Shell completion

How do I...? · Limitations · API reference

cli-fp can generate completion scripts from your registered command tree. Built-in completion supplies commands, subcommands, options, Boolean values, and enum values; it does not currently support dynamic application callbacks. The generated shell function calls the application's internal __complete entrypoint, so candidates always reflect the command tree in the executable that generated the script.

Bash

Start in the directory containing your compiled myapp executable (replace this placeholder with the actual name, for example QuickStartDemo).

./myapp --completion-file > myapp-completion.bash
source ./myapp-completion.bash

Generate the file after the application is compiled. Source it for the current shell; add a source command to your shell configuration only after confirming the script path is stable.

PowerShell

.\myapp.exe --completion-file-pwsh > .\myapp-completion.ps1
. .\myapp-completion.ps1

The scripts register completion for the executable name and common relative invocation forms. Generated executable names, paths, and metadata are quoted for their target shell, including values containing spaces or shell-sensitive characters. PowerShell cycles candidates with Tab; Bash normally lists matches according to its configured completion behavior. Build and try SimpleDemo for a runnable command/subcommand completion example.

Boundaries

Completion-script requests are handled only when they are the first argument. The first non-option token selects a named command; later non-option tokens can select its subcommands. --help works at application and selected-command scope. Reserved -v/--version requests display the application version at root, named and nested scopes, and are offered in command-option completions. Use -d or long-only --verbose for verbosity; -V is also a version request. Historical completion qualification material is retained in the archive, not in the user guide.