Remove codecov, solely use coverage.

[tim-schilling]

CodeCov has become a bit flaky in their evaluation of code coverage.
Using GitHub actions, we can utilize coverage to combine multiple reports.

Thanks to Hynek Schlawack for the approach.
https://hynek.me/articles/ditch-codecov-python/

[tim-schilling]

@matthiask what do you think about dropping codecov? It's probably possible to get the github actions to create/update a comment on the PR with the report generated.

I'm a bit unsure of all this, but I do know we've had a few PRs fail CI due to codecov saying it didn't cover everything.

Additionally, we'd probably need to lower the failure percentage and/or cover existing code paths.

[matthiask]

I read the post by Hynek too. It makes a lot of sense to me. Thank you!

[auvipy]

how is the report looks like?

[tim-schilling]

@auvipy in this implementation you'd have to click on the "Check coverage" action's details link: https://github.com/jazzband/django-debug-toolbar/runs/4514568613?check_suite_focus=true

It contains a table that looks like the following:

Name                                                 Stmts   Miss Branch BrPart  Cover   Missing
------------------------------------------------------------------------------------------------
debug_toolbar/__init__.py                                7      0      2      0   100%
debug_toolbar/apps.py                                   58      3     25      1    95%   78, 116-117
debug_toolbar/decorators.py                             21      0      4      0   100%
debug_toolbar/forms.py                                  25      0      8      0   100%
debug_toolbar/management/__init__.py                     0      0      0      0   100%
debug_toolbar/management/commands/__init__.py            0      0      0      0   100%
debug_toolbar/management/commands/debugsqlshell.py      20      0      4      0   100%
debug_toolbar/middleware.py                             60      1     26      1    98%   32
debug_toolbar/panels/__init__.py                        65      2      6      1    96%   82, 92, 103->exit
debug_toolbar/panels/cache.py                          164     22     44      7    84%   46, 72, 75-77, 80, 83, 87, 113, 117, 121, 125, 129, 133, 137, 141, 163, 221-225, 278->277, 282, 298->297, 300->exit
debug_toolbar/panels/headers.py                         22      0      8      0   100%
debug_toolbar/panels/history/__init__.py                 2      0      0      0   100%
debug_toolbar/panels/history/forms.py                    3      0      2      0   100%
debug_toolbar/panels/history/panel.py                   52      2      8      0    97%   63-65
debug_toolbar/panels/history/views.py                   33      2     12      2    91%   37, 71
debug_toolbar/panels/logging.py                         49      2      8      0    96%   11-12
debug_toolbar/panels/profiling.py                       96      6     32      3    93%   29, 32-33, 43, 102, 110
debug_toolbar/panels/redirects.py                       18      0      6      0   100%
debug_toolbar/panels/request.py                         32      0     10      0   100%
debug_toolbar/panels/settings.py                        16      0      6      0   100%
debug_toolbar/panels/signals.py                         44      3     19      3    90%   56, 79-80, 89->91
debug_toolbar/panels/sql/__init__.py                     2      0      0      0   100%
debug_toolbar/panels/sql/forms.py                       36      4      6      2    86%   31, 40-41, 47
debug_toolbar/panels/sql/panel.py                      164     46     72      7    68%   20-32, 36-48, 68-92, 182-187, 201-205, 207, 211, 215, 223->225, 228-229, 239, 281-282
debug_toolbar/panels/sql/tracking.py                   125      9     48      3    93%   42->exit, 71->exit, 159-160, 193-197, 210, 216
debug_toolbar/panels/sql/utils.py                       51      2     22      2    95%   75, 82
debug_toolbar/panels/sql/views.py                       67      3     22      3    93%   33, 69, 122
debug_toolbar/panels/staticfiles.py                    112      3     36      2    97%   16-17, 42, 177->176
debug_toolbar/panels/templates/__init__.py               2      0      0      0   100%
debug_toolbar/panels/templates/panel.py                124      6     46      7    92%   25->34, 37, 108, 112, 127, 181-182, 185->188
debug_toolbar/panels/templates/views.py                 43     11     10      3    74%   19, 22-23, 30->29, 44-47, 51-56
debug_toolbar/panels/timer.py                           59      8     12      4    80%   11-12, 29-32, 57-59, 63->65, 69->71, 71->93
debug_toolbar/panels/versions.py                        41      1     20      1    97%   58
debug_toolbar/settings.py                               17      0      0      0   100%
debug_toolbar/templatetags/__init__.py                   0      0      0      0   100%
debug_toolbar/toolbar.py                                87      5     28      2    92%   71-79, 98, 127->139
debug_toolbar/urls.py                                    3      0      0      0   100%
debug_toolbar/utils.py                                 153     23     71     13    80%   18-19, 29-30, 78-81, 108, 112->115, 129->128, 158->162, 179-180, 184, 191-192, 200, 204-212, 217, 227, 250, 261->263, 268->270
debug_toolbar/views.py                                  16      0      2      0   100%
------------------------------------------------------------------------------------------------
TOTAL                                                 1889    164    625     67    89%
Coverage failure: total of 89 is less than fail-under=100

[pauloxnet]

I think it woul be better start using a pyproject.toml to store configurations.

[auvipy]

I think it woul be better start using a pyproject.toml to store configurations.

I quite agree with you, but may be in another PR

[tim-schilling]

Any qualms with me setting the failure percentage to 89%?

[matthiask]

Any qualms with me setting the failure percentage to 89%?

I think that's fine. Maybe you could insert a // TODO Raise this, when the testsuite allows it or something so that the percentage appears in editors' to do lists.

[pauloxnet]

Any qualms with me setting the failure percentage to 89% ?

I agree with that. We can also define a coverage goal fro the next release (e.g. 90%)

[tim-schilling]

@pauloxnet can you confirm that all of your concerns with the PR were addressed? I'll squash and merge our commits after I get that green checkmark.

[pauloxnet]

Everything ok for me.

[pauloxnet]

I noticed only now that in this PR we forgot to update the coverage badge in the README.