Skip to main content

Future Plans

In this section, we outline some areas of uncertainty regarding the feasibility of our design in detail. These are the areas we plan to explore next.

Positional Arguments​

Also known as anonymous arguments in core.command.

One of the key areas we need to investigate further is the handling of anonymous arguments, also known as positional arguments in cmdliner and climate. Due to the differences in how these arguments are managed in core.command, translating them may not be as straightforward as it is for named arguments.

However, we have a good intuition that by reducing some of the expressiveness of what we allow to construct, we can solve cases that are sufficient in practice. For example, enforcing that positional arguments be defined in left-to-right order in the specification might be a viable approach.

Tasks​

  • Investigate handling of anonymous arguments in cmdliner and climate (Completed: Aug 2024)
  • Develop a strategy for translating positional arguments in core.command (Completed: Aug 2024)
  • Implement left-to-right order enforcement for positional arguments when compiling to core.command (Completed: Aug 2024)

Targeting stdlib.arg as a runner​

We'd like to write a mini-compiler targeting stdlib.arg as a proof-of-concept showing that it is possible to implement an execution runner for cmdlang that reuses the parsing engine implemented in the standard library.

Tasks​

  • Implemented an execution engine for cmdlang based on stdlib.arg (Completed: Nov 2024)

Generation of Complex Man Pages​

Another area of focus is the generation of complex man pages. cmdliner has excellent support for these. Currently, we have added basic support for one-line summaries of help messages to get started. However, we believe we could reuse most of the design of cmdliner and add it as optional information to the specification language.

The rendering to simple strings could be exported by a standalone cmdliner helper library to reduce dependencies. We could then reuse and integrate this in the translation to core.command and climate. This part is more prospective and may require some coordination with the developers of other libraries.

Tasks​

  • Add support for one-line summaries and readme help pages.
  • Design optional information for specification language based on cmdliner
  • Create a standalone helper library for rendering to simple strings
  • Integrate design into translation to core.command and climate

Auto-Completion​

The third point is by far the most challenging but also a source of significant added value: auto-completion. This has been a highly anticipated feature within the community, and we have recently seen progress in this area, particularly with the climate developers working on integrating support for auto-completion.

We envision a potential hybrid translation mode where a cmdlang command is translated into both cmdliner and climate — cmdliner for runtime execution and climate to generate completion scripts. This approach leaves the question of reentrant completion as future work.

Given the ongoing developments in other libraries regarding auto-completion, we will carefully consider how cmdlang's unique architecture can accommodate and leverage these changes. This will require thoughtful planning and possibly significant adjustments as the landscape evolves.

Tasks​

  • Research auto-completion support in climate
  • Develop a hybrid translation mode for cmdlang commands
  • Implement auto-completion feature