[Dashboard] [Controls] Fix unsaved changes bug on empty dashboard #155648
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Heenawter
added
bug
💚 Build Succeeded
Metrics [docs]Async chunks
Unknown metric groupsESLint disabled line counts
Total ESLint disabled count
To update your PR or re-run it, just comment with: cc @Heenawter |
|
Pinging @elastic/kibana-presentation (Team:Presentation) |
cqliu1
approved these changes
Apr 24, 2023
Heenawter
added a commit
that referenced
this pull request
Apr 28, 2023
Closes #155777 ## Summary The `"replaces panel with incoming embeddable if id matches existing panel"` test was **not** the flaky one, even though it was the one that showed the failure; it was actually one of the previous `"pulls state <...>"` tests that was failing, which was pretty confusing to catch 🤦 Basically, because we are [running the unsaved changes check on load now](#155648), depending on the timing of the `debounce` on the `checkForUnsavedChangesSubject$` subscription it would sometimes run for the `"pulls state <...>"` tests (but not always) - so whenever it **would** run, because the mocked `loadDashboardStateFromSavedObject` was only returning a partial Dashboard input, this would result in trying to get the property of `undefined` when checking for filter changes, panel changes, etc. This is fixed by ensuring that the `loadDashboardStateFromSavedObject` returns a complete Dashboard input, with all the required keys, for all of the relevant tests. Note that, since you can't test Jest unit tests using the flaky test runner, I was able to run it multiple times by surrounding all the tests with the following in order to ensure that it was no longer flaky: ```typescript for (const i of Array(x) .fill(null) .map((_, i) => i)) { describe(`test run ${i}`, () => { <...> // all the tests go here }); }; ``` I did this with `x = 200`, and the entire test suite passed 200 times in a row 👍 ### Checklist - [x] [Unit or functional tests](https://www.elastic.co/guide/en/kibana/master/development-tests.html) were updated or added to match the most common scenarios ### For maintainers - [ ] This was checked for breaking API changes and was [labeled appropriately](https://www.elastic.co/guide/en/kibana/master/contributing.html#kibana-release-notes-process)
kibanamachine
pushed a commit
to kibanamachine/kibana
that referenced
this pull request
Apr 28, 2023
Closes elastic#155777 ## Summary The `"replaces panel with incoming embeddable if id matches existing panel"` test was **not** the flaky one, even though it was the one that showed the failure; it was actually one of the previous `"pulls state <...>"` tests that was failing, which was pretty confusing to catch 🤦 Basically, because we are [running the unsaved changes check on load now](elastic#155648), depending on the timing of the `debounce` on the `checkForUnsavedChangesSubject$` subscription it would sometimes run for the `"pulls state <...>"` tests (but not always) - so whenever it **would** run, because the mocked `loadDashboardStateFromSavedObject` was only returning a partial Dashboard input, this would result in trying to get the property of `undefined` when checking for filter changes, panel changes, etc. This is fixed by ensuring that the `loadDashboardStateFromSavedObject` returns a complete Dashboard input, with all the required keys, for all of the relevant tests. Note that, since you can't test Jest unit tests using the flaky test runner, I was able to run it multiple times by surrounding all the tests with the following in order to ensure that it was no longer flaky: ```typescript for (const i of Array(x) .fill(null) .map((_, i) => i)) { describe(`test run ${i}`, () => { <...> // all the tests go here }); }; ``` I did this with `x = 200`, and the entire test suite passed 200 times in a row 👍 ### Checklist - [x] [Unit or functional tests](https://www.elastic.co/guide/en/kibana/master/development-tests.html) were updated or added to match the most common scenarios ### For maintainers - [ ] This was checked for breaking API changes and was [labeled appropriately](https://www.elastic.co/guide/en/kibana/master/contributing.html#kibana-release-notes-process) (cherry picked from commit 5779773)
kibanamachine
added a commit
that referenced
this pull request
Apr 28, 2023
# Backport This will backport the following commits from `main` to `8.8`: - [[Dashboard] Fix flaky Dashboard create tests (#156085)](#156085) <!--- Backport version: 8.9.7 --> ### Questions ? Please refer to the [Backport tool documentation](https://github.com/sqren/backport) <!--BACKPORT [{"author":{"name":"Hannah Mudge","email":"[email protected]"},"sourceCommit":{"committedDate":"2023-04-28T14:14:12Z","message":"[Dashboard] Fix flaky Dashboard create tests (#156085)\n\nCloses https://github.com/elastic/kibana/issues/155777\r\n\r\n## Summary\r\n\r\nThe `\"replaces panel with incoming embeddable if id matches existing\r\npanel\"` test was **not** the flaky one, even though it was the one that\r\nshowed the failure; it was actually one of the previous `\"pulls state\r\n<...>\"` tests that was failing, which was pretty confusing to catch 🤦\r\n\r\nBasically, because we are [running the unsaved changes check on load\r\nnow](#155648), depending on the\r\ntiming of the `debounce` on the `checkForUnsavedChangesSubject# Backport This will backport the following commits from `main` to `8.8`: - [[Dashboard] Fix flaky Dashboard create tests (#156085)](#156085) <!--- Backport version: 8.9.7 --> ### Questions ? Please refer to the [Backport tool documentation](https://github.com/sqren/backport) <!--BACKPORT \r\nsubscription it would sometimes run for the `\"pulls state <...>\"` tests\r\n(but not always) - so whenever it **would** run, because the mocked\r\n`loadDashboardStateFromSavedObject` was only returning a partial\r\nDashboard input, this would result in trying to get the property of\r\n`undefined` when checking for filter changes, panel changes, etc.\r\n\r\nThis is fixed by ensuring that the `loadDashboardStateFromSavedObject`\r\nreturns a complete Dashboard input, with all the required keys, for all\r\nof the relevant tests. Note that, since you can't test Jest unit tests\r\nusing the flaky test runner, I was able to run it multiple times by\r\nsurrounding all the tests with the following in order to ensure that it\r\nwas no longer flaky:\r\n\r\n```typescript\r\nfor (const i of Array(x)\r\n .fill(null)\r\n .map((_, i) => i)) {\r\n describe(`test run ${i}`, () => {\r\n <...> // all the tests go here\r\n });\r\n};\r\n```\r\n\r\nI did this with `x = 200`, and the entire test suite passed 200 times in\r\na row 👍\r\n\r\n### Checklist\r\n\r\n- [x] [Unit or functional\r\ntests](https://www.elastic.co/guide/en/kibana/master/development-tests.html)\r\nwere updated or added to match the most common scenarios\r\n\r\n\r\n### For maintainers\r\n\r\n- [ ] This was checked for breaking API changes and was [labeled\r\nappropriately](https://www.elastic.co/guide/en/kibana/master/contributing.html#kibana-release-notes-process)","sha":"57797731905b0303eec9059e398388826c88fa5f","branchLabelMapping":{"^v8.9.0$":"main","^v(\\d+).(\\d+).\\d+$":"$1.$2"}},"sourcePullRequest":{"labels":["Team:Presentation","loe:days","release_note:skip","impact:critical","v8.8.0","v8.9.0"],"number":156085,"url":"https://github.com/elastic/kibana/pull/156085","mergeCommit":{"message":"[Dashboard] Fix flaky Dashboard create tests (#156085)\n\nCloses https://github.com/elastic/kibana/issues/155777\r\n\r\n## Summary\r\n\r\nThe `\"replaces panel with incoming embeddable if id matches existing\r\npanel\"` test was **not** the flaky one, even though it was the one that\r\nshowed the failure; it was actually one of the previous `\"pulls state\r\n<...>\"` tests that was failing, which was pretty confusing to catch 🤦\r\n\r\nBasically, because we are [running the unsaved changes check on load\r\nnow](#155648), depending on the\r\ntiming of the `debounce` on the `checkForUnsavedChangesSubject# Backport This will backport the following commits from `main` to `8.8`: - [[Dashboard] Fix flaky Dashboard create tests (#156085)](#156085) <!--- Backport version: 8.9.7 --> ### Questions ? Please refer to the [Backport tool documentation](https://github.com/sqren/backport) <!--BACKPORT \r\nsubscription it would sometimes run for the `\"pulls state <...>\"` tests\r\n(but not always) - so whenever it **would** run, because the mocked\r\n`loadDashboardStateFromSavedObject` was only returning a partial\r\nDashboard input, this would result in trying to get the property of\r\n`undefined` when checking for filter changes, panel changes, etc.\r\n\r\nThis is fixed by ensuring that the `loadDashboardStateFromSavedObject`\r\nreturns a complete Dashboard input, with all the required keys, for all\r\nof the relevant tests. Note that, since you can't test Jest unit tests\r\nusing the flaky test runner, I was able to run it multiple times by\r\nsurrounding all the tests with the following in order to ensure that it\r\nwas no longer flaky:\r\n\r\n```typescript\r\nfor (const i of Array(x)\r\n .fill(null)\r\n .map((_, i) => i)) {\r\n describe(`test run ${i}`, () => {\r\n <...> // all the tests go here\r\n });\r\n};\r\n```\r\n\r\nI did this with `x = 200`, and the entire test suite passed 200 times in\r\na row 👍\r\n\r\n### Checklist\r\n\r\n- [x] [Unit or functional\r\ntests](https://www.elastic.co/guide/en/kibana/master/development-tests.html)\r\nwere updated or added to match the most common scenarios\r\n\r\n\r\n### For maintainers\r\n\r\n- [ ] This was checked for breaking API changes and was [labeled\r\nappropriately](https://www.elastic.co/guide/en/kibana/master/contributing.html#kibana-release-notes-process)","sha":"57797731905b0303eec9059e398388826c88fa5f"}},"sourceBranch":"main","suggestedTargetBranches":["8.8"],"targetPullRequestStates":[{"branch":"8.8","label":"v8.8.0","labelRegex":"^v(\\d+).(\\d+).\\d+$","isSourceBranch":false,"state":"NOT_CREATED"},{"branch":"main","label":"v8.9.0","labelRegex":"^v8.9.0$","isSourceBranch":true,"state":"MERGED","url":"https://github.com/elastic/kibana/pull/156085","number":156085,"mergeCommit":{"message":"[Dashboard] Fix flaky Dashboard create tests (#156085)\n\nCloses https://github.com/elastic/kibana/issues/155777\r\n\r\n## Summary\r\n\r\nThe `\"replaces panel with incoming embeddable if id matches existing\r\npanel\"` test was **not** the flaky one, even though it was the one that\r\nshowed the failure; it was actually one of the previous `\"pulls state\r\n<...>\"` tests that was failing, which was pretty confusing to catch 🤦\r\n\r\nBasically, because we are [running the unsaved changes check on load\r\nnow](#155648), depending on the\r\ntiming of the `debounce` on the `checkForUnsavedChangesSubject# Backport This will backport the following commits from `main` to `8.8`: - [[Dashboard] Fix flaky Dashboard create tests (#156085)](#156085) <!--- Backport version: 8.9.7 --> ### Questions ? Please refer to the [Backport tool documentation](https://github.com/sqren/backport) <!--BACKPORT \r\nsubscription it would sometimes run for the `\"pulls state <...>\"` tests\r\n(but not always) - so whenever it **would** run, because the mocked\r\n`loadDashboardStateFromSavedObject` was only returning a partial\r\nDashboard input, this would result in trying to get the property of\r\n`undefined` when checking for filter changes, panel changes, etc.\r\n\r\nThis is fixed by ensuring that the `loadDashboardStateFromSavedObject`\r\nreturns a complete Dashboard input, with all the required keys, for all\r\nof the relevant tests. Note that, since you can't test Jest unit tests\r\nusing the flaky test runner, I was able to run it multiple times by\r\nsurrounding all the tests with the following in order to ensure that it\r\nwas no longer flaky:\r\n\r\n```typescript\r\nfor (const i of Array(x)\r\n .fill(null)\r\n .map((_, i) => i)) {\r\n describe(`test run ${i}`, () => {\r\n <...> // all the tests go here\r\n });\r\n};\r\n```\r\n\r\nI did this with `x = 200`, and the entire test suite passed 200 times in\r\na row 👍\r\n\r\n### Checklist\r\n\r\n- [x] [Unit or functional\r\ntests](https://www.elastic.co/guide/en/kibana/master/development-tests.html)\r\nwere updated or added to match the most common scenarios\r\n\r\n\r\n### For maintainers\r\n\r\n- [ ] This was checked for breaking API changes and was [labeled\r\nappropriately](https://www.elastic.co/guide/en/kibana/master/contributing.html#kibana-release-notes-process)","sha":"57797731905b0303eec9059e398388826c88fa5f"}}]}] BACKPORT--> Co-authored-by: Hannah Mudge <[email protected]>
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Summary
Before
When loading a dashboard with no panels but at least one control, the
diffingMiddlewarewas not being fired - this caused a bug where making selections in the control (which were not saved into the dashboard) would not trigger unsaved changes on reload/refresh:Screen.Recording.2023-04-24.at.10.29.11.AM.mov
After
This was because, when the dashboard had no children, no Redux state changes were being dispatched to the Dashboard container which meant that the middleware was never triggered - this is fixed by adding
startWith(null)(reference) to thecheckForUnsavedChangesSubject$subscription so that it always fires on load.Screen.Recording.2023-04-24.at.11.01.24.AM.mov
Checklist
For maintainers