Add a section to the readme documenting how to shell out to Just at the
end of a recipe, to simulate dependencies that run after a recipe,
instead of before.
This is certainly not as good as having an explicit syntax for this, but
it's a common question so it would be good to document the current best
workaround.
The public documentation is minimal and doesn't change between
platforms, so we only build them for linux on docs.rs to save
their build machines some cycles.
- Complete variable names after `--set`
- Complete recipe names
- Display recipe signature and body below command line
- Modify completions subcommand to produce enhanced zsh completion script
Trim whitespace at beginning and end of generated completions.
Additionally, since some editors will automatically insert a final
newline into text files, make sure a final newline is present.
Add a subcommand that prints out a space-separated list of the names of
top-level variables in the justfile.
The syntax is:
$ just --variables
a b c
This can be used for any purpose, but is mostly intended for completion
scripts, so that they can get the names of variables without using
`--evaluate`.
Additionally:
- Add `bin/generate-completions` script to regenerate checked-in
completions
- Update dependencies
- Regenerate checked-in completions
Shebang recipes have the somewhat confusing property of being quiet by
default, with `@` before a shebang recipe name causing just to print out
the recipe before executing it.
This is somewhat questionable behavior, since it's the opposite of
linewise recipes, but it should be documented, even if it might change
in the future.
- Add a very basic github pages site that links to the github repo.
- Add new install script in the root of the site
- Document the new script in the readme
Make just print clap-generated shell completion scripts with `--completions`
command. Currently, Bash, Zsh, Fish, PowerShell, and Elvish are supported.
Additionally, the generated completion scripts are checked in to the
`completions` folder.
Improve indentation handling in preparation for implementing inline
submodules. This changes the lexer to only parse freeform text inside
the first indent after a ':', so that just can be extended with new
indented constructs which are not recipe bodies. In addition, the lexer
should now handle multiple levels of indentation correctly.
