Bring back link checks

[The-Compiler]

We have quite some history when it comes to link checks:

• do linkchecks for changelog lint #1722
• Fix/Remove broken links #5613
• Added link checking to tox and release.py #5614
• [release] should not check (all) links #6612
• Skip link checks when doing releases through the bot #6916

So we're currently not running link checking at all. And it shows:

.../adopt.rst:56: WARNING: broken link: fixture.html ()
.../adopt.rst:55: WARNING: broken link: assert.html ()
.../talks.rst:88: WARNING: broken link: http://bruynooghe.blogspot.com/2010/06/pytest-test-generators-and-cached-setup.html (404 Client Error: Not Found for url: http://bruynooghe.blogspot.com/2010/06/pytest-test-generators-and-cached-setup.html)
.../talks.rst:103: WARNING: broken link: http://bruynooghe.blogspot.com/2009/12/skipping-slow-test-by-default-in-pytest.html (404 Client Error: Not Found for url: http://bruynooghe.blogspot.com/2009/12/skipping-slow-test-by-default-in-pytest.html)
.../announce/release-2.2.0.rst:12: WARNING: broken link: http://pytest.org/en/stable/example/how-to/parametrize.html (404 Client Error: Not Found for url: https://pytest.org/en/stable/example/how-to/parametrize.html)
.../adopt.rst:55: WARNING: broken link: https://bitbucket.org/hpk42/pycmd/overview (404 Client Error: Not Found for url: https://bitbucket.org/hpk42/pycmd/overview)
.../announce/release-2.6.0.rst:21: WARNING: broken link: https://bitbucket.org/pytest-dev/pytest/issues?status=new&status=open&sort=-votes (404 Client Error: Not Found for url: https://bitbucket.org/pytest-dev/pytest/issues?status=new&status=open&sort=-votes)
.../history.rst:120: WARNING: broken link: http://web.archive.org/web/20091012022829/http://codespeak.net/py/dist/test/plugin/how-to/monkeypatch.html (404 Client Error: NOT FOUND for url: http://web.archive.org/web/20091012022829/http://codespeak.net/py/dist/test/plugin/how-to/monkeypatch.html)
.../explanation/goodpractices.rst:113: WARNING: broken link: https://blog.ionelmc.ro/2014/05/25/python-packaging/#the-structure> (Anchor 'the-structure>' not found)
.../announce/release-2.3.0.rst:97: WARNING: broken link: https://docs.pytest.org/en/6.0.1/faq.html (404 Client Error: Not Found for url: https://docs.pytest.org/en/6.0.1/faq.html)
.../announce/release-8.0.0rc2.rst:22: WARNING: broken link: https://docs.pytest.org/en/release-8.0.0rc2/changelog.html (404 Client Error: Not Found for url: https://docs.pytest.org/en/release-8.0.0rc2/changelog.html)
.../reference/reference.rst:153: WARNING: broken link: https://github.com/asottile/re-assert#re_assertmatchespattern-str-args-kwargs (Anchor 're_assertmatchespattern-str-args-kwargs' not found)
.../contact.rst:24: WARNING: broken link: https://matrix.to/#/%23pytest:libera.chat (Anchor '/%23pytest:libera.chat' not found)
.../changelog.rst:5757: WARNING: broken link: https://peps.python.org/pep-0506/#recommended-filter-settings-for-test-runners (Anchor 'recommended-filter-settings-for-test-runners' not found)
.../history.rst:99: WARNING: broken link: https://pypi.org/project/py/0.8.0-alpha2/#history/ (Anchor 'history/' not found)
.../contributing.rst:150: WARNING: broken link: https://tox.readthedocs.io/en/latest/config.html#configuration-discovery (Anchor 'configuration-discovery' not found)
.../contact.rst:24: WARNING: broken link: https://web.libera.chat/#pytest (Anchor 'pytest' not found)
.../adopt.rst:57: WARNING: broken link: mark.html ()
.../talks.rst:83: WARNING: broken link: monkeypatch.html ()
.../talks.rst:87: WARNING: broken link: parametrize.html#test-generators ()
.../talks.rst:105: WARNING: broken link: plugins.html ()
.../adopt.rst:47: WARNING: broken link: unittest.html ()
.../changelog.rst:1157: WARNING: broken link: with-setup-nose ()
.../adopt.rst:59: WARNING: broken link: xdist.html ()
.../adopt.rst:56: WARNING: broken link: xunit_setup.html ()
.../explanation/goodpractices.rst:290: WARNING: broken link: https://github.com/pytest-dev/pytest-runner#deprecation-notice (Anchor 'deprecation-notice' not found)
.../history.rst:108: WARNING: broken link: https://github.com/pytest-dev/pytest/blob/main/doc/en/changelog.rst#091 (Anchor '091' not found)
.../changelog.rst:104: WARNING: broken link: https://github.com/python/cpython/blob/51aefc5bf907ddffaaf083ded0de773adcdf08c8/Lib/unittest/case.py#L419-L426 (Anchor 'L419-L426' not found)
.../changelog.rst:6716: WARNING: broken link: https://github.com/thisch (404 Client Error: Not Found for url: https://github.com/thisch)

We should bring it back somehow. According to @webknjaz it got better at doing retries and such, and apparently there's also a distinct status for timeouts:

• linkcheck builder: add a distinct reporting status code for timeouts sphinx-doc/sphinx#11868

From @nicoddemus in #6916:

The idea is good, but I'm not sure if the scheduled based job won't just be ignored in the end.

But what we currently have is worse IMHO...

[nicoddemus]

The idea is good, but I'm not sure if the scheduled based job won't just be ignored in the end.

I reconsidered this, because now we have some scheduled jobs (like plugin update, pre-commit, etc) and history shows that they are not ignored, often sometimes takes time to fix them if they fail.

So a scheduled job might do the trick here in the end.

[ladislav987]

Hi @The-Compiler, I would like to look at it as my first issue.

[webknjaz]

@ladislav987 please feel free to ask questions as you work through the issue or submit a draft PR right away, updating it as you work through stuff.

[nikunjmathur08]

Can I work on this? @The-Compiler

[The-Compiler]

@nikunjmathur08 See the comments right above yours, @ladislav987 has already offered to work on this a week ago.

[ladislav987]

I read all about "link check" history. If I get it right you want use sphinx. Or you want try something else?

[webknjaz]

This is about Sphinx.

[nikunjmathur08]

@The-Compiler just asked because there was no interaction for the past week. No issues if @ladislav987 is already working on it.

[ayush-a-srivastava]

There is already a fix in the tox.ini file, which checks the links. #12474

Is it still not working?