mirror of
https://github.com/cypress-io/cypress.git
synced 2026-01-09 16:10:11 -06:00
201e9f366e
* chore: set up feature/test-burn-in feature branch * feat: add burnIn Configuration option (currently a no-op) (#27377) * feat: add the burnIn Configuration to the config package. Option currently is a no-op * chore: make burn in experimental * chore: set experimentalBurnIn to false by default * feat: add new experimental retries configuration (#27412) * feat: implement the experimental retries configuration options to pair with test burn in * [run ci] * fix cache invalidation [run ci] * fix snapshot added in v13 for module api to include test burn in experimentalflag * chore: fix merge conflict * chore: add burnInTestAction capability (#27768) * add burnInTestAction capability * feat: add burn in capability for cloud * chore: fix snapshot for record_spec * feat: implement experimental retries (#27826) * chore: format the retries/runner snapshot files to make diff easier * feat: implement experimentalRetries strategies 'detect-flake-and-pass-on-threshold' and 'detect-flake-but-always-fail'. This should not be a breaking change, though it does modify mocha and the test object even when the experiment is not configured. This is to exercise the system and make sure things still work as expected even when we go GA. Test updates will follow in following commits. * chore: update snapshots from system tests and cy-in-cy tests that now have the cypress test metadata property _cypressTestStatusInfo. tests have been added in the fail-with-[before|after]each specs to visually see the suite being skipped when developing. * chore: add cy-in-cy tests to verify reporter behavior for pass/fail tests, as well as new mocha snapshots to verify attempts. New tests were needed for this as the 'retries' option in testConfigOverrides currently is and will be invalid for experiment and will function as an override. tests run in the cy-in-cy tests are using globally configured experimentalRetries for the given tested project, which showcases the different behavior between attempts/retries and pass/fail status. * chore: add unit test like driver test to verify the test object in mocha is decorated/handled properly in calculateTestStatus * chore: add sanity system tests to verify console reporter output for experimental retries logic. Currently there is a bug in the reporter where the logged status doesnt wait for the aftereach to complete, which impacts the total exitCode and printed status. * fix: aftereach console output. make sure to fail the test in the appropriate spot in runner.ts and not prematurely, which in turn updates the snapshots for cy-in-cy as the fail event comes later." * chore: address comments from code review * fix: make sure hook failures print outer status + attempts when the error is the hook itself. * chore: improve types within calculateTestStatus inside mocha.ts * Revert "feat: add burnIn Configuration option (currently a no-op) (#27377)" This reverts commitc428443079. * Revert "chore: add burnInTestAction capability (#27768)" This reverts commitae3df1a505. * chore: run snapshot and binary jobs against experimental retries feature branch * chore: add changelog entry (wip) * Revert "fix snapshot added in v13 for module api to include test burn in experimentalflag" This reverts commitbb5046c91e. * Fix system tests * Clear CircleCI cache * Normalize retries config for test execution * Fixed some unit tests * update snapshots for newer test metadata * Fix cy-in-cy snapshots * update snapshots * bump cache version * chore: ensure legacy retry overrides work; reject exp. retries overrides (#28045) * update changelog * flip if statement in experimental retries option validation * refactor invalid experimental retry override for more useful error msg * revert testConfigOverrides snapshot * update snapshots for test override sys test * Update packages/config/src/validation.ts Co-authored-by: Chris Breiding <chrisbreiding@users.noreply.github.com> * succinct changelog entry; links to docs for details * testConfigOverride system test snapshots * Update .github/workflows/update_v8_snapshot_cache.yml Co-authored-by: Ryan Manuel <ryanm@cypress.io> * Update cli/CHANGELOG.md Co-authored-by: Ryan Manuel <ryanm@cypress.io> * Update packages/driver/src/cypress.ts Co-authored-by: Ryan Manuel <ryanm@cypress.io> * updating cache-version * improve typescript usage when appending experimental retry options to experiments in Experimenets.vue * Revert "improve typescript usage when appending experimental retry options to experiments in Experimenets.vue" This reverts commitb459aba882. * refactor test config override validation for experimental retry subkeys * account for error throw differences in browsers in system tests * bump circle cache * bump circle cache again --------- Co-authored-by: astone123 <adams@cypress.io> Co-authored-by: mabela416 <mabel@cypress.io> Co-authored-by: Muaz Othman <muaz@cypress.io> Co-authored-by: Muaz Othman <muazweb@gmail.com> Co-authored-by: Cacie Prins <cacie@cypress.io> Co-authored-by: Cacie Prins <cacieprins@users.noreply.github.com> Co-authored-by: Chris Breiding <chrisbreiding@users.noreply.github.com> Co-authored-by: Ryan Manuel <ryanm@cypress.io> Co-authored-by: Matthew Schile <mschile@cypress.io>
Server
The server is the heart of the Cypress application. All of this code represents the node process running behind the browser application. This node process is responsible for:
- Proxying every byte coming in and out of the browser
- Performing and normalizing automation tasks for each browser
- Coordinating and synchronizing state with the launchpad and driver packages
- Performing node specific tasks on behalf of the driver
- Instantiating and orchestrating nearly every other layer and package
- Spinning up various static file and http servers
- Communicating with our external API's
- Recording videos of run
- Managing mocha reporters
- Managing 3rd party plugins
The driver and the server are the two most complex packages of Cypress.
Developing
To run the Cypress server:
## boots the entire Cypress application
yarn start
Since the server controls nearly every aspect of Cypress, after making changes you'll need to manually restart Cypress.
Since this is slow, it's better to drive your development with tests.
Building
Note: you should not ever need to build the .js files manually. @packages/ts provides require-time transpilation when in development.
yarn workspace @packages/server build-prod
yarn test-unitexecutes unit tests intest/unityarn test-integrationexecutes integration tests intest/integrationyarn test-performanceexecutes performance tests intest/performance
You can also use the test-watch command to rerun a test file whenever there is a change:
yarn test-watch /test/path/to/spec.js
Running individual unit tests
yarn test <path/to/test>
yarn test test/unit/api_spec.js
## or
yarn test-unit api_spec ## shorthand, uses globbing to find spec
Running individual integration tests
yarn test <path/to/test>
yarn test test/integration/cli_spec.js
## or
yarn test-integration cli_spec ## shorthand, uses globbing to find spec
Running e2e/system tests
With the addition of Component Testing,
e2etests have been renamed tosystem-testsand moved to thesystem-testsdirectory.
Updating snapshots
Prepend SNAPSHOT_UPDATE=1 to any test command. See snap-shot-it instructions for more info.
SNAPSHOT_UPDATE=1 yarn test test/unit/api_spec.js
SNAPSHOT_UPDATE=1 yarn test test/integration/cli_spec.js
V8 Snapshots
In order to improve start up time, Cypress uses electron mksnapshot for generating v8 snapshots for both development and production.
Cypress code is automatically set up to run using snapshots. If you want to run Cypress in development without the v8 snapshot (for debugging purposes or to see if there's a problem with the snapshot or the code itself) you can set the environment variable DISABLE_SNAPSHOT_REQUIRE to 1 or true.