Technical Architecture: Shell Autocompletion

Overview

This technical architecture document describes the design of subcommand-level help dispatch in C++23 and the completion generator logic in Bash.

C++23 Subcommand Help Dispatch

In src/app.cpp, App::run() checks for -h or --help in subcommand argument vectors before executing command logic:

bool is_help_requested(const std::vector<std::string> &args) {
  for (const auto &arg : args) {
    if (arg == "-h" || arg == "--help") {
      return true;
    }
  }
  return false;
}

When is_help_requested() is true, execution routes to dedicated help functions:

print_login_help()
print_chat_help()
print_msg_help()
print_send_help()
print_topic_help()

Bash Autocompletion Engine

File: completions/grm.bash

The script hooks into Bash completion via complete -F _grm_completions grm:

Top-Level Commands: Completes login, chat, msg, send, topic, and global flags (-h, --help, -V, --version, -v, --verbose, -d, --debug, -q, --quiet, -c, --config).
Subcommands & Flags:
- grm login $rightarrow$ -p, --phone, -k, --code
- grm chat $rightarrow$ ls
- grm msg $rightarrow$ ls, export, search
- grm msg export <chat_id> $rightarrow$ csv, json
- grm send $rightarrow$ file
- grm send file $rightarrow$ -C, --caption, -t, --topic
- grm topic $rightarrow$ ls