Ankh v2.1.1

• Add Helm 2 warning add TODO for v2 support removal
• Filter out image tags and chart versions
• Fix rollback mode check for jira/slack notifications
• Add new formatting keys for jira/slack notifications

Thanks for the contributions, @dtrupenn and @shortland !

Ankh v2.1.0

• fixed a Nil pointer dereference
• ankh chart create
• Helm 3 compatibility for passing release name
• Match the Helm user experience for packaging/publishing a chart with a version provided on the command line

Ankh v2.0.1

• update formula for 2.0.0
• do a better job at managing newlines and yaml document markers in filter (#80)
• kubectl: fix pod selection with --no-prompt
• ensure chart version included in slack notification (when selected with prompt)
• ensure chart version included in jira summary/description (when selected with prompt)
• support notifying about multiple chart releases at once
• account for chart being a local path instead of remote version

Ankh v2.0.0

Finally leaving beta.

• Fix a bug where --filter was poorly implemented and cut off yaml documents that existed as string literals in some places.
• Fix a bug where pod selection would crash if there were no pods (eg: ankh logs on a chart with no backing pods)

Ankh v2.0.0-rc.0

Fix a few bugs, improve some language, deprecate some behaviors.

Getting closer to a real 2.0

v2.0.0-beta.7

Refactor a lot of things very seriously.

Temporarily break explain while it refactor is in progress.

Fixed a bug where large kubectl output leads to a read deadlock on stdout and stderr.

Support for go modules

Lots of new features that will be finalized in v2.0.0

Ankh v2.0.0-beta.6

• Rename helm registry to the more aptly named repository
• Support registry/repository overrides on the command line when running ankh image ... or ankh chart ...
• Change image parsing code to use upstream docker code, for consistency.
• Add jira ticket support
• Add environment overrides for ANKH_HELM_COMMAND and ANKH_KUBECTL_COMMAND
• Changed logs/exec chart behavior to select pods by match selector instead instead of metadata labels. Label selection was a hack.
• Changed get chart behavior to select objects based on name instead of metadata labels. Label selection was a hack.
• Added more information to pod columns when selecting a pod fo logs/exec etc. This makes it much easier to decide which pod you want to look at, especially during a rollout.

Ankh v2.0.0-beta.5

• Fix --no-prompt to reconcile missing configs
• Fix namespace precedence
• Fix packaging logging during ankh chart publish
• Use /tmp as a DataDir by default. Generate a more random directory name.
• Prompt for a chart when there are neither charts nor dependencies found. Fix up some docs.
• Use a WorkingPath to properly read local charts when resolving dependencies in other directories
• Look for tagKey in DefaultValues to avoid prompting
• The slack message needs to be done after the prompt for versions is done
• We only want it to send one slack message, not one for every context

Ankh v2.0.0-beta.4

• Honor per-chart wildcard labels. Plan on deprecating get/logs/exec etc over an Ankh file that contains multiple charts.
• Gracefully fail when no Helm registry is set and a chart version is missing.
• Only log PUT responses for ankh chart publish with verbose logging.
• Fix an oddity/bug where the generated kubeconfig has a single context, but an empty current-context.
• Refine signal catching to only happen when running tail -f or pods -w.
• Fix a bug where --set would silently (unless verbose logging) ignore key=value pairs where value had extra equal signs.
• Customizable slack messages

Ankh v2.0.0-beta.3

• Read ankh.yaml from within a chart for key metadata
• Add slack support
• Bring back easier local chart support with --chart-path