine, we can use our own judgement for cases like this and still merge

[wshanks]

Looks good to me now!

I think you are right about codecov already being optional in the way that I meant. In any case, it is passing now 🙂

[newville]

Looks great.
Does anyone object to merging this?

[wshanks]

I think we had enough consensus to merge here!