* chore: 🤖 remove packageManager field from package.json
* chore: 🤖 migrate yarn resolutions to pnpm overrides
* chore: 🤖 generate pnpm-lock.yml by using: pnpm import yarn.lock
* chore: 🤖 remove yarn.lock file
* chore: 🤖 migrate workspace configuration
* chore: 🤖 remove .yarn and .yarnrc.yml config
* chore: 🤖 update .copyright.hcl to ignore root pnpm files
* chore: 🤖 migrate husky pre-commit hook from yarn to pnpm
* chore: 🤖 migrate package.json scripts from yarn to pnpm
* chore: 🤖 update rose package.json & build to use explicit deps
* chore: 🤖 remove unused cp metadata.json command in pjson script
The cp metadata.json to dist was added in:
d19a251
The addons/rose/metadata.json no longer exists so it can't be copied
* chore: 🤖 migrate electron-app package.json scripts to pnpm
* chore: 🤖 create pnpm workspace for electron-app w/ hoisted deps
* chore: 🤖 migrate electron-app yarn lockfile to pnpm lockfile
Used "pnpm import yarn.lock"
* chore: 🤖 migrate electron-app overrides
node-gyp isn't needed since the fork is used via @electron/node-gyp
see: https://github.com/electron/node-gyp?tab=readme-ov-file#node-gyp---nodejs-native-addon-build-tool
* chore: 🤖 add trusted deps to be built to onlyBuiltDependencies
This is a pnpm feature that increases security. By defining these
dependencies they are being trusted to allow install hooks,
including postinstall. See:
https://pnpm.io/settings#onlybuiltdependencies
* chore: 🤖 add electron-app pnpm files to copyright header ignore
* chore: 🤖 remove yarn references from .*ignore files
* chore: 🤖 migrate --yarn to --pnpm for ember-cli-update.json
This isn't clearly documented but something that is supported
and can be found in other ember apps:
https://github.com/search?q=--pnpm+ember-cli-update.json&type=code
* chore: 🤖 migrate ci and related scripts to use pnpm
The cache is migrated to use the cache supported by
actions/setup-node@v4, an action we are already using. This is the
recommended way by pnpm:
https://pnpm.io/continuous-integration#github-actions
This commit also ensures that all github actions are using
consistent versions of actions/setup-node and pnpm/action-setup
* docs: ✏️ update documentation to use pnpm instead of yarn
* chore: 🤖 use explicit asciinema-player dependency
Because we override styles on asciinema-player we depend on their
styles. What might be a minor non-breaking change in their version could
(and did) break our style expectations in our overrides.
* chore: 🤖 run pnpm install and commit lockfile changes
* chore: 🤖 Remove ember-inflector from being overidden
https://github.com/hashicorp/boundary-ui/pull/2828#discussion_r2110380711
* chore: 🤖 remove highlight.js from pnpm overrides
https://github.com/hashicorp/boundary-ui/pull/2828#discussion_r2109847938
* chore: 🤖 remove additional unnecessary files from ignore files
* docs: ✏️ fix pnpm doc links
* chore: 🤖 use corepack to specify packageManager in package.json
* chore: 🤖 let pnpm action in ci use packageManager for version
https://github.com/pnpm/action-setup?tab=readme-ov-file#version
* docs: ✏️ fix broken pnpm website links
* remove rose::button files and imports
* remove rose btn utilities
* remove btn class from app.scss
* delete rose btn styling, component files
* update test
* remove playground.hbs, team agreed
* refactor dummy/app/templates/index.hbs to use hds btn
* remove rose form input component folder
* remove component integration test
* remove _input styling and import ref
* remove import from addons
* remove app.scss form input styling
* replace form.input with Hds::Form::TextInput::Field
* remove assertions for disabled state, remove Hds replacement
* remove test that checks input field that no longer exists
* remove rose::form::input component from playgrouund.hbs
* remove @use 'input'
* remove rose::icon test
* remove .rose-icon class styling from ui/admin
* remove .rose-icon class styling from ui/desktop
* remove .rose-icon class styling from addons/
* remove @import 'icon'
* remove icon.js
* feat: 🎸 scope picker component (#2646)
* feat: 🎸 scope picker component
---------
Co-authored-by: lisbet-alvarez <lisbet-alvarez@users.noreply.github.com>
Co-authored-by: Zhihe Li <zhihe.li@hashicorp.com>
* refactor: 💡 use hds side nav w/ hds frame instead of rose (#2688)
* refactor: 💡 use hds side nav w/ hds frame instead of rose
* test: 💍 clean up selectors
* fix: 🐛 specificity of style override
* refactor: 💡 use hds side nav list instead of target and portal
* refactor: 💡 revert to using portal component & update scss
* docs: ✏️ update comment
* refactor: 💡 use custom class names for containers
* refactor: 💡 pr feedback (use @hasSidebar arg)
* refactor: 💡 fix onboarding page & fix settings title
* refactor: 💡 remove unneeded template file (pr feedback)
* test: 💍 fix tests due to merge conflicts
* test: 💍 add missing test selector due to merge conflict
* test: 💍 fix more broken tests due to merge conflicts
* chore: 🤖 update yarn.lock & update hds version to 4.18.1
* chore: 🤖 fix e2e test cases due to side nav refactor (#2751)
* chore: 🤖 fix e2e test cases due to side nav refactor
* test: 💍 fix e2e test for desktop
---------
Co-authored-by: lisbet-alvarez <lisbet-alvarez@users.noreply.github.com>
Co-authored-by: Zhihe Li <zhihe.li@hashicorp.com>
* chore: 🤖 add test selectors to roles list tests
* chore: 🤖 add test selectors for roles global-scope tests
* chore: 🤖 add TABLE_RESOURCE_LINK to global-scope tests
* test: 💍 change exists to isVisible
There was an issue in tests where events fired in
Hds::FlyOut and Hds::Modal were not being dispatched.
These close events were depended on by test waiters
to know when the element was removed. In recent
versions of chrome, when ran headless and on macos,
these events were not firing. This meant the tests
waited indefinitely on developers local machines,
until the tests timed out. This issue was seen
across multiple macbooks and even on the hds repo.
This issue was *only* experienced in headless
chrome on recent macos.
It was **not** happening in:
* macos chrome headed locally
* linux chrome headless in CI
This made the issue only impact people running tests
locally, getting obscure timeouts. This issue is fixed
by the addition of this flag:
--disable-backgrounding-occluded-windows
The other flags added in this commit are not required
for the fix but should help keep a browser "active",
which was loosely what researching this bug hinted at
the cause of finding the issue fix.
* chore: 🤖 add .eslintignore for e2e-tests directory
* chore: 🤖 add auto-fixable eslint changes
* chore: 🤖 make js tasks more explicit to avoid race conditions
avoid having js files be in two separate concurrent branches where tasks
could result in race-condition writes
* chore: 🤖 expand .eslintignore to include more generated dirs
* chore: 🤖 fix additional lint-staged tasks
Fix additional lint-staged tasks that could result in race conditions
* refactor: 💡 Replace rose::separator comp with hds::separator
This PR replaces the rose::separator with the hds::separator and removes
an integration test that is no longer needed
✅ Closes: ICU-15646
* retire rose::separator and remove styling
* remove separator import
* remove separator.js
hashicc
Carlos Alcaide Corvo
Priya Patel
Lauren
Dharini Jeeva
Cameron Perera
Lisbet Alvarez
