doc/ref/spec: consider definitions for closedness

Definitions were excempted from closedness checks.

With this change, it can be guaranteed that if a template
writer does not make an existing field more specific
and a user doesn't embed the template, the changes
cannot break a user.

Consider this template:
	-- pkg.com/foo/foo.cue --
	#Def: {}

And its usage:

	-- bar.cue --
	import "pkg.com/foo/foo.cue"

	def: #Def & {
	  #foo: 2
	}

The introduction of #foo is currently allowed, as
closedness is not checked for definitions. So
if the original template is no augmented as such

	-- pkg.com/foo/foo.cue --
	#Def: {
	  #foo: 1
	}

it breaks the user.

This means though, that it breaks the property that if
a template writer changes their template, it may
conflict with a definition that a user may have added.

This already holds for regular fields, but that is to be
expected. Users can still add definitions by embedding
the struct. But the same warning applies in this case
for regular fields and definitions, making it more
consistent.

Breakage can also occur if a template writer makes a
template more specific. Again here also, this also holds
for regular fields and this also makes definitions more
consistent with regular fields.

Change-Id: I1e2aedcf6cc4b17f17f9b4a742ca7aac238ae382
Reviewed-on: https://cue-review.googlesource.com/c/cue/+/6721
Reviewed-by: Marcel van Lohuizen <mpvl@golang.org>
1 file changed
tree: 3b7be02758f63326d326ce385ad919e2fc41dc97
  1. .github/
  2. cmd/
  3. cue/
  4. cue.mod/
  5. cuego/
  6. doc/
  7. encoding/
  8. internal/
  9. pkg/
  10. tools/
  11. .dockerignore
  12. .gitattributes
  13. .gitignore
  14. .goreleaser.yml
  15. AUTHORS
  16. codereview.cfg
  17. CONTRIBUTING.md
  18. Dockerfile
  19. gen.go
  20. genpkgtests.go
  21. go.mod
  22. go.sum
  23. LICENSE
  24. README.md
  25. tmp.cue
  26. tools.go
README.md

GoDoc Github Go Report Card GolangCI Go 1.13+ platforms

The CUE Data Constraint Language

Configure, Unify, Execute

CUE is an open source data constraint language which aims to simplify tasks involving defining and using data.

It is a superset of JSON, allowing users familiar with JSON to get started quickly.

What is it for?

You can use CUE to

  • define a detailed validation schema for your data (manually or automatically from data)
  • reduce boilerplate in your data (manually or automatically from schema)
  • extract a schema from code
  • generate type definitions and validation code
  • merge JSON in a principled way
  • define and run declarative scripts

How?

CUE merges the notion of schema and data. The same CUE definition can simultaneously be used for validating data and act as a template to reduce boilerplate. Schema definition is enriched with fine-grained value definitions and default values. At the same time, data can be simplified by removing values implied by such detailed definitions. The merging of these two concepts enables many tasks to be handled in a principled way.

Constraints provide a simple and well-defined, yet powerful, alternative to inheritance, a common source of complexity with configuration languages.

CUE Scripting

The CUE scripting layer defines declarative scripting, expressed in CUE, on top of data. This solves three problems: working around the closedness of CUE definitions (we say CUE is hermetic), providing an easy way to share common scripts and workflows for using data, and giving CUE the knowledge of how data is used to optimize validation.

There are many tools that interpret data or use a specialized language for a specific domain (Kustomize, Ksonnet). This solves dealing with data on one level, but the problem it solves may repeat itself at a higher level when integrating other systems in a workflow. CUE scripting is generic and allows users to define any workflow.

Tooling

CUE is designed for automation. Some aspects of this are:

  • convert existing YAML and JSON
  • automatically simplify configurations
  • rich APIs designed for automated tooling
  • formatter
  • arbitrary-precision arithmetic
  • generate CUE templates from source code
  • generate source code from CUE definitions (TODO)

Download and Install

Install using Homebrew

Using Homebrew, you can install using the CUE Homebrew tap:

brew install cuelang/tap/cue

Install From Source

If you already have Go installed, the short version is:

go get -u cuelang.org/go/cmd/cue

This will install the cue command line tool.

For more details see Installing CUE.

Learning CUE

The fastest way to learn the basics is to follow the tutorial on basic language constructs.

A more elaborate tutorial demonstrating of how to convert and restructure an existing set of Kubernetes configurations is available in written form.

References

Contributing

Our canonical Git repository is located at https://cue.googlesource.com.

To contribute, please read the Contribution Guide.

To report issues or make a feature request, use the issue tracker.

Changes can be contributed using Gerrit or Github pull requests.

Contact

You can get in touch with the cuelang community in the following ways:


Unless otherwise noted, the CUE source files are distributed under the Apache 2.0 license found in the LICENSE file.

This is not an officially supported Google product.