v0.15.0

• Retry applying same update on top of new resource if diff hasnt changed
  • If resulting diff has not changed, it's guranteed that we are still applying "same" change
• Use JSON instead of YAML for last applied configuration to fit better in max annotation val size
  • It was reported (#48) that some CRDs are very large, hence violate above max val size when serialized as YAML
• Add --dangerous-ignore-failing-api-services flag as a failsafe hatch for deploy/delete/inspect operations
  • This may be helpful when some cluster APIs are misconfigured
• Prevent users from accidently overwriting kapp state records with kapp
  • More on state kept by kapp: https://github.com/k14s/kapp/blob/master/docs/state-namespace.md
• Garbage collect older app changes at the end of kapp deploy
  • setting can be controller via --app-changes-max-to-keep flag
• Add apps.kappctrl.k14s.io/v1alpha1 waiter to integrated with https://github.com/k14s/kapp-controller

Misc

• Fix deploy cmd help to correctly show -a label:name=key usage
• Shrink last change column headers in kapp ls command
• Improved error message when app record (backed by ConfigMap) cannot be parsed
• Use (cluster) instead of <cluster> in UI

Docs

• Add docs for dangerous flags
• Add dev notes with design principles
  • https://github.com/k14s/kapp/blob/master/docs/dev.md
• Add hint in FAQ about anonymous fifo usage to allow stdin to be used for confirmation
• Mention that update strategy may be dangerous to some types of resources
  • https://github.com/k14s/kapp/blob/master/docs/apply.md#controlling-apply-via-resource-annotations

8d3744e9b01c0c2ec4526f5c86ed838854b56806c656b83fc0eff78229ac68f2  ./kapp-darwin-amd64
70bee927591f848f79bdf0888034cb55ed4f6b0e8389aea1f23264effc59160b  ./kapp-linux-amd64
995cb5c9021fa64ba905bcdb09ebbbaa470b78283cfbd7091c99859b5c5a1e34  ./kapp-windows-amd64.exe

(Installation instructions: https://k14s.io/#install-from-github-release)