Change send-coverage-info and send-test-results jobs so it could be triggered on workflow_dispatch event:

Run workflow from here and check if it could trigger those jobs.

Replace BRO_ADDR environment variable with a JSON config file.

• BRO_ADDR environment variable is deprecated, modify the README.md too.
• The JSON config is in:
  • ~/.bro/backend_config.json for the backend server, it contains the following field:
    • listen_addr, listen address for the backend server, defaults to null.
  • ~/.bro/config.json for the CLI client, it contains the following field:
    • backend_addr, backend address for used by the CLI client, defaults to null.
• On running, both the backend server and the CLI client it will do the following:
  • Check if JSON file exists, load if exists, else create a new file.
  • then, for backend server, if listen_addr is null, ask prompt from the user to input the listen address, if user does not input anything, it will be defaults to :320, save config.
  • the same apply for the CLI client, backend_addr, and localhost:320.

Rename this Threeal Bot project into Bro project.

Why?

• Bro is nice, your friendly buddy 🤜 🤛 .
• It's short, you could just run bro to call the client.
• https://www.urbandictionary.com/define.php?term=Bro
• Threeal is my name 👀 .

Use file based database (maybe sqlite?).

• Put database in ~/.bro/log.db (this one only store logs).
• Modify log functions (log.Printf, log.Fatalf, etc.) so it could pass logs to that database.

The logs will be stored for each session (start -> exit), so there are two tables:

• List of session.
• List of logs.

The logs at least should store the following data:

• Time.
• Session Id.
• Topic (could be empty for now).
• Content.

While session for now just store start time of the session.

Rename the following commands:

• client -> bro.
• backend -> bro-backend.

Modify guide in the README.md too.

Currently bro client does not show any command when running bro --help.

It's possibly happen because we use flag package but doesn't specify flag like what it normally should do.
Maybe using other package could fix this too?

Include the source codes that generated by Protobuf to this repository.

Remove this line in .gitignore too:

https://github.com/threeal/threeal-bot/blob/ee17c9b6a8043b737b63ced69868ff3b5b52a986/.gitignore#L2

• Add a systemd config for backend server in this project under service/threeal-bot.service.

  See this for an example of systemd config.

• Add a bash script to install and uninstall that config.

  Install simply means copy service/threeal-bot.service to /lib/systemd/system/.

Utility tests seems failed when ran using gotestsum.

https://github.com/threeal/bro/actions/runs/4040612690/jobs/6946425105

If gotestsum is running the test before generating the test summary (instead of just generating the test summary), maybe let's consider to use gotestsum during the unit-test job (instead of during the send-test-results job).

See this. Apparently, Coveralls still requires this project to be checked out when uploading the coverage report.

Add a status command in the Bro Backend for checking if service of Bro Backend is running.

• This command is available on Linux only. See this.
• This command check if:
  • The config file is exist.
  • The service is running.

Bro config should save the default value after we just press the enter during config prompt.
This way we will only be asked once, only during the initialization of config file.

This line won't work if this script is being called from different directory (service/threeal-bot.service is not portable, should be relative to current script location)
https://github.com/threeal/threeal-bot/blob/42b89bcd65b3cb71b80f7527aef3bbf308370a3c/service/svc.sh#L43

And let's move this script to the project root directory.

Add spin command to run Bro backend process.
So running Bro backend now will use bro-backend spin instead of just bro-backend.

Also, add --help option in bro-backend to list all available options and commands.

Add a config command in both bro and bro-backend cmds.

It will work as follow:

• bro config listen_addr, will output current value for listen_addr.
• bro config listen_addr 12130, will overwrite value for listen_addr to 12130.

So the "Run Tests" step in the test.yml workflow is not failing if the coverage is not 100%. Should be failed so we know if we need to write more tests for functions in those modules.