Standard Toolchain Setup

Guide: https://docs.espressif.com/projects/esp-idf/en/stable/esp32/get-started/linux-macos-setup.html#

ESP_IDF

I directly download the archive with all the submodules included: https://github.com/espressif/esp-idf/releases/tag/v5.3.1

This archive can also be downloaded from Espressif’s download server: https://dl.espressif.com/github_assets/espressif/esp-idf/releases/download/v5.3.1/esp-idf-v5.3.1.zip

cd into the unzip folder (The installation will fail in conda venv)

1
2
3

conda deactivate
./install.sh
. ./export.sh

In ~/.zshrc

1

alias get_idf='. $HOME/esp/esp-idf-v5.3.1/export.sh'

Then each time I need to setup esp32 development environment, I only need to type get_idf.

Some useful commands

1
2
3
4
5
6

idf.py set-target esp32
idf.py menuconfig
idf.py build
idf.py -p /dev/ttyUSB0 -b 115200 flash
idf.py -p /dev/ttyUSB0 -b monitor
idf.py -p /dev/ttyUSB0 flash monitor # combine together

PlatformIO

事实证明pio是最方便的。。
事前安装过pio的vscode插件，直接打开pio的esp32项目就直接可以编译上传以及查看串口监视器。

夹爪控制方式

使用 pyRobotiqGripper

但是只兼容 linux 电脑的串口，所以部署在笔记本上并且创建一个局域网服务器供台式机调用。

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32

import pyRobotiqGripper
import time
from flask import Flask, request

app = Flask(__name__)
gripper = pyRobotiqGripper.RobotiqGripper()
gripper.activate()

@app.route('/open_gripper', methods=['POST'])
def open_gripper():
    # 执行脚本
    # import subprocess
    # subprocess.call(['/path/to/your/script.sh'])
    gripper.open()
    return 'gripper open'
@app.route('/close_gripper', methods=['POST'])
def close_gripper():
    # 执行脚本
    # import subprocess
    # subprocess.call(['/path/to/your/script.sh'])
    gripper.close()
    return 'gripper closed'

if __name__ == '__main__':
    app.run(host='0.0.0.0', port=5000)

# gripper = pyRobotiqGripper.RobotiqGripper()
# gripper.activate()
# gripper.close()
# gripper.open()
# time.sleep(3)
# gripper.close()

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19

(base) cyl@arch ~/450> python gripper_test.py
Activation completed. Activation time :  1.9049019813537598
 * Serving Flask app 'gripper_test'
 * Debug mode: off
WARNING: This is a development server. Do not use it in a production deployment. Use a production 
WSGI server instead.
 * Running on all addresses (0.0.0.0)
 * Running on http://127.0.0.1:5000
 * Running on http://192.168.2.111:5000
Press CTRL+C to quit
192.168.2.103 - - [20/Jul/2024 16:26:01] "POST /open_gripper HTTP/1.1" 200 -
192.168.2.103 - - [20/Jul/2024 16:26:05] "POST /close_gripper HTTP/1.1" 200 -
192.168.2.103 - - [20/Jul/2024 16:26:07] "POST /close_gripper HTTP/1.1" 200 -
192.168.2.103 - - [20/Jul/2024 16:26:10] "POST /close_gripper HTTP/1.1" 200 -
192.168.2.103 - - [20/Jul/2024 16:26:13] "POST /close_gripper HTTP/1.1" 200 -
192.168.2.103 - - [20/Jul/2024 16:26:16] "POST /close_gripper HTTP/1.1" 200 -
192.168.2.103 - - [20/Jul/2024 16:26:18] "POST /close_gripper HTTP/1.1" 200 -
192.168.2.103 - - [20/Jul/2024 16:26:21] "POST /close_gripper HTTP/1.1" 200 -
192.168.2.103 - - [20/Jul/2024 16:26:24] "POST /close_gripper HTTP/1.1" 200 -

台式机通过

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19

# 定义要发送的命令和URL
laptop_ip = "192.168.2.111"
close_url = "http://"+laptop_ip+":5000/close_gripper"
open_url = "http://"+laptop_ip+":5000/open_gripper"

# 使用curl命令通过POST请求发送命令
curl_close_command = [
    'curl', 
    '-X', 'POST', 
    close_url, 
]

curl_open_command = [
    'curl', 
    '-X', 'POST', 
    open_url, 
]

print(subprocess.run(curl_open_command, capture_output=True, text=True))

来控制

After a certain upgrade of hyprland, I can no longer copy&paste across Wayland & XWayland apps, which is very annoying.
The relatied issue: https://github.com/hyprwm/Hyprland/issues/6132
Maybe fixed in: https://github.com/hyprwm/Hyprland/pull/6086

According to 6132 issue, some provide a walk around. I tailored it to adapt it to my system.

Create a shell in ~ directory, named clipsync.sh

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54

#!/usr/bin/env sh
# Two-way clipboard syncronization between Wayland and X11.
# Requires: wl-clipboard, xclip, clipnotify.
#
# Usage:
#   clipsync.sh watch - run in background.
#   clipsync.sh kill - kill all background processes.
#   echo -n any | clipsync.sh insert - insert clipboard content fron stdin.
#
# Workaround for issue:
# "Clipboard synchronization between wayland and xwayland clients broken"
# https://github.com/hyprwm/Hyprland/issues/6132

# Updates clipboard content of both Wayland and X11 if current clipboard content differs.
# Usage: echo -e "1\n2" | clipsync insert
insert() {
  # Read all the piped input into variable.
  value=$(cat)
  wValue="$(wl-paste)"
  xValue="$(xclip -o -selection clipboard)"

  notify() {
    notify-send -u low -c clipboard "$1" "$value"
  }

  if [ "$value" != "$wValue" ]; then
    notify "Wayland"
    echo -n "$value" | wl-copy
  fi

  if [ "$value" != "$xValue" ]; then
    notify "X11"
    echo -n "$value" | xclip -selection clipboard
  fi
}

watch() {
  # Wayland -> X11
  wl-paste --type text --watch "/home/cyl/clipsync.sh" insert &

  # X11 -> Wayland
  while clipnotify; do
    xclip -o -selection clipboard | ~/clipsync.sh insert
  done &
}

kill() {
  pkill wl-paste
  pkill clipnotify
  pkill xclip
  pkill clipsync
}

"$@"

alias in ~/.zshrc:

1

alias clipsync="~/clipsync.sh"

enable it by running

1

clipsync watch

kill all by running

1

clipsync kill

Self-start, configured in ~/.config/hyprland/hyprland.conf

1

exec-once = clipsync watch

REMEMBER TO REMOVE ALL THESE STUFF WHEN HYPRLAND HAS FIXED THE ISSUE

Solved in ISSUE 6086

System used: Archlinux

Pre-requirements

• c language server (for completion, diagnostics), e.g. clangd
• A code editor that use language server, e.g. vscode, vim/neovim

Install Related Softwares

STM32CubeMX

STM32CubeMX is mainly responsible for generating the project with your configuration.

For Distro like Ubuntu/Debain, you can go to the ST official site
Or you can install the software through distro repository

1

yay -S stm32cubemx

For Arch, you need to modify the AUR repository (I mean, maybe the maintainer doesn’t do a good job).
The URL for the repository:https://aur.archlinux.org/packages/stm32cubemx

First clone the repository

1

git clone https://aur.archlinux.org/stm32cubemx.git

Modify the required jdk version in file stm32cubemx.sh
from exec archlinux-java-run --min 17 -- -jar /opt/stm32cubemx/STM32CubeMX "$@" to exec archlinux-java-run --min 17 --max 20 -- -jar /opt/stm32cubemx/STM32CubeMX "$@"

Then build and install the STM32CubeMX

1

makepkg --noconfirm --skipinteg -si

Since STM32CubeMX is not compatible with jdk22 (which is the default jdk that arch is currently using), you need to install jdk17 through yay -S jdk17-openjdk

Then you can start STM32CubeMX by running stm32cubemx, and hopefully, everything is fine.

Compiler

Use arm-none-eabi-gcc

1
2

yay -S arm-none-eabi-gcc
yay -S arm-none-eabi-newlib

Debugger

Use OpenOCD to burn and debug STM32 through STLink v2 (the blue USB device provided by us).

1

yay -S openocd

Setup Your STM32 Project

Open your STM32CubeMX, follow the instruction of Lab1.pdf to configure your project.

NOTE: In Project Manage -> Project -> Project Settings -> Toolchain / IDE, use Makefile/CMake.

Generate the code and go to the project directory (with Makefile/CMakeLists.txt in the directory).

Then you need to generate the compile_commands.json for clangd to recognize the project.

Makefile

1

bear -- make

CMake

1

cmake -S ./ -B ./build

Build Project

Makefile

1

make

Then target binary file is ./build/<Project Name>.bin

CMake

1

cmake --build ./build

Then target binary file is ./build/<Project Name>.elf

Load to STM32F103C8T6

Use OpenOCD to load the binary file to the board.

1

sudo openocd -f /usr/share/openocd/scripts/interface/stlink.cfg -f /usr/share/openocd/scripts/target/stm32f1x.cfg -c "program ./build/<Project Name>.bin reset exit 0x8000000"

Result

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21

Open On-Chip Debugger 0.12.0
Licensed under GNU GPL v2
For bug reports, read
	http://openocd.org/doc/doxygen/bugs.html
Info : auto-selecting first available session transport "hla_swd". To override use 'transport select <transport>'.
Info : The selected transport took over low-level target control. The results might differ compared to plain JTAG/SWD
Info : clock speed 1000 kHz
Info : STLINK V2J37S7 (API v2) VID:PID 0483:3748
Info : Target voltage: 3.222587
Info : [stm32f1x.cpu] Cortex-M3 r1p1 processor detected
Info : [stm32f1x.cpu] target has 6 breakpoints, 4 watchpoints
Info : starting gdb server for stm32f1x.cpu on 3333
Info : Listening on port 3333 for gdb connections
[stm32f1x.cpu] halted due to debug-request, current mode: Thread
xPSR: 0x01000000 pc: 0x08000dc8 msp: 0x20005000
** Programming Started **
Info : device id = 0x20036410
Info : flash size = 64 KiB
** Programming Finished **
** Resetting Target **
shutdown command invoked

NOTE: In different Distro, the cfg file for OpenOCD may locate in different directories. You need to find it by yourselves.

By the way, if you use CMake

Note: When uploading binary file to STM32, it’s recommended to use .bin file instead of .elf file.
Please use the following script to convert the .elf to .bin and upload.

1
2
3
4

cmake -DCMAKE_EXPORT_COMPILE_COMMANDS=ON .
cmake --build  ./build 
arm-none-eabi-objcopy  -O binary -S ./build/*.elf ./build/target.bin
sudo openocd -f /usr/share/openocd/scripts/interface/stlink.cfg -f /usr/share/openocd/scripts/target/stm32f1x.cfg -c "program ./build/target.bin reset exit 0x8000000"

Debug

You have three possible choices. I recommend using Ozone.

openocd+gdb+gdbfrontend(not recommended)

reference:
https://rohanrhu.github.io/gdb-frontend/tutorials/embedded-debugging/

openocd+vscode+PlatformIO (not recommended)

reference:
https://blog.csdn.net/qq_41757528/article/details/127741620

Segger Ozone!!!

reference:
https://blog.csdn.net/weixin_41572450/article/details/124710818

Maybe the best debug tool for stm32

To use segger ozone, you need a different linker called jlink (originally we use st-link v2). You need to buy this linker first (maybe on Taobao or Amazon).

Install Ozone through:

1

yay -S ozone

Setup of ozone project:

• Start Ozone
• Choose Device
  • Device: STM32F103C8
  • Register Set: Cortex-M3
  • Peripherials (optional): /opt/SEGGER/Ozone/Config/Peripherals/STM32F103xx.svd
• Connection Settings
  • Target Interface: SWD
  • Target Interface Speed: 4MHz
  • Host Interface: USB
• Program File: select the binary file you have built (.elf is recommended).

Debug:

set some breakpoints and watch some variables of your interest.
Press the green “power” icon on the upper left corner to start (upload the program and start the debugging process)
Press the blue “play” icon besides “power” to continue.

The source code repository:https://github.com/unpbook/unpv13e

Preparation

1

git clone https://github.com/unpbook/unpv13e

Configure the makefile for your system:

1

CC=gcc CFLAGS=-w CPPFLAGS=-w ./configure    

In archlinux, if you use ./configure directly, you will get Wimplicit compile error in the following steps.

Build the dependence library.

1
2
3
4
5
6

cd lib
make
cd ../libfree
make
cd ../libroute
make

You can test by using the sample program

1
2
3

cd ../intro
make daytimetcpcli
./daytimetcpcli 127.0.0.1

If you get error

1

connect error: Connection refused

You need to install xinetd, configure it and start the service

1
2
3

yay -S xinetd
nvim /etc/xinetd.d/daytime # set `disable` from `yes` to `no`
systemctl start xinetd

And run daytimetcpcli again, you will get something like

1

16 MAY 2024 14:09:07 CST

Then, you are all set.

Which are the build-in commands?

A shell built-in command is contained in the shell itself (implemented by shell author) and runs on the same shell without creating a new process, e.g. cd, pwd, exit, alias.

By contrast: For a regular(linux) command, it is run from a binary file (found in $PATH or specified file) by forking the existing process in a subshell, e.g. ls, cat, rm.

How to implement

1. Judge whether is build-in command

Assume we have already got the parsed arguments

char * parsedArgs[MAXPARSE].

For example, a input char * input = "cd .." can be parsed as char * parsedArgs[MAXPARSE] = {"cd", ".."}

Then, we just need to judge whether parsedArgs[0] is contained in our build-in command set {"exit", "cd", "pwd" ...} (by using strcmp)

2. Implementation

In this project (as of milestone 2) , we need to implement exit, pwd, cd.

exit

In Bash, exit is a built-in command used to terminate the current shell session (same in our mumsh).

Each running shell is a process, so we can use the exit() in stdlib.h to terminate the shell process, achieving the goal of exiting the shell.

1
2
3
4

void exec_exit(){
    printf("exit\n");
    exit(0); // 0 for exit normally, otherwise abnormally
}

pwd

Running a crude (or minimal?) shell and have no idea where you are? Just type pwd and it will print the name of the current/working directory.

In C, we can implement the similar functionality using the getcwd() in unistd.h.

if success, getcwd() will return the pointer to the result, else ( e.g. no enough space) return NULL

1
2
3
4

void exec_pwd(){
    char buf[1024]; // memory space need for storage
    printf("%s\n", getcwd(buf,sizeof(buf)));
}

However, the working directory is not guaranteed to be shorter than 1024, so we may need to dynamically allocate the buffer to suit the length.

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17

int size = 100;
char *buffer = malloc(size);

while (getcwd(buffer, size) == NULL) {
    if (errno == ERANGE) {  // buffer too small
        size *= 2;
        buffer = realloc(buffer, size);
    } else {  // other error
        ...
        free(buffer);
        return 1;
    }
}

printf("%s\n", buffer);

free(buffer);

cd

cd is used for changing working directory.

In C, we can implement the similar functionality using the chdir() in unistd.h.

if success, chdir() will return 0, else ( e.g. no such directory) return -1

1
2
3

void exec_cd(char * path){
    chdir(path);
}

Very straight forward, but if we want to implement more advanced features of the cd command such as changing to the home directory or the previous directory, we would need to handle these cases manually.

For example, you could check if the argument is ~ or -, and then call chdir() with the appropriate path

• Install package
  • For .deb
  • For .AppImage
• Fix package
• Manage process
• Useful command
  • grep
    • option
    • Example

Install package

For .deb

1

sudo dpkg -i *.deb

For .AppImage

1
2

chmod +777 *.AppImage
./*.AppImage

Fix package

1
2
3
4
5

sudo apt -f install
sudo apt update
sudo apt install [dependencies]
...
sudo apt install [package]

Manage process

See all the process:

1

ps aux

Find the target pid

1

pgrep [partial name] 

1

kill [pid]

Useful command

grep

Search for confirmed strings in file

1

grep [option] [pattern] [file]

option

• -i: ignore capitalizing
• -v: inverse search
• -n: show the line num
• -f: reveral search of files in folder
• -i: only print the name od file
• -c: only print the line num

Example

display the battery percentage

1

upower -i /org/freedesktop/UPower/devices/battery_BAT0 | grep percentage

Related software

• Awesome wm
• compfy (picom with animation, no longer maintained)
• polybar with polybar-theme (need many requirements i.e. rofi)
• wezterm

Awesome

github
config
config directory:~/.config/awesome/
doc

youtube

Execute commands on spawn

in the config file rc.lua

1
2
3

awful.spawn("picom --experimental-backend")
awful.spawn("polybar")
awful.spawn.with_shell("xxx.sh")