Creating Shell Commands

Here, we will create a sample program that implements a custom shell command.

Building and Flashing the Program

Create a new Pico SDK project named customcmd-definition.

Create Pico SDK Project

1. In VSCode, run >Raspberry Pi Pico: New Pico Project in the command palette.

2. In the dialog below, select C/C++.

3. Create a project with the following settings:

   • Name ... Enter the project name.
   • Board type ... Select your board type.
   • Location ... Select the parent directory where the project directory will be created.
   • Stdio support ... Leave Console over USB unchecked when you use LABOPlatform or other USB features because they conflict with each other. You can enable it by editing the CMakeLists.txt file later.
   • Code generation options ... Check Generate C++ code.

Open Existing Pico SDK Project

Open the project folder in VSCode using one of the following methods:

• In a command prompt, change the current directory to the project folder and execute code ..

• In a Explorer, choose the project folder, push Alt+D to focus the address bar, and execute code ..

  

If the folder is already prepared as a Pico SDK project, just proceed with editing and building the project.

If not, you will see the following message at the bottom right corner of VSCode.

Click Yes and you will see the following window.

Click Import and the project will be prepared as a Pico SDK project.

When the Do you want to import this project as Raspberry Pi Pico project? message disappears before you click Yes, you can reveal it by clicking the icon at the bottom right corner of VSCode.

Clone the pico-jxglib repository from GitHub so the direcory structure looks like this:

├── pico-jxglib/
└── customcmd-definition/
    ├── CMakeLists.txt
    ├── customcmd-definition.cpp
    └── ...

Clone the Repository

Change the current directory to the parent directory where you want to clone the pico-jxglib repository and run the following commands:

$ git clone https://github.com/ypsitau/pico-jxglib.git
$ cd pico-jxglib
$ git submodule update --init --recursive

pico-jxglib is updated almost daily. If you've already cloned it, run the following command in the pico-jxglib directory to get the latest version:

$ git pull

A directory from a Git repository can safely be moved to another location even after cloning.

Add the following lines to the end of CMakeLists.txt:

target_link_libraries(customcmd-definition
    jxglib_Shell jxglib_Serial jxglib_ShellCmd_Basic)
add_subdirectory(${CMAKE_CURRENT_LIST_DIR}/../pico-jxglib pico-jxglib)

Enable UART or USB stdio as described below.

Enable UART and/or USB stdio

Find the following lines in CMakeLists.txt:

pico_enable_stdio_uart(your-project 0)
pico_enable_stdio_usb(your-project 0)

You can set the value to 1 to enable stdio or 0 to disable it.

• pico_enable_stdio_uart() enables UART stdio that uses GPIO0 (UART0 TX) and GPIO1 (UART0 RX).
• pico_enable_stdio_usb() enables USB stdio that uses the USB interface. Make sure to disable USB stdio when you use USB features because they conflict with each other.

Edit customcmd-definition.cpp as follows:

#include "pico/stdlib.h"
#include "jxglib/Serial.h"
#include "jxglib/Shell.h"

using namespace jxglib;

ShellCmd(customcmd, "tests command line arguments")
{
    for (int i = 0; i < argc; i++) {
        tout.Printf("argv[%d] \"%s\"\n", i, argv[i]);
    }
    return Result::Success;
}

int main(void)
{
    ::stdio_init_all();
    Serial::Terminal terminal;
    terminal.Initialize();
    Shell::AttachTerminal(terminal);
    for (;;) {
        Tickable::Tick();
    }
}

Build and flash the program to the board.

Build and Flash

The simplest way to build and flash the program is to use the UF2 file generated by the build process. Pico board, a USB cable, and a computer are all you need to get started. Here are the steps to build and flash the program:

1. Pressing F7 on VSCode will build the project. If this is the first time you build the project, you will see the dialog shown below. Select Pico Using compilers: ... and the build process will start.

   

2. After building, you can find the generated UF2 file in the build directory.

3. Connect your Pico to the computer using a USB cable while holding the BOOTSEL button, and it will appear as a mass storage device. Copy the generated UF2 file to this device to flash it. No need to mind the destination directory, just copy it to the root directory of the device.

If you have a debug probe like this, you can also flash the program using OpenOCD and GDB. This is the recommended method for development, as it allows you to debug the program while running it on the board!

Running the Program

Open a terminal emulator to connect it.

Setup Tera Term for LABOPlatform

Tera Term is available here.

From the menu bar, select [File] - [New Connection...] to open the dialog below:

pico-jxgLABO or a firmware that links jxglib_LABOPlatform provides two USB serial ports: one for terminal use and the other for applications such as logic analyzers and plotters. Select one of the ports and press Enter key in the terminal. When successfully connected, you will see a prompt in the terminal.

L:/>

Run the command with no arguments:

>customcmd
argv[0] "customcmd"

Run the command with some arguments:

>customcmd arg1 arg2 arg3
argv[0] "customcmd"
argv[1] "arg1"
argv[2] "arg2"
argv[3] "arg3"

Run the command with arguments that contain spaces:

>customcmd "A quick brown" "fox jumps over" "the lazy dog"
argv[0] "customcmd"
argv[1] "A quick brown"
argv[2] "fox jumps over"
argv[3] "the lazy dog"

The custom command also appears in the help list:

>help
.               executes the given script file
about-me        prints information about this own program
about-platform  prints information about the platform
customcmd         tests command line arguments
d               prints memory content at the specified address
dump            prints memory content at the specified address
echo            prints the given text
echo-bin        generates binary data from the given arguments
help            prints help strings for available commands
history         prints the command history
logout          logs out the current user
password        changes the password file
prompt          changes the command line prompt
set             set environment variable
sleep           sleeps for the specified time in milliseconds
ticks           prints names and attributes of running Tickable instances

Program Explanation

A shell command is defined using the ShellCmd macro, which has the following syntax:

ShellCmd(symbol, "help")
{
    // command implementation
    return Result::Success;
}

symbol is the name of the command, and "help" is the help string that describes the command. The help string is displayed when the user types help command in the shell.

If the command name contains characters that are not allowed in C++ identifiers, you can use the ShellCmd_Named macro instead, which allows you to specify the command name as a string:

ShellCmd_Named(symbol, "name", "help")
{
    // command implementation
    return Result::Success;
}

The following variables are passed to the command program:

• Printable& tout ... Printable instance. Used to puts texts to the terminal or to a file through redirection.
• Printable& terr ... Printable instance. Used to puts texts to the terminal regardless of redirection.
• int argc ... Number of arguments
• char** argv ... Argument strings. argv[0] contains the command name itself and argv[1] to argv[argc-1] contain the command line arguments.

The function should return Result::Success if there is no error, or Result::Error if an error occurs.

You do not need to register commands. When you create a command with the ShellCmd macro, it is automatically registered with the shell. With this mechanism, you can add commands simply by linking the source file that implements the command to the main program.