Conversation
|
current output: results: can also suppress descriptions for quick view of all commands: comprehensive output: |
|
I'm not entirely sure this is worth it? I feel like the LLM is pretty good at getting the right info from the --help docs and just a single call to I don't think a human will read |
|
the md or the like is definitely worth it. In particular its also worth it to help people make skills etc where they can grab the subcommand text they want. The tool call usage for rv is still not terribly efficient, especially rediscovering. Mangen might also be viable if we can use it to print out in a similar fashion, but i don't have time to explore that too much unless there is a good known reason it would be better. I would not expect many people to actually use/add true manpages to add to the system itself. I also want to use this to autogenerate some docs to dump into rv-docs for each version so people can just get a simple rendering of all the commands |
|
going to merge this puppy and see how it plays out in real life - may revisit, though also going to mark this as experimental formatting wise |
Add rv docs command for programmatic CLI documentation generation.
Motivation
CLI documentation is traditionally discovered incrementally by running --help on each command and subcommand. This creates friction for:
Solution
Leverage clap's introspection APIs to generate complete CLI documentation programmatically:
Design objectives
Structured data for tooling
Bug Fix
Also fixes a missing #[clap(long)] attribute on no_r_environment in MigrateSubcommand::Renv that caused clap to panic when iterating commands.