Skip to content

Dynamic Namespaces #1132

Open
Open
@StartAutomating

Description

@StartAutomating

Dynamic Namespaces

Why should a dynamic language have static namespaces?

If a developer can dynamically define a variable, alias, or function with literally any name, and items can be queried efficiently enough, it makes a startling amount of sense to be able to dynamically define a namespace, too.

PipeScript should embrace dynamic namespacing.

That is, it should be easy to define a namespace in PipeScript, and the namespaces should be as dynamic as possible.

To that end, let us start to think of a dynamic namespace as "a pattern for names".

Justification

In many ways, it already has (see #452). Modules can define CommandTypes, which use a name and pattern to identify themselves.

This is a somewhat awkward name, as it overlaps with the built-in concept of Command Types.

By starting to call these namespaces, the concept should be clear to most programmers.

By letting them be as dynamic as the language they live in, they should be incredibly powerful.

Architecture

A namespace should be able to be self-described in any [PSObject], provided it has the typename 'Namespace'.

Core namespace support will be implemented in .types.ps1xml

Then any object can choose to become a namespace, simply by adding Namespace to it's .pstypenames.

Putting it very simply:

Once namespaces are easy to declare, they are easy to work with.

Implementation

Dynamic namespaces will bring some core changes to PipeScript:

Any namespace will be able to perform the following methods:

Any namespace will have the following ScriptProperties:

Any namespace will have the following ScriptMethods:

Namespaces should have fairly good formatting:

Metadata

Metadata

Labels

Projects

No projects

Milestone

No milestone

Relationships

None yet

Development

No branches or pull requests

Issue actions