* Add kubeconfig string to ClusterCreateRequest
* Update generated files
* Copy and adapt cluster management logic into db
* Add service account deletion to db
* Return errors from new DB methods
* Adapt InstallClusterManagerRBAC for db
* Update errors in db
* Return error if it occurs from db
* Integrate code to (un)install cluster manager
* Use invalid argument instead of failed precondition
* Set bearer token if error is nil
* Rm cluster RBAC install from CLI
* Rm cluster mgmt install from e2e
* Rm common/install.go
* Move install components into server/cluster, thanks @jessesuen
* Rm unneeded ctx arg
* Restore common/installer.go
* Replace all quoted percent-s with percent-q
* Refactor common/installer.go with error returns
* Return errors rather than exiting fatally
* Return proper number of args
* Slim down cluster methods again
* Simplify, simplify, simplify
* Return gRPC error if RBAC could not be installed
* Issue log entries, not print statements
* Fix log import
* Update generated files
* Refactor
* Major cleanup
* Unmarshal context now
* Put claims check after bearer token insertion
* Initial work to use Kubernetes manifest to create a cluster
* Pass context name now
* Wire up prototype
* Add missing parameter for e2e test
* Just read file directly
* Change how we read cluster server
* Support more attributes from localconfig
* Update generated files
* Support incluster flag
* Comment out unused field for now
* Rm previous NewCluster function
* Unmarshal kube config successfully
* Handle insecure clusters, too
* Use existing logic to get config, thanks @jessesuen
* Revert cluster.go to master version
* Update invocations of RBAC installation
* Fix e2e invocation
* Don't remove management account, thanks @jessesuen
* Fix missing error check in e2e test
* Fix missing clientset arg in e2e fixture
* Create kubeclientset for kubeconfig, thanks @jessesuen
* Make use of dex refresh tokens and store them into local config
* API client will automatically redeem OIDC refresh token if auth token expired.
* Stop the practice of reissuing/resigning non-expiring dex claims in API server.
* Add ExpiresAt seconds
Per NumericDate having resolution of seconds at https://tools.ietf.org/html/rfc7519#page-6
* Rename expires for clarity; update comments
* Don't use different possible values for now
* Use intermediate variable for expires value
* Add pseudocode comments to session manager
* Update password storage
* Factor out LocalUsers
* Fix compile errors
* Add claim checks
* Support expiry on ReissueClaims tokens
* Set location to UTC for tokens
* Add logging for username
* Fix issuedAt type assertion
* Set mtime to UTC location
* Set second param on mgr.Create
* Update output for sync
* Major refactor
* Reduce verbosity
* Reduce duplicated code some more, thanks @jessesuen
* Move printout
* Move printout to success, not failure
* Revert "Move printout to success, not failure"
This reverts commit 3a6863d8f497c02bd381cf9ed6ff4a642c8bdcb5.
* Print final status on success _or_ failure
* Adjust printouts with frankenparameters
* Major refactor of data pipelining, thanks @jessesuen
* Refactor app state change printouts
* Fix number of Sprintf args
* Use previous format for keys, rather than hash
* Rename res => hook for clarity
* Don't print app resources initially, thanks @jessesuen
* Refactor Fprintf call to Fprintln
* Rename waitUntilOperationCompleted, thanks @jessesuen
* Refactor to merge data on update
* Default to updated for new resource states
* Use map for fields that actually change
* Don't let flapping lead to duplicate printouts
* Simplify caching mechanism
* Add ksonnet version to version endpoint
I needed to move config.go out of the cli package to fix a circular dependency.
* Remove ksonnetVersion field from the ArgoCD version struct
* create User Api
* update UsersPasswordRequest to UpdatePasswordRequest
* update UserResponse to UpdatePasswordResponse
* current password only needs to be entered once
* add ability to update password whenneeded
* refactor structire so settingsverifier in settings util
* consolidate more
* move MakeSignature test
* re kick off build
* reretrigger build
* move hook resource state into SyncResult (from operation state)
* fix rollback to use apply based sync
* re-assess sync/health status between each sync phases
* PostSync hook should wait until application is Healthy (resolves#363)
* check for keys on server startup
* move manifest to it's own folder
* revert Gopkg.lock changes
* add default password warnings
* update getting started docs
* remove install dependency from e2e test
* fix test pathing
* readding 02* manifests
* set url to blank as default
* update sso docs
* update getting_started to include namespace
* make defaultSetting internal
* remove extra check, should be caught by settingsMgr.GetSettings() error check
* fix manifests path
* error if configmap is missing, but replace if secret missing
* fix getting started
* set password to hostname
* update comment for initializeSettings
* remove unneeded bitbucket.webhook.uuid
* Gopkg.lock modified
* Add timeout to sync and rollback commands
* Make defaultCheckTimeoutSeconds package-global
* Add cancel to context for rollback, sync
* Add cancel after timeout to rollback/sync
* Assign appName earlier in sync
* Don't unnecessarily reassign ctx
* Use full func for timeout after all
* Try applying wait logic
* Add sync/health flags to rollback/sync
* Slight cleanup to realign with spec
* Clean up, still broken
* Adapt waitUntilOPerationCompleted, thanks @jessesuen
* Log fatal immediately after timeout
* Move fatal log back
* Reduce diff further
* Rm two blank lines
* Add upsert field to cluster create request
* Update generated files
* Add idempotence check to cluster
* Add command-line flag for upsert cluster
* Fix error logging in repository, cluster, application create
* Check Server instead of Name
* Add upsert field for repo creation
* Update generated files
* Update repo specs
* Redact existing repo
* Use one exit point to reduce chance of redaction error
* Set error to nil when appropriate
* Handle claims more properly
* Process error better
* Fix comparison, rm unneeded assignment
* Use pointers for repo RBAC name
* Apply repoRBACName in two other places
* DRY
* Don't nest unnecessarily
* Revert repoRBACName change to simplify diff
* Rearrange repo upsert check, thanks @jessesuen
* Issue #295 - implement app destination permissions validation
* Apply reviewer notes. Use project to check application access. Update project access checks
* Use GetProject() instead of project to make sure default value is inferred
* Apply reviewer notes
* Add status field to resource details
* Update generated code
* Set up const message responses
* Check number of resources requiring pruning
* Fix imports
* Use string, thanks @alexmt
* Update generated code
* introduce rbac library around casbin
* supports claims enforcement by iteration through user's groups
* supports filtering of resources by level of access
* policy loader and automatic updates from configmap
* support for builtin and userdefined policies
