Adrien and I added this is in https://github.com/certbot/certbot/pull/6590 in response to https://github.com/certbot/certbot/issues/6582 which I wrote. I now personally think these tests are way more trouble than they're worth.
In almost all cases, the versions pinned in `tools/requirements.txt` are used. The two exceptions to this that come to mind are users using OS packages and pip. In the former, the version of our dependencies is picked by the OS and do not change much on most systems. As for pip, [we only "support it on a best effort basis"](https://eff-certbot.readthedocs.io/en/stable/install.html#alternative-2-pip).
Even for pip users, I'm not convinced this buys us much other than frequent test failures. We have our tests configured to error on all Python warnings and [we regularly update `tools/requirements.txt`](https://github.com/certbot/certbot/commits/master/tools/requirements.txt). Due to that, assuming our dependencies follow normal conventions, we should have a chance to fix things in response to planned API changes long before they make their way to our users. I do not think it is necessary for our tests to break immediately after an API is deprecated.
I think almost all other failures due to these tests are caused by upstream bugs. In my experience, almost all of them will sort themselves out pretty quickly. I think that responding to those that are not or planned API changes we somehow missed can be addressed when `tools/requirements.txt` is updated or when someone opens an issue. I personally don't think blocking releases or causing our nightly tests to fail is at all worth it here. I think removing this frequent cause of test failures makes things just a little bit easier for Certbot devs without costing us much of anything.
* generate multiarch images for non-architecture tags
* lock docker build to legacy docker buider, and bugfix
* rename deploy.sh to deploy_by_arch.sh
* Update documentation related to multiarch Docker
* Consistent IFS value with respect to other scripts
Co-authored-by: humanoid2050 <humanoid2050@monolith>
Co-authored-by: Brad Warren <bmw@users.noreply.github.com>
* certbot-ci: boulder will now only supports port 80 for http-01
* forgot to actually use the http_01_port argument
* print the port the proxy listens on
* try allow binding to privileged ports
* fix cover tox envs
* make test work on all Pythons
* Remove unused import
Co-authored-by: alexzorin <alex@zorin.id.au>
Co-authored-by: alexzorin <alex@zorin.id.au>
This is in response to the thread starting at https://github.com/certbot/certbot/pull/9330#issuecomment-1320416069.
In addition to this, I plan to add the following text to the step of the release instructions that tells you to wait until Azure Pipelines for the release has finished running:
> Some jobs such as building our snaps can take a long time to complete, however, if the process seems hung, you can cancel the build and then rerun the failed jobs. To do this, click on the build for the release in the link above, make sure you're logged into Azure Pipelines, and then use the cancel/rerun buttons in the top right of the web page.
This PR:
* Deletes the 2.0 pre-release pipeline
* Causes 1.x releases to be released to Docker Hub without updating the latest tag, PyPI, and the candidate and stable channels of the snap store
* Causes 2.x releases to be released to Docker Hub, PyPI, the beta channel of the snap store, and our Windows installer
We could potentially look into how to continue to do 1.x Windows installer releases through GitHub releases and tech ops tooling, but I personally don't think it's worth it right now.
This PR DOES NOT do anything about progressive snap releases. I think we can revisit this when/if we decide (how) to do them.
* Remove external mock dependency
This also removes the "external-mock" test environment
* remove superfluous ignores
* remove mock warning ignore from pytest.ini
* drop deps on mock in oldest, drop dep on types-mock
Co-authored-by: Alex Zorin <alex@zorin.id.au>
* try the easy thing of just doing what the error message says
* temporarily add deploy stage to extended tests to see if it uploads properly
* follow instructions on https://forum.snapcraft.io/t/snapcraft-authentication-options/30473
* just run the packaging jobs for speed
* fix formatting
* import changes from test- branch and revert temporary changes
* Update instructions in deploy-stage.yml
It was previously 5.5 hours, which was just to have an exception thrown
before Azure's 6 hour timeout. Generally we aren't seeing this step take
more than 45 minutes, so 90 minutes seems like more than enough.
* Remove Windows 2016 environment, generate 64 bit installer
* Add note to changelog
* Use win_amd64 as installer suffix
* Bump PYTHON_BITNESS to 64
* Require 64 bit Windows for the installer_build job
* Update certbot install path
* update windows test name
* Base installer suffix on PYTHON_BITNESS again
* Update changelog to request users uninstall old version
The reason I want to do this is many of the targets of `test_sdists.sh` use Python 3.6 which [has reached its EOL](https://www.python.org/dev/peps/pep-0494/#lifespan). We could instead just stop running the test on these systems or install a newer version of Python 3 outside of OS packaging, but instead I decided to look into why we have these tests to begin with.
I introduced these tests many years ago in https://github.com/certbot/certbot/pull/4089 as a fix for https://github.com/certbot/certbot/issues/4044. Essentially the problem was the way packagers ran tests and the way we ran tests were slightly different. This difference could cause test failures when distros tried to run tests on our packages.
Since I did this, [we've switched to telling packagers to run tests using `pytest` like we do](https://github.com/certbot/certbot/blob/5e76669c5026732d1440e39e93b58220ef7bd354/certbot/docs/packaging.rst#notes-for-package-maintainers) and we've greatly reduced our reliance on OS packaging through things like `snap`.
Because of this, I think we should stop running this test, reducing our reliance on the heavy "test farm tests", and simplifying our CI pipeline. I think future problems here is quite unlikely and even if we have them, it should only affect tests on our non-primary distribution mechanisms which I think is a very minor concern.
When reviewing this PR, it's probably worth noting that I just replaced `targets.yaml` with the contents of `apache2_targets.yaml` since the Apache 2 tests are the only runs we're running with this change.
Fixes https://github.com/certbot/certbot/issues/9058.
The changes to the CI config are equivalent to the ones made in https://github.com/certbot/certbot/pull/8460.
Other than ignoring some warnings raised by botocore, the main additional work that had to be done here was switching away from using `distutils.version.LooseVersion` since the entire `distutils` module was deprecated in Python 3.10. To do that, I took a few different approaches:
* If the version strings being parsed are from Python packages such as Certbot or setuptools, I switched to using [pkg_resources.parse_version](https://setuptools.pypa.io/en/latest/pkg_resources.html#parsing-utilities) from `setuptools`. This functionality has been available since [setuptools 8.0 from 2014](https://setuptools.pypa.io/en/latest/history.html#id865).
* If the version strings being parsed are not from Python packages, I added code equivalent to `distutils.version.LooseVersion` in `certbot.util.parse_loose_version`.
* The code for `CERTBOT_PIP_NO_BINARY` can be completely removed since that variable isn't used or referenced anywhere in this repo.
* add python 3.10 support
* make some version changes
* don't use looseversion in setup.py
* switch to pkg_resources
* deprecate get_strict_version
* fix route53 tests
* remove unused CERTBOT_PIP_NO_BINARY code
* stop using distutils in letstest
* add unit tests
* more changelog entries
It seems that all required pre-compiled wheels to install Certbot on Python 3.9 on Windows are present.
This PR upgrades Windows tests on Python 3.9 and repackages the installer on this version of Python.
This PR is a new approach for fixing #8732 based on the discussions occurred in the first PR #8877.
This PR upgrades python-augeas to the latest version, and avoids tests failure of Windows because of this upgrade. To do so it leverages the [tox multi-platform feature](https://tox.readthedocs.io/en/latest/example/platform.html) and modifications to `tools/venv.py` in order to not install and not test `certbot-apache` on Windows.
* Unpin python-augeas and upgrade current pinnings
* Do not install certbot-apache in Windows dev environments
* Introduce tox specific win packages + remove certbot compatibility on windows
* Add libaugeas to sphinx build
* Redefine lint and mypy targets
* Keep the lint and mypy environments
I think we should use our `pip_install*` scripts wherever we can and I'm not quite sure yet if I'd call `repoze.sphinx.autointerface` unmaintained.
* use pip_install_editable
* update sphinx comment
Since Saturday the CI pipeline is failing due to several Sphinx errors. See https://dev.azure.com/certbot/certbot/_build/results?buildId=3928&view=logs&j=d74e04fe-9740-597d-e9fa-1d0400037dfd&t=dde413a4-f24c-59a0-9684-e33d79f9aa02
First, the build of certbot-dns-google is failing because of a particular configuration. It seems that this configuration has been written here to activate the support of the RST instruction `.. code-block:: json` in documentation. However, it does not seem to be necessary for a similar situation in certbot-dns-route53 documentation. So let's try to remove it and fix the Sphinx builds.
Second, Sphinx builds were not pinning dependencies, so Sphinx 4.x (that has been released yesterday) started to be used in the pipeline. Sadly this new version is not compatible with the plugin `repoze.sphinx.autointerface`, used to extract documentation from `zope.interface`. So I fixed the pinning and also explicitly pin Sphinx to 3.5.x for now.
Technically speaking the second action is sufficient to fix the first error, but I keep the dedicated solution because it improves the documentation in my opinion.
This situation could be fixed by not requiring `repoze.sphinx.autointerface`, but this is possible only if we remove `zope.interface` from Certbot. Luckily I started the work few days ago ;).
* Remove explicit lexer call in certbot-dns-google doc builds.
* Write a valid JSON file in the documentation
* Apply constraints to sphinx build environments
* Pin Sphinx to 3.5.4
* Update dependencies
* Pin traitlets
* Move version.py to tests/letstest since it's used by test_sdists.sh
* Delete unused components of certbot-auto
* Remove test_leauto_upgrades.sh and references to it
* Remove test_letsencrypt_auto_certonly_standalone.sh and references to it
* Remove outstanding references to certbot-auto
* Remove references to letsencrypt-auto
* find certbot in the correct directory
* delete letsencrypt-auto-source line from .isort.cfg since that directory no longer contains any python code
* remove (-auto) from certbot(-auto)
* delete line from test
* Improve style for version.py
Fixes#8425
This PR upgrades mypy to the latest version available, 0.812.
Given the advanced type inference capabilities provided by this newer version, this PRs also fixes various type inconsistencies that are now detected. Here are the non obvious changes done to fix types:
* typing in mixins has been solved using `Protocol` classes, as recommended by mypy (https://mypy.readthedocs.io/en/latest/more_types.html#mixin-classes, https://mypy.readthedocs.io/en/stable/protocols.html)
* `cast` when we are playing with `Union` types
This PR also disables the strict optional checks that have been enable by default in recent versions of mypy. Once this PR is merged, I will create an issue to study how these checks can be enabled.
`typing.Protocol` is available only since Python 3.8. To keep compatibility with Python 3.6, I try to import the class `Protocol` from `typing`, and fallback to assign `object` to `Protocol` if that fails. This way the code is working with all versions of Python, but the mypy check can be run only with Python 3.8+ because it needs the protocol feature. As a consequence, tox runs mypy under Python 3.8.
Alternatives are:
* importing `typing_extensions`, that proposes backport of newest typing features to Python 3.6, but this implies to add a dependency to Certbot just to run mypy
* redesign the concerned classes to not use mixins, or use them differently, but this implies to modify the code itself even if there is nothing wrong with it and it is just a matter of instructing mypy to understand in which context the mixins can be used
* ignoring type for these classes with `# type: ignore` but we loose the benefit of mypy for them
* Upgrade mypy
* First step for acme
* Cast for the rescue
* Fixing types for certbot
* Fix typing for certbot-nginx
* Finalize type fixes, configure no optional strict check for mypy in tox
* Align requirements
* Isort
* Pylint
* Protocol for python 3.6
* Use Python 3.9 for mypy, make code compatible with Python 3.8<
* Pylint and mypy
* Pragma no cover
* Pythonic NotImplemented constant
* More type definitions
* Add comments
* Simplify typing logic
* Use vararg tuple
* Relax constraints on mypy
* Add more type
* Do not silence error if target is not defined
* Conditionally import Protocol for type checking only
* Clean up imports
* Add comments
* Align python version linting with mypy and coverage
* Just ignore types in an unused module
* Add comments
* Fix lint
* Pin pynsist
* Update dependencies
* Set windows installer a proper python project
* Optimize usage of the venvs
* Add windows-installer when venv is set up
* Fix call
* Remove env marker
Fixes#8700
Now that `snapcraft remote-build` truly uses new builds for each call, we can split the builds to have a dedicated Azure job for each target architecture. This PR does that.
* Split snap_build job on each architecture
* Also parallelize the publish_snap jobs over each architecture
Fixes#8389#8584.
This PR makes the necessary modifications to officially drop Python 2 support in the Certbot project.
I did not remove the specific Python 2 compatibility branches that has been added in various places in the codebase, to reduce the size of this PR and this will be done in a future one
* Update classifiers and python_requires in setup.py
* Remove warnings about Python 2 deprecation
* Remove Azure jobs on Python 2.7
* Remove references to python 2 in documentation
* Pin dnspython to 2.1.0
* Update changelog
* Remove warning ignore
Fixes https://github.com/certbot/certbot/issues/8580.
With this PR, it should now be possible to run the oldest tests natively on Linux, at least when using an older version of Python 3, which hasn't been possible in a long time. Unfortunately, this isn't possible on macOS which I opened https://github.com/certbot/certbot/issues/8589 to track.
You can see the full test suite running with these changes at https://dev.azure.com/certbot/certbot/_build/results?buildId=3283&view=results.
I took the version numbers for the packages I updated by searching for the oldest version of the dependency I think we should try and support based on the updated comments at the top of `oldest_constraints.txt`. While kind of annoying, I think it'd be a good idea for the reviewer to double check that I didn't make a mistake with the versions I used here.
To find these versions, I used https://packages.ubuntu.com, https://packages.debian.org, and a CentOS 7 Docker image with EPEL 7 installed. For the latter, not all packages are available in Python 3 yet (which is something Certbot's EPEL package maintainers are working on) and in that case I didn't worry about the system because I think they can/will package the newest version available. If they end up hitting any issues here when trying to package Certbot on Python 3, we can always work with them to fix it.
* remove py27 from oldest name
* update min cryptography version
* remove run_oldest_tests.sh
* upgrade setuptools and pyopenssl
* update cffi, pyparsing, and idna
* expand oldest_constraints comments
* clarify oldest comment
* update min configobj version
* update min parsedatetime version
* quote tox env name
* use Python 3.6 in the oldest tests
* use Python 3.6 for oldest integration tests
* properly pin asn1crypto
* update min six version
* set basepython for a nicer error message
* remove outdated python 2 oldest constraints
* clean up some Sphinx warnings
* first attempt at a doc-test pipeline job
* fix formatting
* fix test name
* set env for bash
* try bash vs script
* maybe it didn't like me setting 'PATH'...derp
* drop use of venv
* sphinx-build isn't a py script
* try activating venv
* docs: remove unused html_static tags
* clean up final sphinx build errors for certbot
* clean up final sphinx build errors for acme
* better names for docs pipeline
* fix spelling
* add docs_extras to setup.py
* remove temp doc-testing pipeline; add template to main.yml
* rearrange pipeline execution; run sphinx builds in one job
* add documentation note to compat.os
* add uninstall.rst as a sub-toctree to avoid build error
Now that we have a new pipstrap script with recent version of pip, dependencies for Windows can be resolved correctly on Python 3.8.
This PR enables tests on Python 3.8, and package Certbot for Windows on Python 3.8 also. I do not move up to Python 3.9 since some dependencies (`cryptography`, `pynacl`) do not provide wheels for Python 3.9 yet on Windows, which would require a complete C++ build system to compile them.
* Enable windows tests on Python 3.8 and package it on Python 3.8 also.
* Upgrade pynsist, nsis and pywin32, remove old workarounds
Co-authored-by: Brad Warren <bmw@users.noreply.github.com>
