Migrate lint/format tooling to Ruff, update CI matrix, and stabilize cross-platform/firefox tests

[yuriverweij]

This branch replaces legacy lint/format tooling with Ruff, updates CI and runtime compatibility targets, and cleans up platform/browser-specific test behavior so checks are consistent across environments.

Running the utests, atests, linting, and formatting should now work consistently on Windows and Unix based devices.

Major changes

• Replaced Flake8 and Black with Ruff for linting and formatting.
• Added/updated Invoke tasks for linting and formatting in tasks.py
• ❗Added/updated Ruff lint and format checks to GitHub workflow
• ❗ Configured lint/format workflow to run both format and lint steps, then fail the job if either step fails.
• ⚠️ Re-enabled Firefox testing in CI and marked one unstable mouse-over acceptance case as a known issue for Firefox in mouse.robot
• ⚠️ Re-enabled previously skipped Windows unit tests and removed unneeded import, see Review restriction on running ApprovalTests under Windows #1884
• Updated minimum supported Python version to 3.10 and aligned GitHub workflows accordingly.
• Refactored type hints to Python 3.10+ syntax (for example, union operator style).
• Replaced collections.namedtuple usages with typing.NamedTuple in affected unit-test helpers/constructors.
• Applied import-order and formatting cleanups required by Ruff rules.
• Added SKIP_ON_WINDOWS tag for some atests that would not evaluate successfully due to Windows file paths and logging changes.

Documentation updates

• Updated contributor guidance for new lint/format commands in CONTRIBUTING.rst

Behavioral impact

• No intended product/runtime behavior changes in SeleniumLibrary keywords.
• Main impact is development workflow, CI behavior, typing/style modernization, and test-platform handling.

CI and quality checks

• Ruff format check now runs in check mode with diff output.
• Ruff lint check runs in the same workflow.
• Workflow reports both step outcomes and fails if one or both fail.
• Firefox is re-enabled with known-issue handling for the unstable case.

Notes

• atest/ folder linting is not enabled yet, needs work. (formatting is enabled)
• The Firefox known issue should be tracked for eventual removal once root cause is resolved.
• SKIP_ON_WINDOWS tests are still valid, but produce different results on windows, or fail because of unix-oriented paths in the tests. This concerns ~12 tests and I thought the risk minimal to skip these on Windows.

[Copilot]

Pull request overview

This PR migrates the project’s Python lint/format tooling to Ruff, updates CI/runtime compatibility targets (Python 3.10+), and adjusts unit/acceptance tests so results are more consistent across platforms/browsers (notably Windows + Firefox).

Changes:

• Replaced Black/Flake8 with Ruff (config in pyproject.toml, dev deps, Invoke tasks, and a dedicated GitHub Actions workflow).
• Updated CI matrices and acceptance-test runner behavior (Windows exclusion tags, Firefox re-enabled with known-issue tagging).
• Modernized codebase typing/style for Python 3.10+ and refactored tests/helpers accordingly.

Reviewed changes

Copilot reviewed 98 out of 100 changed files in this pull request and generated 8 comments.

Show a summary per file

File	Description
utest/test/utils/test_xpath_escape.py	Removes Windows skip and updates imports around approval tests.
utest/test/utils/test_types.py	Updates list construction style compatible with Ruff formatting.
utest/test/utils/test_package.py	Converts unittest assertions to pytest-style asserts.
utest/test/translation/test_translation.py	Import ordering + pytest fixture decorator normalization.
utest/test/translation/init.py	Package marker for translation tests.
utest/test/robotframework_seleniumlibrary_translation_list/init.py	Formatting cleanup for translation list helper.
utest/test/robotframework_seleniumlibrary_translation_fi/init.py	Formatting cleanup for FI translation helper.
utest/test/locators/test_windowmanager.py	Moves to pytest.raises/assert style and minor mock behavior adjustments.
utest/test/locators/test_elementfinder.py	Ruff-driven import cleanup, pytest fixture normalization, stronger exception matching.
utest/test/keywords/test_windowmananger_window_info.py	Import order fixes and pytest-style asserts.
utest/test/keywords/test_webdrivercache.py	Converts exception testing to pytest.raises and updates asserts/imports.
utest/test/keywords/test_waiting_stale_element_refereance_exception.py	Import order cleanup.
utest/test/keywords/test_tablekeywords.py	Import ordering cleanup.
utest/test/keywords/test_selenium_service_parser.py	Removes Windows skips and refactors assertions/formatting for approval tests.
utest/test/keywords/test_selenium_options_parser.py	Removes Windows skips and refactors formatting/assertions.
utest/test/keywords/test_screen_shot.py	Import ordering + spacing cleanup.
utest/test/keywords/test_runonfailure_from_lib.py	Import ordering cleanup.
utest/test/keywords/test_press_keys.py	Import ordering and removes Windows skip decorators.
utest/test/keywords/test_keyword_arguments_waiting.py	Removes unused variable in test.
utest/test/keywords/test_keyword_arguments_selectelement.py	Whitespace cleanup.
utest/test/keywords/test_keyword_arguments_formelement.py	Fixture normalization and variable rename to avoid keyword conflicts.
utest/test/keywords/test_keyword_arguments_element.py	Import ordering and formatting; adjusts ActionChains mocking formatting.
utest/test/keywords/test_keyword_arguments_browsermanagement.py	Import ordering and pytest-style asserts.
utest/test/keywords/test_javascript.py	Removes Windows skips and approval-test gating for Windows.
utest/test/keywords/test_input_text_file_decorator.py	Adjusts asserts to use is None.
utest/test/keywords/test_firefox_profile_parsing.py	Removes Windows skip and normalizes attribute detection formatting.
utest/test/keywords/test_expectedconditions.py	Renames test class and method; formatting.
utest/test/keywords/test_cookie.py	Fixture decorator normalization and whitespace cleanup.
utest/test/keywords/test_click_modifier.py	Tightens exception assertions using match=.
utest/test/keywords/test_browsermanagement.py	Import ordering, better error assertions, formatting and variable naming.
utest/test/keywords/IGNOREtest_webdrivercreator_service_log_path.py	Replaces namedtuple with NamedTuple; formatting and stubbing updates.
utest/test/keywords/IGNOREtest_webdrivercreator_executable_path.py	Mockito import ordering and formatting cleanup.
utest/test/keywords/IGNOREDtest_webdrivercreator.py	Mockito import ordering, stronger exception matching, minor lint suppressions.
utest/test/entry/test_entry_point.py	Import ordering cleanup.
utest/test/api/test_plugins.py	Converts helper namedtuple to NamedTuple; converts many asserts to pytest style.
utest/test/api/test_plugin_keyword_tags.py	Updates plugin class imports and converts assertions to pytest style.
utest/test/api/test_plugin_documentation.py	Removes Windows skips to allow running under normalized LF files.
utest/test/api/test_filepath_unusual_characters.py	Removes Windows skip to allow running under normalized LF files.
utest/test/api/test_event_firing_webdriver.py	Adds pytest.raises match assertions and converts asserts to pytest style.
utest/test/api/test_accessing_keywod_methods.py	Converts unittest raises/asserts to pytest equivalents.
utest/test/api/plugin_with_event_firing_webdriver.py	Renames class to CapWords and keeps backward-compatible alias.
utest/test/api/plugin_tester.py	Renames class to CapWords and keeps backward-compatible alias.
utest/test/api/my_lib_wrong_name.py	Renames class to CapWords.
utest/test/api/my_lib_not_inherit.py	Renames class to CapWords and keeps backward-compatible alias.
utest/test/api/my_lib_args.py	Renames class to CapWords and keeps backward-compatible alias.
utest/test/api/my_lib.py	Renames class and doc headings; keeps backward-compatible alias.
utest/test/api/approved_files/PluginDocumentation.test_many_plugins.approved.txt	Updates approved output to reflect renamed plugin classes/headings.
utest/run.py	Switches to pytest.main() and logs suppressed errors via logging.
tasks.py	Adds Ruff-based lint/format Invoke tasks; updates atest command to use sys.executable.
src/SeleniumLibrary/utils/types.py	Python 3.10 typing, minor formatting/flow simplification.
src/SeleniumLibrary/utils/events/event.py	Simplifies Selenium major version parsing.
src/SeleniumLibrary/utils/events/init.py	Import ordering and __all__ ordering cleanup.
src/SeleniumLibrary/utils/init.py	Consolidates type exports import and adjusts noqa placement.
src/SeleniumLibrary/locators/windowmanager.py	Replaces namedtuple with NamedTuple; simplifies control flow and naming.
src/SeleniumLibrary/locators/elementfinder.py	Uses PEP604 union types; improves exception chaining; minor simplifications.
src/SeleniumLibrary/locators/customlocator.py	Uses callable() and simplifies return flow.
src/SeleniumLibrary/keywords/window.py	PEP604 typing, import ordering, minor logic simplification and doc cleanup.
src/SeleniumLibrary/keywords/webdrivertools/webdrivertools.py	Refactors browser mapping, exception chaining, formatting, and service/options parsing.
src/SeleniumLibrary/keywords/webdrivertools/sl_file_detector.py	Simplifies boolean return and ensures explicit None return from _get_sl.
src/SeleniumLibrary/keywords/waiting.py	PEP604 typing; adjusts control flow for limit is None branches.
src/SeleniumLibrary/keywords/selectelement.py	PEP604 typing, simplifies return logic, stricter zip formatting.
src/SeleniumLibrary/keywords/screenshot.py	PEP604 typing, formatting cleanups, explicit None returns when no browser is open, PDF printing improvements.
src/SeleniumLibrary/keywords/runonfailure.py	PEP604 typing and refactor of keyword resolution logic.
src/SeleniumLibrary/keywords/javascript.py	Replaces namedtuple with NamedTuple; simplifies marker parsing flow.
src/SeleniumLibrary/keywords/formelement.py	PEP604 typing and exception chaining improvements.
src/SeleniumLibrary/keywords/expectedconditions.py	Improves error handling + messages; formatting and typing updates.
src/SeleniumLibrary/keywords/element.py	PEP604 typing; replaces namedtuple with NamedTuple; formatting and minor refactors.
src/SeleniumLibrary/keywords/cookie.py	PEP604 typing; simplifies branching; minor formatting adjustments.
src/SeleniumLibrary/keywords/browsermanagement.py	PEP604 typing; import reorg; exception chaining and warning message formatting.
src/SeleniumLibrary/keywords/alert.py	PEP604 typing; exception chaining; minor message formatting cleanup.
src/SeleniumLibrary/errors.py	Formatting-only adjustment.
src/SeleniumLibrary/entry/translation.py	PEP604 typing; local import with Ruff suppression; list type hint update.
src/SeleniumLibrary/entry/get_versions.py	Import ordering cleanup.
src/SeleniumLibrary/entry/main.py	PEP604 typing; replace prints with click.echo; minor cleanup.
src/SeleniumLibrary/base/librarycomponent.py	PEP604 typing and import ordering.
src/SeleniumLibrary/base/context.py	PEP604 typing and return type update for element lists.
src/SeleniumLibrary/init.py	PEP604 typing; replaces namedtuples with NamedTuples; refactors parsing helpers and messaging.
setup.py	Bumps python_requires to >=3.10.
requirements-dev.txt	Replaces Black/Flake8 with Ruff and updates dev tool pins.
pyproject.toml	Adds Ruff lint/format configuration.
docs/index.html	Updates documented supported Python versions.
atest/run.py	Adds Windows exclusion tag handling; formatting cleanup.
atest/resources/testlibs/get_driver_path.py	Minor formatting/quoting updates.
atest/acceptance/open_and_close.robot	Adds SKIP_ON_WINDOWS to a flaky/invalid-on-Windows case.
atest/acceptance/multiple_browsers_service.robot	Adds SKIP_ON_WINDOWS and adjusts known-issue tags.
atest/acceptance/multiple_browsers_options.robot	Adds SKIP_ON_WINDOWS tags to Windows-incompatible option parsing tests.
atest/acceptance/keywords/textfields.robot	Makes Chrome-only option conditional to avoid failures on non-Chrome browsers.
atest/acceptance/keywords/page_load_timeout.robot	Adds SKIP_ON_WINDOWS tag.
atest/acceptance/keywords/mouse.robot	Marks unstable mouse-over case as known issue for Firefox.
atest/acceptance/entry_point.robot	Normalizes path separators for cross-platform execution.
atest/acceptance/create_webdriver.robot	Adds SKIP_ON_WINDOWS tag.
atest/acceptance/2-event_firing_webdriver/event_firing_webdriver.robot	Adds SKIP_ON_WINDOWS tag.
atest/acceptance/1-plugin/OpenBrowserExample.py	Formatting for long signature line.
README.rst	Updates documented supported Python versions.
CONTRIBUTING.rst	Updates contributor guidance for Ruff + Invoke tasks, and adjusts docs task name.
.github/workflows/Select.yml	Expands matrix (adds Firefox) and updates actions/steps.
.github/workflows/LintFormatCheck.yml	Adds Ruff lint/format workflow using Invoke tasks.
.github/workflows/CI.yml	Updates Python matrix targets and browser setup actions.
.gitattibutes	Adds LF normalization rule for approval test approved files (but filename is misspelled).
.flake8	Removes legacy Flake8 configuration.

Comments suppressed due to low confidence (1)

.github/workflows/Select.yml:47

• This step is labeled “Setup ${{ matrix.config.browser }} browser” but always runs browser-actions/setup-chrome, even when the matrix browser is firefox. Either make browser setup conditional (Chrome vs Firefox) or rename the step so it doesn’t imply it’s configuring the selected browser.

    - name: Setup ${{ matrix.config.browser }} browser
      uses: browser-actions/setup-chrome@v2
      with:
        chrome-version: latest
        install-dependencies: true
      id: setup-chrome

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[Copilot]

Pull request overview

Copilot reviewed 97 out of 99 changed files in this pull request and generated 8 comments.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[emanlove]

If this test is designed to check chrome then should we be skipping it instead of running a non-test?

[yuriverweij]

Not a test, A keyword that is being used bu the tests, in case of Firefox/non-chrome this gives an error, since experimental options does not exist.

[emanlove]

What is the RF standard for dealing with paths? I know I had once recommended switching / to ${/} but I recall someone else giving a strong argument for these to remain / noting that it will work for both Windows and Linux/MacOS (although I don't recall the underlying mechanisms. I thought this would fail on windows).

[emanlove]

Less concerned about this specific PR. I am good with changing this but trying to understand when we lint what the recommended results should be.

[yuriverweij]

I think for most paths I would (and did) not change this to ${/}, since windows will be able to handle /. In this specific case it did make a difference because it's output from the stdout that is string compared in the test.

[emanlove]

Why are these skip on windows?

[emanlove]

[An aside, I had always wanted an easy to document but clear documented explanation as to why for wa particular test we would want to skip on a certain OS. Or if a known issues what change would trigger a reevalaution for the skip/ignore reason. That way when one came along latter one could evaklute if this was still the case.]

[yuriverweij]

These atests test on specific lines in the logging. Due to Windows specific stuff there are additional debug lines at the start which messes these up. So we could fix them, but this would mean the assertions will have to change quite a lot. Since this is not windows-specific functionality I figured this would not pose much of a risk if we skip these on Windows.

[emanlove]

I don't like this Linter rule. I think ity make it less clear. Does the linter accept ... instead?

[yuriverweij]

No, also not accepted.
I can add # noqa: PIE790 to suppress the linting rule here.

[emanlove]

Does this affect any downstream usage of this method? Or is all this doing is explicitly setting the return to None?

[yuriverweij]

Does not affect at all. Just explicitly returns None, instead of doing it anyway if the for/if loop does not return anything. So normally would return None anyway if it reached here, but difficult to debug since it's not explicit.

[emanlove]

Do we want this as a class attribute (as was) or should it be an instance attribute (as changed to)?

[yuriverweij]

You're right, this would make more sense as a class attribute.

[emanlove]

leaving note to self to relook at this ..

[emanlove]

Is this a linter syntax or an AI syntax?

[yuriverweij]

Fom linter (rule B904), basically to better debug where the exception originated. In this case: instead of getting an assertion error that has something timeout in it. you'd get the explanation with it that the timeout exception caused the assertion error, preserving the full traceback

[emanlove]

Why does it in one instance put all the arguments into one line. And then elsewhere want one argument per line?

[yuriverweij]

Basically comes down to 80 character line limit. If it exceeds it splits all the arguments to its own line. (all or nothing)

[emanlove]

This is an example oif why i think I like the syntax Optional[str] over str | None. That is if those are reversed as here, None | str that would seem to have a different meaning.

Maybe these are equivalent anyways and it is just that in the Optional case it has a default value .. 🤔

[yuriverweij]

These are 100% equivalent. None | str, str | None, and Optional[str] all produce the same type at runtime. Union order is irrelevant for type checking.

Optional[str] is literally just an alias for Union[str, None] and the UP007 rule converts that to str | None

Readability and convention say to put primary type (str) first before None, so we could update that, but no functional difference.