Rather than hard-coding what a primitive is, we will only look for
templates, and let resourcifier error when kubectl fails to understand
anything else.

Adding missed new file.

bootstrap.sh should look for the namespace file local to the script.

Adds support for "dm deployed-instances <type>:<version>"

When you deploy with <type>:<version>, the type actually seen by DM is
the template URL. While the template URL still shows up as the type when
performing a "dm deployed-types", this at least lets you look for
instances of a known type.

Refactoring the type URL code in client a bit.

Refactoring getTypeUrl.

Deployment name can now be set when doing "dm deploy".

Pretty print JSON responses from service.

Add try/catch handler for Jinja exceptions so we can better annotate errors, to include the file name.

errors, to include the file name.

The default is still either template file name or <type>-<version>.
Override using the "name" flag.

Adds Wordpress example

Simplify Travis Config

Porcelain type deployments deserve to be named too man.

introduce state of a resource to Resource in types.

Put DM into its own namespace

Supports describing a type

fix namespace in install.yaml

Implements correct Expand in API and client

This endpoint takes a template and returns a manifest containing only
the expanded config and layout.

CreateDeployment will now create a deployment before expansion, and will
fail the deployment on expansion error.

1) by type name: dm describe redis:v1
2) by URL: dm describe https://raw.githubusercontent.com/kubernetes/deployment-manager/master/templates/redis/v1/redis.jinja

Describing a type returns the template schema, which includes
descriptive information as well as the schema for properties on the
type.

Treat delete not_found failure as a success

Update code for terminology rationalization.

Update documentation to rationalize the terminology

Do not delete the deployment nor resources when a deployment fails