Programming language ranking

lambda expression

(parameter) => function body

#include <iostream>
#include <vector>
#include <algorithm>

int main() {
    std::vectornumbers = {1, 2, 3, 4, 5};

    // Use lambda expression to calculate the sum of even numbers
    int sum = 0;
    std::for_each(numbers.begin(), numbers.end(), [&sum](int n) {
        if (n % 2 == 0) sum += n;
    });

    std::cout << "The sum of even numbers: " << sum << std::endl;
    return 0;
}

# Use lambda expression to calculate the sum of two numbers
add = lambda x, y: x + y
print(add(5, 10)) # Output: 15

using System;
using System.Collections.Generic;
using System.Linq;

class Program {
    static void Main() {
        List numbers = new List{ 1, 2, 3, 4, 5 };
        
        // Use a Lambda expression to filter out even numbers and calculate the sum
        int sum = numbers.Where(n => n % 2 == 0).Sum();
        
        Console.WriteLine($"The sum of even numbers: {sum}");
    }
}

Reflection

concept

Common uses

• Serialization/Deserialization (JSON/XML ↔ Object)
• Dependency Injection (DI)/Inversion of Control (IoC)
• ORM (data table ↔ object field correspondence)
• Plug-in/module dynamic loading
• Testing framework (automatic scan testing, Mock)
• Validation and automatic generation of forms/documents

Advantages and Disadvantages

• advantage:High flexibility, can reduce boilerplate code, and supports universal framework design.
• shortcoming:Lower performance, reduced type safety, broken visibility (private members can be accessed), and increased maintenance complexity.

Language support status

• Full support:Python、C#、Java、Kotlin、JavaScript/TypeScript、Ruby、PHP、Go
• Partial support:Swift（Mirror）、Scala
• Not supported (no runtime reflection):C++ (to C++23), Rust (replace with derive macro or custom code generation)

Typical example

# Python: List and read object properties
class User:
    def __init__(self): self.id = 0; self.name = "Ann"
u = User()
for k, v in vars(u).items():
    print(k, v) # id 0 / name Ann

// C#: Get the field/property and read the value
using System;
using System.Reflection;

var t = typeof(MyType);
foreach (var p in t.GetProperties(BindingFlags.Instance|BindingFlags.Public|BindingFlags.NonPublic))
    Console.WriteLine($"{p.Name}={p.GetValue(obj)}");

// JavaScript: dynamic enumeration and calling
const obj = { x: 1, y: 2, add(){ return this.x + this.y; } };
for (const [k,v] of Object.entries(obj)) console.log(k, v);
console.log(obj["add"]()); // 3

// Java: Reading/writing fields using reflection
import java.lang.reflect.*;
class U { private int id = 42; }
U u = new U();
Field f = U.class.getDeclaredField("id");
f.setAccessible(true);
System.out.println(f.getInt(u)); // 42

// Go: reflect visit structure
import "reflect"
func Dump(v any) {
    val := reflect.ValueOf(v)
    for i := 0; i < val.NumField(); i++ {
        name := val.Type().Field(i).Name
        fmt.Println(name, val.Field(i).Interface())
    }
}

best practices

• Only use it when there is a real need for elasticity (framework/tool ​​layer) to avoid abuse of core hot paths.
• Access to sensitive or private members must be carefully controlled and follow the principle of least privilege.
• Add caching (such as caching member information) to reduce reflection overhead.
• Use type annotations/attributes to clarify the intent.

When to avoid

• Performance critical path (can be replaced by pre-generated code or generics/templates).
• When strong type safety and predictable behavior are required (use explicit interfaces/abstractions instead).

Checks whether any structure is all zeros

Concept summary

• Can arbitrary structures be checked "automatically"Depends on whether the language is supportedreflection/introspection（reflection/introspection）。
• Dynamic languages ​​(such as Python, JavaScript, Ruby, PHP) are easiest; static languages ​​require Reflection API (such as C#, Java, Kotlin, Go).
• C++ (up to C++23) and Rust do not have built-in runtime reflection, so you need to manually write checking logic for custom types or use macros/derive to assist.

Language support comparison

Python example

def is_all_zero(obj, eps=1e-6):
    if isinstance(obj, (int, float)): # Basic numerical value
        return abs(obj) < eps
    elif isinstance(obj, (list, tuple, set)): # sequence/set
        return all(is_all_zero(x, eps) for x in obj)
    elif isinstance(obj, dict): # Dictionary
        return all(is_all_zero(v, eps) for v in obj.values())
    elif hasattr(obj, "__dict__"): # General objects
        return all(is_all_zero(v, eps) for v in vars(obj).values())
    else:
        return False

# test
class point:
    def __init__(self, x=0, y=0): self.x, self.y = x, y
class Line:
    def __init__(self, p1=None, p2=None):
        self.p1 = p1 or Point()
        self.p2 = p2 or Point()

print(is_all_zero([[0,0],[0,0]])) # True
print(is_all_zero(Line(Point(0,0),Point(0,0)))) # True
print(is_all_zero(Line(Point(1,0),Point(0,0)))) # False

JavaScript example

function isAllZero(obj, eps = 1e-6) {
  const isNumZero = n => Math.abs(n) < eps;
  if (typeof obj === "number") return isNumZero(obj);
  if (Array.isArray(obj)) return obj.every(v => isAllZero(v, eps));
  if (obj && typeof obj === "object")
    return Object.values(obj).every(v => isAllZero(v, eps));
  return false;
}

console.log(isAllZero([[0,0],[0,0]])); // true
console.log(isAllZero({x:0, y:{z:0}})); // true
console.log(isAllZero({x:0.000001, y:0})); // depends on eps

C# (.NET) Example

using System;
using System.Linq;
using System.Reflection;

public static class ZeroCheck {
    public static bool IsAllZero(object obj, double eps = 1e-6) {
        if (obj == null) return true;
        switch (obj) {
            case int i: return i == 0;
            case long l: return l == 0;
            case float f: return Math.Abs​​(f) < eps;
            case double d: return Math.Abs​​(d) < eps;
        }

        var t = obj.GetType();
        if (t.IsArray)
            return ((Array)obj).Cast<object>().All(x => IsAllZero(x, eps));

        //Scan fields and attributes
        foreach (var f in t.GetFields(BindingFlags.Instance|BindingFlags.Public|BindingFlags.NonPublic))
            if (!IsAllZero(f.GetValue(obj), eps)) return false;

        foreach (var p in t.GetProperties(BindingFlags.Instance|BindingFlags.Public|BindingFlags.NonPublic))
            if (p.CanRead && !IsAllZero(p.GetValue(obj, null), eps)) return false;

        return true;
    }
}

Go examples

package main

import (
    "math"
    "reflect"
)

func IsAllZero(v interface{}, eps float64) bool {
    val := reflect.ValueOf(v)
    switch val.Kind() {
    case reflect.Float32, reflect.Float64:
        return math.Abs(val.Float()) < eps
    case reflect.Int, reflect.Int8, reflect.Int16, reflect.Int32, reflect.Int64:
        return val.Int() == 0
    case reflect.Uint, reflect.Uint8, reflect.Uint16, reflect.Uint32, reflect.Uint64:
        return val.Uint() == 0
    case reflect.Slice, reflect.Array:
        for i := 0; i < val.Len(); i++ {
            if !IsAllZero(val.Index(i).Interface(), eps) { return false }
        }
        return true
    case reflect.Map:
        for _, k := range val.MapKeys() {
            if !IsAllZero(val.MapIndex(k).Interface(), eps) { return false }
        }
        return true
    case reflect.Struct:
        for i := 0; i < val.NumField(); i++ {
            if !IsAllZero(val.Field(i).Interface(), eps) { return false }
        }
        return true
    case reflect.Pointer, reflect.Interface:
        if val.IsNil() { return true }
        return IsAllZero(val.Elem().Interface(), eps)
    default:
        return false
    }
}

C++/Rust limitations and practices

• C++ (to C++23): No runtime reflection. This is done by writing for each custom typeisAllZero, or use template/macro to help expand members.
• Rust: There is also no runtime reflection, often withderiveMacros automatically generate implementations for types (such as custom traitsIsAllZero）。

Practical suggestions

• Please add tolerance when processing floating point numbers.eps, to avoid accuracy issues leading to misjudgments.
• If the structure is large, consider providing explicitisAllZero()method to avoid costly reflection scans.
• Unit tests available in static languages ​​(C++/Rust) ensure that the checking logic for each new structure has covered all members.

message queue

Definition and function

Message Queue (MQ) is an architectural pattern for asynchronous communication between software systems. It allows independent applications or services to exchange information by sending and receiving messages without having to call each other directly or rely on each other's real-time status. The core role of MQ is to act as a middle layer that temporarily stores messages until the intended recipient is ready to process them.

Main components

• Sender (Producer/Publisher):An application that generates and posts messages to a queue or topic. After delivering the message, the sender usually does not wait for a reply and can continue to perform his or her task.
• Receiver (Consumer/Subscriber):An application that fetches messages from a queue or topic and processes them. The receiver can start or stop at any time according to needs without affecting the operation of the sender.
• Message Broker:The core server of the MQ system. It is responsible for receiving, storing, routing and delivering messages to the correct recipient. Common Broker implementations include RabbitMQ, Kafka, ActiveMQ, etc.

Common message patterns

Key Benefits of Message Queuing

• Asynchronous Processing:Allow time-consuming operations to be executed asynchronously in the background, improving front-end response speed and user experience.
• System Decoupling:Eliminate direct dependencies between services. The sender does not need to know the location and status of the receiver, making the system more resilient and maintainable.
• Load Leveling/Peak Shaving:In high-concurrency scenarios, MQ acts as a buffer to smoothly distribute a large number of instant requests to back-end services for processing, preventing the back-end system from crashing due to overload.
• Reliability and Redundancy:Messages are usually stored persistently. Even if the receiver temporarily fails, the message will not be lost and can continue to be processed after the system is restored.

Conclusion

Message queues are the cornerstone of modern distributed systems, microservices architectures, and high-availability applications. It significantly improves the system's resilience, scalability, and robustness by introducing indirection in time and space.

Message Queuing vs. HTTP API

Differences in communication modes

Applicable situation comparison

Advantages and application scenarios of Message Queuing (MQ)

• Traffic peak clipping:Handle sudden large requests, such as order submissions during flash sales or large promotions. MQ acts as a buffer to protect backend services from being overwhelmed by transient traffic.
• Asynchronous processing:Perform long-term or non-immediate operations, such as sending welcome letters after user registration, log processing, generating reports, etc.
• System decoupling:In the microservice architecture, the life cycles and errors of different services are isolated to prevent the failure of service A from affecting service B.
• Event driven architecture:When an event occurs (such as an order status change), multiple downstream systems need to be notified for simultaneous processing.

Advantages and application scenarios of HTTP API

• Instant interaction:Scenarios that require immediate response, such as user login, real-time inventory query, and real-time confirmation of financial transactions.
• Request-response pattern:Scenarios that require two-way interaction and take the next step based on the server's reply.
• Simplicity:For simple CRUD (add, check, modify, delete) operations, the development and testing of HTTP APIs is relatively straightforward and fast.
• Client application:Front-end applications such as web pages or mobile apps usually need to directly send synchronization requests to the back-end and update the interface.

in conclusion

MQ and HTTP API are not replacement technologies, but are designed for different problems.

• If you need immediate results, require synchronization, and the interaction between services is simple and clear, you should choose HTTP API.
• Message queues should be used if you need to handle high concurrency, perform time-consuming operations, maximize system decoupling and fault tolerance.

In modern distributed systems, the two modes often coexist and cooperate with each other to meet different business needs.

Application examples suitable for message queues

1. Asynchronous task processing

In many network applications, some operations are time-consuming, and if users are made to wait simultaneously, it will result in a poor user experience. MQ allows these tasks to be executed asynchronously.

• User registration and notification:When a new user completes registration, the core system only needs to send a "User Registration Successful" message to the queue and immediately respond to the client. Subsequent tasks, such as sending a welcome email, generating initial profile information, or issuing points, are taken out of the queue by the consumer on the back end and executed slowly.
• Long-term calculation and report generation:The user requests to generate a complex report containing large amounts of data or perform large-scale data calculations. The system pushes the request to MQ and immediately informs the user that "the report is being generated and will be notified by email when completed." The background service completes the calculation asynchronously.

2. Traffic peak shaving and buffering

MQ is critical in handling transient high-traffic scenarios and can protect backend services from crashing.

• E-commerce rush sale/flash sale activities:During high-traffic promotions, tens of thousands of order submission requests can overload databases and back-end order processing services. MQ receives and queues all order requests. The backend service pulls requests from the queue and processes them at a steady rate it can handle, smoothing out peak pressures.
• Payment system:Payment requests go to MQ first. Even if the payment gateway is briefly overloaded, the request is not lost but is safely queued waiting to be processed.

3. System decoupling and microservice communication

In complex distributed systems and microservice architectures, MQ is used to isolate services and reduce interdependence.

• Order status change notification:When the "Order Service" successfully creates a new order, it sends a "New Order Created" event message to the MQ topic. Multiple services that are interested in this event (such as "Inventory Service" that needs to deduct inventory, "Logistics Service" that needs to prepare for shipment, and "Points Service" that needs to calculate points) will receive this message and perform corresponding operations independently.
• Data synchronization:When an important data in the main database is updated, a change event can be sent to MQ, allowing other downstream systems (such as search indexing services, data warehouses) to subscribe to and not update their own copies of the data synchronously to achieve eventual consistency.

4. Log collection and monitoring

Collect large amounts of log data from front-end applications or servers into a centralized processing system.

• Log collection:Hundreds of application servers continuously generate log data. These logs are not written directly to the analytics repository, but are first sent to MQ (for example using Apache Kafka). Centralized log processing services (such as Logstash) consume logs from MQ, clean, transform and index them, and finally store them in storage such as Elasticsearch. This isolates the process of log sending and processing.

5. Data streaming processing

In particular, high-throughput MQ/streaming platforms like Apache Kafka are ideal for handling continuous streams of real-time data.

• Internet of Things (IoT) data:Millions of sensors and devices continuously generate temperature, location or status data. MQ can efficiently receive these data streams and pass them to real-time monitoring systems or data analysis pipelines.

HTTP API transfers images and videos

Challenges of transferring large binary data

The main challenges in using HTTP API to transfer large binary data (Binary Data) such as images and videos are:

• Size limit:Individual files are often large and may exceed the Request Body size limit set by the server or network.
• Time limit:Transferring large files takes a long time and can easily cause HTTP request timeout (Timeout).
• Efficiency and Coding:Traditional JSON or XML are not suitable for transmitting binary data and require efficient encoding.

HTTP API method to upload (send) pictures/videos

1. Use multipart/form-data format (most commonly used for file upload)

This is the most standard and common way for a browser or client application to upload files.

• principle:Divide the entire HTTP request body into multiple parts. Each part can contain different field data (such as text, numbers) or a complete file.
• advantage:Highly versatile, most web frameworks and servers natively support processing this format. Multiple files and related metadata can be uploaded in one request.
• Content-Type example: multipart/form-data; boundary=YourBoundaryString

2. Use binary body directly

If you only need to upload a single file, you can directly use the binary content of the file as the subject of the request.

• principle:The body of the request contains only the raw bytes of the file (Raw Bytes).
• advantage:Simple and direct, no additional form boundary (Boundary) overhead, high transmission efficiency.
• Content-Type example:picture:image/jpegorimage/png;video:video/mp4。

3. Use Base64 encoding (not recommended for large files)

Convert binary data into ASCII strings and embed them into text formats such as JSON or XML for transmission.

• advantage:Easy to embed into standard JSON requests, no changes requiredContent-Type, APIs for handling simple text formats are also supported.
• shortcoming:Base64 encoding will increase the file size by approximately 33% and is **extremely unsuitable** for transmitting large videos or high-quality images. It will significantly increase transmission time and server decoding overhead.

HTTP API method to download (receive) pictures/videos

Downloading binary data is relatively simple. The server directly returns the original binary content of the file as the body of the HTTP response (Response Body).

• Server response:The server must set the correct HTTP header:
  • Content-Type：The MIME type of the file must be specified, e.g.image/jpegorvideo/mp4。
  • Content-Disposition：Usually set toattachment; filename="example.mp4", instructs the browser to download the content as a file rather than display it directly.

Optimization of video transmission: chunked transmission and streaming

For particularly large files (especially videos), the following techniques are recommended to improve reliability and efficiency:

• Chunked Upload:The client divides the large file into multiple smaller chunks (Chunks), uploads them respectively through HTTP requests, and is finally assembled by the server. This avoids a single request timeout and facilitates resumption.
• Byte Range Requests:When downloading, the client can use HTTP HeaderRangeRequest a specific fragment of an archive (e.g.Range: bytes=100-200). This is very important for video streaming (Streaming), allowing the player to download only the part that needs to be played, and supports fast forward/rewind.

Streaming key technologies

1. Transport protocols (Protocols)

The core of streaming transmission lies in how to transmit audio and video content from the server to the client efficiently, stably and with low latency.

• HTTP Live Streaming (HLS)：Developed by Apple, it is one of the most popular streaming standards available. It cuts the video into small $\texttt{.ts}$ (MPEG Transport Stream) or $\texttt{.fmp4}$ (Fragmented MP4) file segments, and uses a $\texttt{.m3u8}$ playlist to manage these segments.
  • advantage:Highly compatible (supported by almost all devices and browsers) and easily accelerated via standard HTTP cache and CDN.
• MPEG-DASH (Dynamic Adaptive Streaming over HTTP)：An international standard that functions similarly to HLS, also uses HTTP transport and fragmentation technology, and uses $\texttt{.mpd}$ files as media descriptions.
  • advantage:Highly flexible and functional, it's a popular choice for Android devices and many non-Apple platforms.
• WebRTC (Web Real-Time Communication)：Designed specifically for **ultra-low latency** instant messaging, such as video conferencing, live interaction, and online games. It usually uses UDP for data transfer.
  • advantage:Extremely low latency (milliseconds), peer-to-peer (P2P) or near-peer-to-peer transmission.
• RTMP (Real-Time Messaging Protocol)：Traditional Adobe Flash protocol. Although it has been gradually replaced by HLS/DASH, it is still widely used for live streaming (pushing from the encoder to the streaming server).

2. Core technologies and concepts

Adaptive Bitrate Streaming (ABR)

This is the cornerstone of modern streaming services. The server will encode the same piece of video content into multiple versions of different qualities (bit rate, resolution).

• principle:The client (player) will **dynamically** switch to the highest quality version that can play smoothly based on its own network bandwidth and CPU load.
• benefit:Provide the best user experience and avoid buffering and lag.

Content Delivery Network (CDN)

CDN is indispensable for streaming services targeting global users.

• principle:Distribute and cache audio and video content (HLS/DASH clips) to edge node servers around the world.
• benefit:Users obtain content from the geographically closest server, significantly reducing latency, increasing transfer speeds, and offloading the load on the original server.

Audio and video codecs (Codecs)

Used to compress and decompress audio and video data to reduce file size.

• video:H.264 (AVC), H.265 (HEVC), AV1. H.265 and AV1 offer higher compression efficiency than H.264, delivering the same image quality in a smaller file size.
• News:AAC (Advanced Audio Coding) is the most commonly used audio codec for streaming.

3. Content protection and monetization

• Digital Rights Management (DRM):Used to protect copyrighted content from illegal copying and dissemination, ensuring that only authorized users can play it. Common standards include Widevine (Google), PlayReady (Microsoft), and FairPlay Streaming (Apple).
• Server-Side Ad Insertion (SSAI):Before the streaming content reaches the client, the ad fragments are spliced ​​with the main content on the server side, effectively bypassing the client's Ad Blocker.

Shell

Bash - $?

In Bash, $? represents the exit status code (Exit Status) of the last executed command. This is an integer value usually used to determine whether the previous command was executed successfully.

The meaning of exit status code

• 0: Indicates that the command was successfully executed.
• Non-0: Indicates that the command execution failed. Different non-0 values ​​can represent different error types.

Example 1: Checking successfully executed commands

#!/bin/bash
ls
echo "The exit status code of the previous command was: $?"

In this example,lsThe command will list the directory contents. After successful execution,$?The value will be 0.

Example 2: Check for failed commands

#!/bin/bash
ls /nonexistent-directory
echo "The exit status code of the previous command was: $?"

In this example,lsAttempting to list a directory that does not exist will cause the command to fail.$?The value of will be a non-zero number.

Example 3: Using exit status code for conditional judgment

#!/bin/bash
cp file1.txt /some/directory/
if [ $? -eq 0 ]; then
    echo "File copied successfully."
else
    echo "File copy failed."
fi

This example determines which message to display based on the exit status of the command.

Bash if

basic grammar

if [condition]; then
    instructions
fi

Use AND

To satisfy multiple conditions simultaneously in if , use:

1. Use-a(outdated but available)

if [ "$a" -gt 0 -a "$b" -gt 0 ]; then
    echo "a and b are both greater than 0"
fi

2. Use[[ ]] &&(recommend)

if [[ "$a" -gt 0 && "$b" -gt 0 ]]; then
    echo "a and b are both greater than 0"
fi

3. Use multiple if combinations&&

if [ "$a" -gt 0 ] && [ "$b" -gt 0 ]; then
    echo "a and b are both greater than 0"
fi

Use OR

It can be executed as long as one of the conditions is true:

1. Use-o(outdated but available)

if [ "$a" -eq 0 -o "$b" -eq 0 ]; then
    echo "a or b is 0"
fi

2. Use[[ ]] ||(recommend)

if [[ "$a" -eq 0 || "$b" -eq 0 ]]; then
    echo "a or b is 0"
fi

3. Use multiple if combinations||

if [ "$a" -eq 0 ] || [ "$b" -eq 0 ]; then
    echo "a or b is 0"
fi

Mix AND and OR

Can be used with parentheses to control precedence:

if [[ ( "$a" -gt 0 && "$b" -gt 0 ) || "$c" -eq 1 ]]; then
    echo "a and b are both greater than 0, or c is equal to 1"
fi

Common usage of NOT

In Bash,ifInstructions are usually paired with[ ](test command) or[[ ]](Extended Test Command) is used. To express NOT, use an exclamation point!。

if [ ! condition ]; then
    # Code to be executed when the condition is false
fi

1. Check whether the file or directory exists

This is the most common usage, for example: create a file if it "doesn't exist".

# If no file named config.txt exists
if [ ! -f "config.txt" ]; then
    echo "File does not exist, creating..."
    touch config.txt
fi

# If the directory does not exist
if [ ! -d "/var/log/myapp" ]; then
    mkdir -p "/var/log/myapp"
fi

2. String comparison

Check whether the string is "not equal to" or "not empty".

STR="hello"

# Check if the variable is not equal to "world"
if [ "$STR" != "world" ]; then
    echo "String does not match"
fi

# Check if the string is not empty (! -z is equivalent to -n)
if [ ! -z "$STR" ]; then
    echo "There is data in the variable"
fi

3. Numerical comparison

Although the numerical value is relatively-ne(not equal), but can also be matched!use.

NUM=10

if [ ! "$NUM" -eq 20 ]; then
    echo "The number is not equal to 20"
fi

Conditional logic comparison table

Advantages of using double brackets [[ ]]

In more modern Bash scripts, it is recommended to use[[ ]], which is processing!More powerful and less error-prone when combined with logic.

# Combine multiple conditions: if it is not a directory and cannot be read
if [[ ! -d $path && ! -r $path ]]; then
    echo "Invalid path or insufficient permissions"
fi

Things to note

• use[[ ]]timely support&&and||
• use[ ]It is recommended to use multiple conditional statements together&& / ||
• Variables must be enclosed in quotation marks to avoid errors caused by null values.

Bash conditional branching

basic grammatical structure

caseStatement is used to compare the value of a variable with multiple patterns (patterns), and its function is similar to that in other languages.switch. Its grammatical structure is as follows:

case variable in
    Mode 1)
        # Execute instructions
        ;;
    Mode 2)
        # Execute instructions
        ;;
    *)
        # Default execution command (similar to default)
        ;;
esac

Common usage examples

1. Simple menu judgment

This is the most common usage, performing different actions based on user input.

read -p "Please enter (yes/no): " input

case "$input" in
    "yes")
        echo "You selected Yes"
        ;;
    "no")
        echo "Did you choose No"
        ;;
    *)
        echo "Input error"
        ;;
esac

2. Combine multiple patterns (OR)

use|Symbols allow multiple modes to execute the same block of instructions.

read -p "Please enter the month abbreviation: " month

case "$month" in
    Jan|Feb|Mar)
        echo "Season 1"
        ;;
    Apr|May|Jun)
        echo "Season 2"
        ;;
    *)
        echo "other months"
        ;;
esac

3. Use wildcards

Pattern matching supports extended symbols like file paths (e.g.*, ?, [a-z]）。

read -p "Please enter a character: " char

case "$char" in
    [a-z])
        echo "This is lowercase letter"
        ;;
    [A-Z])
        echo "This is a capital letter"
        ;;
    [0-9])
        echo "This is a number"
        ;;
    *)
        echo "This is a special symbol or multiple characters"
        ;;
esac

Explanation of key symbols

Precautions for use

• Variable reference:It is recommended to put variables in double quotes (e.g."$var") to avoid syntax errors when the variable has a null value.
• Order sensitive: casewill be compared from top to bottom. Once the match is successful and encounter;;will stop, so please put the specific mode in the general mode (such as*)Before.

Case sensitivity

Default behavior: case sensitive. In a standard Bash environment,caseThe statement isCase sensitiveof. This means "A" and "a" are treated as different patterns.

read -p "Enter a character: " char
case "$char" in
    a) echo "This is lowercase a" ;;
    A) echo "This is capital A" ;;
esac

Method 1: Use the nocasematch option

If you want the entirecaseThe statement ignores case and can be usedshoptcommand enablednocasematchset up. This will affect all subsequent string comparisons andcasejudge.

# Enable ignoring case
shopt -s nocasematch

read -p "Enter yes or YES: " input
case "$input" in
    yes) echo "Match successful (ignore case)" ;;
esac

# Turn off ignoring case (restore default)
shopt -u nocasematch

Method 2: Manually specify the range in the pattern

If you don't want to change the global setting, defining the case range directly in the schema is the most standard and consistent approach.

read -p "Enter a letter: " char
case "$char" in
    [aA])
        echo "You entered a or A"
        ;;
    [b-zB-Z])
        echo "You entered additional letters"
        ;;
esac

Method 3: Combine pipe symbols (OR)

For specific words, you can use|to list all possible spellings.

case "$input" in
    stop|STOP|Stop)
        echo "stop execution"
        ;;
esac

Comparison of setting methods

Bash operation precedence order

Use parentheses inside (( ))

Arithmetic extensions in Bash(( ))Internally, you can just use parentheses( )to explicitly specify the precedence of operations, consistent with the logic of C or most programming languages. This is very important when dealing with complex logic (combining AND, OR, addition, subtraction, multiplication and division).

# Example: (A and B) or C
if (( (MIN > 91 && MIN< 110) || FORCE_MODE == 1 )); then
    echo "條件成立"
fi

precedence rules

exist(( ))Within, the order of precedence of operations follows standard mathematical and logical conventions:

1. Items in brackets: ( )Highest priority.
2. Multiplication and division: *, /, %。
3. Addition and subtraction: +, -。
4. Comparison operations: <, >, <=, >=。
5. Logical operations: &&(AND) takes precedence over|| (OR)。

Complex condition example

Suppose you need to determine: the value must be within the range, and (must be a specific multiple or in forced mode):

if (( MIN > 91 && (MIN % 5 == 0 || FORCE_MODE == 1) )); then
    {
        echo "Trigger complex conditional judgment"
        IS_SHUTDOWN=true
    } |& tee -a "$LOGFILE"
fi

Comparisons and recommendations

Bash exit usage

exit command and status code

In a Bash script,exitUsed to terminate script execution and return a status code (Exit Status) to the parent program.

• exit 0：Represents "success" or "normal end".
• exit 1 (or non-zero number):Represents "failure" or "an error occurred". Different numbers usually represent different types of errors (range 0-255).

if [ -f "config.txt" ]; then
    echo "File exists"
    exit 0
else
    echo "Error: Profile not found"
    exit 1
fi

Get exit status code

After executing a script or instruction, you can use special variables$?to get the previous programexitstatus code.

./script.sh
echo "Script end status: $?"

Use backtick `` or $() to get the output value

when you call.shfile and want to "output the contents" of it (i.e.echoWhen storing the contents of a variable, you can use backticks or$(). This is called Command Substitution.

Sample script (get_name.sh)

#!/bin/bash
# This is the output content and will be captured
echo "Ubuntu_User"
# This is the end state and will not be captured by backticks
exit 0

Call from another script or terminal

# Use backticks
RESULT=`./get_name.sh`

# Use $() (recommended modern writing)
RESULT=$(./get_name.sh)

echo "The result obtained is: $RESULT"

Important difference: return value vs output content

If you want to get both at the same time, you can refer to the following example:

OUTPUT=$(./script.sh)
STATUS=$?

echo "Content is: $OUTPUT"
echo "The status code is: $STATUS"

Bash return usage

Basic definition of return

In Bash,returninstructions are specifically used toFunctionor throughsourceThe script to execute. It stops execution of the current function or script and returns the specified status code to the caller, but does not close the current shell program.

1. Use return within a function

This is the most common usage.returnWhat is returned is the status code (0-255), not the string data.

Sample script (func_test.sh)

#!/bin/bash

# define function
check_file() {
    if [ -f "$1" ]; then
        return 0 # success
    else
        return 1 # failed
    fi
}

# Call function
check_file "test.txt"
RESULT=$?

if [ $RESULT -eq 0 ]; then
    echo "File exists"
else
    echo "File does not exist"
fi

2. Use return (with source) in the .sh file

If you want the variables and environment to remain in the current terminal after a script is executed, you usually usesource. Must be used at this timereturnrather thanexit。

callee (sub_script.sh)

# Simulate logical judgment
if [ "$USER" != "root" ]; then
    echo "Insufficient permissions, only for root use"
    # If you use exit, you will directly close the terminal. If you use return, you will only stop the execution of this script.
    return 1
fi

export APP_STATUS="Ready"
return 0

caller (directly in CMD or another script)

source ./sub_script.sh
echo "Script returns status: $?"
echo "Get environment variables: $APP_STATUS"

3. How to obtain the "return value" through the return mechanism

becausereturnOnly numbers (status codes) can be returned. To obtain strings or large amounts of data, it is recommended to combineechoandcommand substitution。

callee (get_data.sh)

#!/bin/bash
calculate() {
    local val=$(( $1 + $2 ))
    # Output results to stdout
    echo "$val"
    # Return execution status code
    return 0
}

calculate 10 20

caller

# Get the output content
DATA=$(./get_data.sh)
# Get the status code of return
STATUS=$?

echo "The calculation result is: $DATA"
echo "Execution status: $STATUS"

Usage summary and comparison

Bash check if variable is empty

In Bash, you can use conditional judgment to check whether a variable is empty. Here are several common ways:

Example 1: Using-zCheck if variable is empty

#!/bin/bash

var=""
if [ -z "$var" ]; then
    echo "The variable is empty"
else
    echo "The variable is not empty"
fi

-zUsed to check whether the variable is empty. If the variable is empty, the condition is true.

Example 2: Use-nCheck if the variable is not empty

#!/bin/bash

var="some value"
if [ -n "$var" ]; then
    echo "The variable is not empty"
else
    echo "The variable is empty"
fi

-nUsed to check whether the variable is not empty. If the variable has a value, the condition is true.

Example 3: Using double quote comparison to check if a variable is empty

#!/bin/bash

var=""
if [ "$var" == "" ]; then
    echo "The variable is empty"
else
    echo "The variable is not empty"
fi

This method directly compares the variable with the empty string to check whether the variable is empty.

Array structure in Bash

Introduction

Bash supports two types of arrays:

• Indexed Array:Access with numerical index
• Associative Array:Access with custom string keys (requires Bash 4.0 or above)

index array

Defined way

fruits=("apple" "banana" "cherry")

Read elements

echo "${fruits[0]}"     # apple
echo "${fruits[1]}"     # banana

list all elements

echo "${fruits[@]}"

Get array length

echo "${#fruits[@]}"

Add new element

fruits+=("date")

Traverse elements

for fruit in "${fruits[@]}"; do
    echo "$fruit"
done

associative array

Definition and usage (requires Bash 4+)

declare -A capital
capital["Taiwan"]="Taipei"
capital["Japan"]="Tokyo"

Access and traverse

echo "${capital["Japan"]}" # Tokyo

for key in "${!capital[@]}"; do
    echo "The capital of $key is ${capital[$key]}"
done

Things to note

• Index array index starts from 0
• It is recommended that element values ​​be enclosed in double quotes to avoid whitespace character errors.
• Associative arrays must be useddeclare -Adeclare

in conclusion

Bash array is an important structure for storing multiple pieces of data, and is suitable for data processing such as parameter lists, directory sets, and key-value pairs.

Bash merge two arrays

basic grammar

combined=("${array1[@]}" "${array2[@]}")

This will merge all elements from both arrays into a new arraycombined。

Complete example

array1=("apple" "banana")
array2=("cherry" "date")

combined=("${array1[@]}" "${array2[@]}")

echo "Combined array:"
for item in "${combined[@]}"; do
    echo "$item"
done

Merge into original array

array1+=("${array2[@]}")

This willarray2content is added directly toarray1rear.

Things to note

• Please be sure to use"${array[@]}"form to prevent elements from being expanded incorrectly
• If there are blanks or special characters in the array, it is necessary to wrap them in double quotes.

Bash determines whether a string begins with a certain word

Use string pattern comparison (recommended)

str="hello world"

if [[ "$str" == hello* ]]; then
    echo "String starts with hello"
fi

illustrate:use[[ ]]Support shell pattern comparison,*Represents any character.

Use regular expressions

if [[ "$str" =~ ^hello ]]; then
    echo "String starts with hello"
fi

illustrate: ^Indicates the beginning,[[ ]]in=~is a regular operation.

usecaseNarrate

case "$str" in
  hello*) echo "String starting with hello" ;;
  *) echo "does not start with hello" ;;
esac

illustrate: caseIt is a good tool for processing string patterns, concise and highly readable.

useexpr(Old style of writing)

if expr "$str" : '^hello' &> /dev/null; then
    echo "String starts with hello"
fi

illustrate:useexprRegular matching function, suitable for older versions of the shell.

Things to note

• It is best to put quotes around variables to prevent empty string errors.
• [[ ]]Compare[ ]More complete functions, recommended to use
• use=~Regular expressions can be used, but remember not to add quotes

Bash determines whether a file or directory exists

Determine whether it exists:-e

-eUsed to check whether a file or directory exists (regardless of type).

FILE="/etc/passwd"

if [ -e "$FILE" ]; then
    echo "$FILE exists"
else
    echo "$FILE does not exist"
fi

Determine whether the directory does not exist:! -d

-dUsed to check whether it is a directory,!It is a negative operation.

DIR="/tmp/myfolder"

if [ ! -d "$DIR" ]; then
    echo "$DIR does not exist, creating..."
    mkdir -p "$DIR"
else
    echo "$DIR already exists"
fi

Other commonly used condition options

• -e: Whether the file or directory exists (no matter what type)
• -f: Whether it is a "general file"
• -d: Whether it is a "directory"
• -L: Whether it is a "symbolic link"

Example: Combining inspection and creation

PATH_TO_CHECK="/home/user/config"

if [ -e "$PATH_TO_CHECK" ]; then
    echo "$PATH_TO_CHECK already exists"
else
    echo "$PATH_TO_CHECK does not exist and is being created..."
    mkdir -p "$PATH_TO_CHECK"
fi

suggestion

• If you only need to know "whether there exists or not" → use-e
• If you need to specifically confirm that it is a directory or file → use-dor-f
• Variables must be enclosed in double quotes to avoid null or space errors.

List all subdirectories and store them in an array

Example: Store to Array

target_dir="/path/to/your/dir"
subdirs=()

while IFS= read -r -d $'\0' dir; do
    subdirs+=("$dir")
done < <(find "$target_dir" -mindepth 1 -maxdepth 1 -type d -print0)

illustrate

• find: List subdirectories (no recursion)
• -mindepth 1:Exclude own directory
• -maxdepth 1:Only list the first level
• -type d:Only list directories
• -print0matchread -d $'\0': Handle paths with spaces

Example output

for dir in "${subdirs[@]}"; do
    echo "$dir"
done

Alternative simple writing (if there are no special characters)

subdirs=( "$target_dir"/*/ )

This way of writing uses wildcards to match subdirectories, but it cannot exclude files or handle spaces and special characters.

Get the actual path containing ~

Method 1: Use eval to expand

path="~/myfolder/file.txt"
realpath "$(eval echo "$path")"

Method 2: Replace with $HOME

path="~/myfolder/file.txt"
realpath "$(echo "$path" | sed "s|^~|$HOME|")"

Method 3: Use readlink

path="~/myfolder/file.txt"
readlink -f "$(eval echo "$path")"

Recommended usage

realpath "$(eval echo "$path")"

grep

basic search

grep "keyword" file name

Ignore case

grep -i "keyword" file name

Show line number

grep -n "keyword" file name

recursive search

grep -r "keyword" directory path

Only show matching text

grep -o "keyword" file name

Also display file name

grep -H "keyword" file name

Two sets of keywords appear together (AND)

grep "keyword 1" file name | grep "keyword 2"

Any one of the two keywords appears (OR)

#Method 1: Regular expression
grep -E "Keyword 1 | Keyword 2" file name

#Method 2: Multiple -e parameters
grep -e "keyword 1" -e "keyword 2" file name

Solve binary file matches problem

#Method 1: Force the file to be treated as text
grep -a "keyword" file name

#Method 2: Convert the file encoding to UTF-8 and then search
iconv -f original encoding -t UTF-8 file name | grep "keyword"

# Example (convert from BIG5 to UTF-8)
iconv -f BIG5 -t UTF-8 file name | grep "keyword"

Common combinations

grep -rin "keyword" directory path

-print0 with read -d $'\0'

Instructions for use

When processing file names or paths that contain spaces or special symbols (such as blanks, quotation marks, newlines), the traditionalfindDistribution linexargsorreadMisjudgment may occur.

-print0Allowablefindwith null(\0) characters as output delimiters, whileread -d $'\0'Can accurately read null-delimited content to ensure accurate separation of each file/path.

Basic example: collect all subdirectories

subdirs=()
while IFS= read -r -d $'\0' dir; do
    subdirs+=("$dir")
done < <(find . -type d -print0)

illustrate

• -print0: End each result with a null character without line breaks.
• -d $'\0'：readRead one null-terminated item at a time
• IFS=: Prevent spaces from being treated as field separators
• -r: Avoid skipping characters during processing

Example 1: List the absolute paths of all .txt files

txt_files=()
while IFS= read -r -d $'\0' file; do
    txt_files+=("$file")
done < <(find "$(pwd)" -type f -name "*.txt" -print0)

Example 2: Send files to a command for processing (security)

find . -type f -print0 | while IFS= read -r -d $'\0' file; do
    echo "Processing: $file"
    # your_command "$file"
done

Example 3: Process only directories containing spaces

find . -type d -print0 | while IFS= read -r -d $'\0' dir; do
    if [[ "$dir" == *" "* ]]; then
        echo "Directory containing spaces: $dir"
    fi
done

Why not use line breaks?

Traditional usage is as follows:

find . -type f | while read file; do ...

However, if the file name contains a newline, it will causereadWrong parsing, even multiple lines are misjudged as multiple files.

in conclusion

• -print0andread -d $'\0'The safest way to handle any file name
• Avoid processing errors caused by spaces, special characters, or newlines
• Suitable for Bash array collection, loop processing, xargs matching-0Waiting scene

While read only obtains one directory problem

Problem description

The following writing method is used in some Bash or Cygwin environments:Read only the first item：

while IFS= read -r -d $'\0' dir; do
    subdirs+=("$dir")
done < <(find "$target_dir" -mindepth 1 -maxdepth 1 -type d -print0)

This is usually because `<(find ...)` 在某些系統（如 Cygwin）不是「真正的檔案描述符」，造成 `read` 無法持續讀取。

Correction: Use pipeline with `read -d ''` (or `readarray -d`)

Method 1: Use `find | while` pipeline combination-print0

subdirs=()
find "$target_dir" -mindepth 1 -maxdepth 1 -type d -print0 | while IFS= read -r -d $'\0' dir; do
    subdirs+=("$dir")
done

**Note**: If you want `subdirs` to be available in the main program, this method is not suitable (because `while ... |` subshell cannot return the array)

Method 2: Change to array and process again

mapfile -d '' -t subdirs < <(find "$target_dir" -mindepth 1 -maxdepth 1 -type d -print0)

mapfile(orreadarray) can correctly read null-separated strings into arrays.
This is the most reliable way to keep it in the main shell.

Example output

for dir in "${subdirs[@]}"; do
    echo "Found directory: $dir"
done

Summarize

• mapfile -d '' -tis processing-print0The safest and most Bash-native way
• Avoid using pipelines which would result in array scope being limited to subshells
• Note that some older versions of Bash or non-Linux environments may not support `<( ... )` 正常工作

Read data from variable contents (use read)

Assume that there are multiple lines of text in the variable

var="line1
line2
line3"

Use while read to read line by line

while IFS= read -r line; do
    echo "Read: $line"
done <<< "$var"

Commentary

• IFS=Avoid trim whitespace
• -rAvoid escape characters from being parsed
• <<<is a here-string, treats the variable content as input to while

Single line read

read -r first_line <<< "$var"
echo "First line: $first_line"

read as array

readarray -t lines <<< "$var"
for line in "${lines[@]}"; do
    echo "$line"
done

Find the directory starting from the string and sort by date

Complete instructions (sorted by directory modification time)

find . -mindepth 1 -maxdepth 1 -type d -name "abc*" -printf "%T@ %p\n" | sort -nr | cut -d' ' -f2-

illustrate

• -mindepth 1 -maxdepth 1: Only list subdirectories under the current directory
• -name "abc*": Name starts with abc
• -printf "%T@ %p\n": Print out the modification time (timestamp) and directory name
• sort -nr: Sort by time value from new to old
• cut -d' ' -f2-: Remove the timestamp and only display the path

Example output

./abc_latest
./abc_old
./abc_2020

If you need to save it to a Bash array

readarray -t abc_dirs << <(
  find . -mindepth 1 -maxdepth 1 -type d -name "abc*" -printf "%T@ %p\n" |
  sort -nr | cut -d' ' -f2-
)

for dir in "${abc_dirs[@]}"; do
  echo "Found: $dir"
done

Note

• -printfExclusive function for GNU find, default in macOSfindNot supported, availablegfindsubstitute
• %T@Represents "last modification time (seconds)". If you need to create a time, additional tools are required.

Bash checks if storage device is writable

Method 1: Check if the directory has write permissions

DIR="/mnt/usb"

if [ -w "$DIR" ]; then
    echo "$DIR is writable"
else
    echo "$DIR is not writable"
fi

-wTo check whether the directory has "write permission". However, if the directory itself is writable but the actual device is read-only (for example, mount is readonly), this method may not work.

Method 2: Actually try to write to the test file

DIR="/mnt/usb"
TESTFILE="$DIR/.write_test"

if touch "$TESTFILE" 2>/dev/null; then
    echo "$DIR is writable"
    rm "$TESTFILE"
else
    echo "$DIR is not writable"
fi

This method is the most reliable and can detect the mount status or whether the physical device is actually writable.

Method 3: UsemountInstruction to check whether the mount is read-only

MNT="/mnt/usb"

if mount | grep "$MNT" | grep -q "(ro,"; then
    echo "$MNT is mounted read-only"
else
    echo "$MNT is a writable mount"
fi

This method needs to check whether the device is read-only (ro) method.

suggestion

• If insurance is required, use the "Method 2" actual writing test method.
• Can be used with udev or mount to automatically check.

Shell redirection usage instructions

1. Standard output redirection: `>`

In the shell, use `>` to redirect a command's standard output (stdout) to a file or device. If the file already exists, the contents will be overwritten.

echo "Hello" > output.txt

This line of instruction will"Hello"writeoutput.txtfile.

2. Add output redirection: `>>`

Using `>>` will append standard output to the end of the specified file without overwriting the original content.

echo "Hello again" >> output.txt

This line of command will"Hello again"append tooutput.txtthe end of.

3. Standard error redirection: `2>`

In the shell, `2>` is used to redirect standard error (stderr) to a specified location. For example:

ls non_existent_file 2> error.log

This line of command will write the error message toerror.logfile.

4. Add error redirection: `2>>`

If you do not want to overwrite the error message file, you can use `2>>` to append the error message to the end of the file.

ls non_existent_file 2>> error.log

This line of command will append the error message toerror.log。

5. Redirect standard output and error at the same time: `&>`

Use `&>` to redirect both standard output and standard error to the same file or device.

command &> all_output.log

This line of command willcommandAll output (standard output and error) is written toall_output.log。

6. Merge standard error to standard output: `2>&1`

`2>&1` merges standard error into standard output to facilitate unified management. For example:

command > output.log 2>&1

This line of instruction will write to both standard output and erroroutput.log。

7. Disable all output: `>/dev/null 2>&1`

If you don't want any output to be displayed, you can direct all output to/dev/null,like:

command >/dev/null 2>&1

This line of command willcommandAll output is discarded.

Convert and output UTF-8 encoded files

in useteeWhen the command appends output to a file, you can passiconvConvert the output to UTF-8 encoding, ensuring the file contents are saved in UTF-8. Below are specific instructions and examples.

Command format

The following saves the output as UTF-8 encodedteeInstruction format:

• command: The command you want to execute.
• iconv -t utf-8:Converts the output to UTF-8 encoding, where-tRepresents "target encoding".
• tee -a output.txt: Append the converted output tooutput.txtdocument.

example

The following example demonstrates how tolsThe output of the command is written in UTF-8 encoding tooutput.txt：

After executing this instruction,output.txtThe content will be saved in UTF-8 encoding to avoid encoding errors.

Windows cmd

Opening method

• Search for "cmd" or "command prompt" in the Start menu
• useWin + Rentercmdand press Enter
• Enter in the folder path columncmdCan be opened directly from this path

Common commands

• dir: Display files and folders in the current directory
• cd: Switch directories, for examplecd C:\Users
• cls:Clear screen content
• copy: Copy files, for examplecopy a.txt b.txt
• del: Delete files
• mkdir:Create a new folder
• rmdir: Delete folder
• echo: Output text, for exampleecho Hello
• exit: Close command prompt characters

Advanced operation

• Redirect output:dir > file.txtOutput results to file
• Pipeline operations:dir | find "txt"filter containstxtthe result
• Batch file: create.batFile executes multiple commands at once
• Environment variables: usesetView and set variables

Administrator mode

• Search for "cmd", right-click and select "Run as administrator"
• Can be used likeipconfig /flushdnsorsfc /scannowWaiting for system maintenance instructions

Execute the second command after cmd /k

Execute multiple commands in the same window

Can be used&、&&or||To continue the instructions:

cmd /k "First command & Second command"

• &: Regardless of whether the first one succeeds or fails, the next one will be executed.
• &&: Only execute the next one if the first one succeeds
• ||: Only execute the next one if the first one fails

Call other batch files

After the first instruction addcall, you can continue to execute another batch file:

cmd /k "First command & call second.bat"

Use /c to close after executing

If you don't need to keep the window open, you can use/c：

cmd /c "First command & Second command"

/cwill automatically close the window after executing all commands, and/kThe window will be retained.

CMD variable

What are CMD variables?

In the Windows Command Prompt (CMD), variables are namespaces used to store data or settings. They can be stored in the environment (environment variables) or defined and used in the batch file (regional variables).

Variable type

• Environment Variables:

  Variables set by the system or user are available throughout the operating system session. For example:%PATH%、%TEMP%。

  View environment variables:usesetOrder.

  set

  Set/modify environment variables (current CMD session only):

  set MY_VAR=Hello

• Batch Local Variables:

  In the batch file (.bator.cmdfile)setThe variables defined by the command are only valid during the execution of this batch file.

• For Loop Variables:

  in useforVariables defined when looping, used to iterate over files, folders, or numeric sequences. Usually named with a single letter preceded by two percent signs, for example:%%A(in a batch file) or a single percent sign, for example:%A(When executed directly from the command line).

• Parameter Variables:

  Parameters passed to the batch file when executing it. For example:%1is the first parameter,%2is the second parameter,%0Is the path and name of the batch file itself.

How to use variables

Read variable value

To get the value of a variable, you usually need to add percent signs before and after the variable name.%。

set NAME=CMD User
echo My name is %NAME%

Output:My name is CMD User

Set regional variables in batch files

@echo off
set MESSAGE=This is a batch message
echo%MESSAGE%
pause

Set a variable, but only if it does not exist

You can useIF DEFINEDorIF NOT DEFINEDstatement to check whether the variable already exists or is defined (not empty).

@echo off

rem checks whether the variable MY_DEFAULT_VALUE is defined (not empty)
IF NOT DEFINED MY_DEFAULT_VALUE (
    set MY_DEFAULT_VALUE=Default_Setting
    The echo variable MY_DEFAULT_VALUE is undefined and has been set to the default value.
) ELSE (
    echo variable MY_DEFAULT_VALUE already exists, the value is: %MY_DEFAULT_VALUE%
)

rem is executed once, the variable will be set to Default_Setting
If rem is executed again, it will show that it already exists.

set MY_DEFAULT_VALUE=Existing_Setting

rem check again
IF NOT DEFINED MY_DEFAULT_VALUE (
    set MY_DEFAULT_VALUE=Another_Default
    echo variable MY_DEFAULT_VALUE is undefined, set to Another_Default.
) ELSE (
    echo variable MY_DEFAULT_VALUE already exists, the value is: %MY_DEFAULT_VALUE%
)

pause

Another neat trick is to take advantage of variable augmentation's default value functionality (but this is mainly used for loops and parameter processing, and in standardsetThe logic of "set if not set" is not directly applicable to the command. therefore,IF NOT DEFINEDThis is the most standard approach. )

Enable delayed environment variable expansion (Delayed Expansion)

In loops or conditional statements, standard percent sign variable expansion occurs once during statement parsing. In order to be able to read the new value of the variable on each iteration of the loop, you need to enable lazy expansion and use the exclamation mark!to reference variables.

@echo off
setlocal enabledelayeexpansion
setCOUNTER=0
:LOOP
    if %COUNTER% LSS 5 (
        set /A COUNTER+=1
        echo current count: !COUNTER!
        goto LOOP
    )
endlocal

Special variables/system variables

Advanced usage of parameter variables

For parameter variables (such as%1) or a loop variable (such as%%A), you can use modifiers to extract different parts of the path:

echo full path: %~1
echo drive: %~d1
echo path: %~p1
echo file name: %~n1
echo file extension: %~x1

This is useful when working with archive paths.

Filter keywords in CMD

Use the find command

• Not available in CMDgrep, but you can use the built-infindcommand achieves a similar effect.
• grammar:Command | find "keyword"

example

dir | find "txt" :: Only show files containing "txt"
ipconfig | find "IPv4" :: Only show lines containing IPv4
tasklist | find "chrome" :: Filter out executable programs containing chrome

Common parameters

• /I: Ignore case, for examplefind /I "error"
• /V: Display lines that do not contain keywords
• /C: Only display the number of matching rows

Multiple keyword search

• Can connect multiplefindMeet the AND condition:

type log.txt | find "Error" | find "2025"

• To achieve the OR condition, you can usefindstr：

type log.txt | findstr /I "error warning fail"

Advanced: use findstr

• findstrIt is a CMD version of the advanced search tool that supports multiple keywords and regular expressions.
• example:

dir | findstr /R ".*\.txt$" :: Use regular expressions to find .txt files
type log.txt | findstr /I "timeout error fail"

CMD automatically executes the setting method when starting

Concept Note

• In Linux, it will be automatically executed when logging into a new bash~/.bash_profileor~/.bashrc。
• There is no exact equivalent of ".bash_profile" in Windows CMD, but similar effects can be achieved in several ways.

Method 1: Use AutoRun to log in values

• When CMD starts, it will check the system login table forAutoRunSettings will be executed automatically if there is any content.

HKEY_CURRENT_USER\Software\Microsoft\Command Processor
HKEY_LOCAL_MACHINE\Software\Microsoft\Command Processor

• Create or modify string valuesAutoRun, the content can be set to the batch file path to be executed:

AutoRun = "C:\Users\YourName\cmd_startup.bat"

• This batch file will be executed automatically every time CMD is opened.

Method 2: Modify CMD shortcut

• Right-click the CMD shortcut → Properties → Modify the "Target" field:

%SystemRoot%\System32\cmd.exe /k "C:\Users\YourName\cmd_startup.bat"

• /kExecute the batch file after startup and keep the window open.

Method 3: Use login to start the batch file

• Put the startup batch file into:

%AppData%\Microsoft\Windows\Start Menu\Programs\Startup

• This will be executed automatically when logging in to a Windows account, but this is an automatic startup at the system level, not every time CMD is opened.

Correspondence instructions

Automatically set the batch file of CMD AutoRun

Function description

• This batch file will check the login pathHKEY_CURRENT_USER\Software\Microsoft\Command Processorexists.
• If there is no such pathAutoRunvalue, it will be automatically created and set as the specified batch file path.

Sample batch file content

@echo off
setlocal

:: Set the batch file path to be executed by AutoRun
set "AUTORUN_PATH=C:\Users\%USERNAME%\cmd_startup.bat"
set "REG_PATH=HKCU\Software\Microsoft\Command Processor"

echo Check CMD AutoRun setting...

:: Check whether the login code exists
reg query "%REG_PATH%" >nul 2>&1
if errorlevel 1 (
    echo [Information] The Command Processor code cannot be found, creating...
    reg add "%REG_PATH%" /f >nul
)

:: Check if AutoRun is set
reg query "%REG_PATH%" /v AutoRun >nul 2>&1
if errorlevel 1 (
    echo [Information] AutoRun not found, creating new value...
    reg add "%REG_PATH%" /v AutoRun /d "%AUTORUN_PATH%" /f >nul
    echo [Done] AutoRun has been set to: %AUTORUN_PATH%
) else (
    echo [OK] AutoRun already exists and has not been changed.
)

endlocal
pause

illustrate

• reg query: Used to check whether the login key or value exists.
• errorlevel: Can be used to determine whether the query result is successful (0 means exists, 1 means does not exist).
• reg add: Create a login key or value.
• /f: No prompt for forced coverage.
• can be%AUTORUN_PATH%Change it to the batch file path that you want to automatically execute when CMD starts.

Determine batch file running terminal

Sample batch file

@echo off

:: Detect PowerShell
if defined PSModulePath (
    echo is currently executed in PowerShell
    goto :eof
)

:: Detecting Cygwin
if defined CYGWIN (
    echo is currently executed in Cygwin
    goto :eof
)
if defined SHELL (
    echo is currently executed in Cygwin (SHELL=%SHELL%)
    goto :eof
)

::Default is CMD
echo is currently executed in CMD

Judgment basis

• PowerShell: Environment variables when the batch file is executed in PowerShellPSModulePathwill exist
• Cygwin: Usually there isCYGWINvariable, orSHELL=/bin/bash
• CMD: If none of the above conditions are met, it is a traditional command prompt character.

PowerShell

Opening method

• Search for "PowerShell" in the start menu
• useWin + XSelect "Windows PowerShell" or "Windows Terminal"
• Enter in the folder path columnpowershellCan be opened directly from this path

Common commands

• Get-ChildItem: List files and folders in the current directory (similar todir)
• Set-Location: Switch directory (abbreviationcd)
• Clear-Host:Clear screen content (abbreviationcls)
• Copy-Item:Copy files or folders
• Remove-Item: Delete files or folders
• New-Item:Create a new file or folder
• Write-Output: Output text (abbreviationecho)
• Exit:Close PowerShell

Advanced operation

• Pipeline processing:Get-Process | Where-Object {$_.CPU -gt 100}
• Redirect output:Get-ChildItem > file.txt
• Use aliases: e.g.ls、cp、rmCan be used directly
• Variable usage:$var = "Hello", output$var
• Functions and scripts: can be created.ps1File execution multiple commands

Administrator mode

• Search for "PowerShell", right-click and select "Run as administrator"
• Can be executed asGet-EventLog、Set-ExecutionPolicyWaiting for system maintenance instructions

List all variables in PowerShell

command method

• Get-Variable: Displays all variables in the current PowerShell session
• Get-ChildItem variable:: View all variables through variable driver

example

# List all variables
Get-Variable

# Show only variable names
Get-Variable | Select-Object Name

# Display the value of a specific variable
Get-Variable PATH | Format-List *

#Use variable driver method
Get-ChildItem variable:

illustrate

• Get-VariableBy default, fields such as name, value, description, etc. will be output.
• variable drivervariable:Can be used as a path to manipulate variables, for exampleGet-Content variable:PATH
• To filter, you can matchWhere-Object,For example:
  Get-Variable | Where-Object { $_.Name -like "P*" }Only show variables starting with P

C and C++ languages

High efficiency

• Closely integrated with the underlying hardware, it can directly operate memory and hardware resources.
• The code execution speed is fast and is suitable for application scenarios that require high performance.

static type system

• Check type errors during compilation to reduce the occurrence of execution errors.
• Supports type conversion for flexible programming.

Resource control

• Provides fine memory management capabilities, allowing developers to manually configure and release memory.
• Suitable for developing embedded systems and real-time applications that require efficient resource utilization.

Grammatical flexibility

• The syntax is concise and provides highly controlled structural design capabilities.
• Supports multiple programming paradigms, including procedural and object-oriented.

Wide range of applications

• C: Used for system programming, driver, and embedded development.
• C++: used for game development, graphics processing, financial calculations and large-scale software development.

Cross-platform

• Available on most operating systems and hardware architectures.
• While providing high performance, it can meet cross-platform needs.

C/C++ preprocessor directives

Overview

Preprocessor Directives are commands executed by a program module called "Preprocessor" before the C or C++ program code is officially processed (compiled) by the compiler. These instructions are preceded by a pound sign (#) and not preceded by a semicolon (;) at the end.

The role of the preprocessor is to perform text replacement, file inclusion, conditional compilation and other operations to generate the final "Translation Unit" for use by the compiler.

Main command categories and examples

1. File Inclusion

Used to insert the contents of other files into the current file.

• #include <filename>: Used to include standard library header files. The preprocessor searches for files in the compiler's default system header file path.
• #include "filename": Used to contain user-defined header files. The preprocessor will first search the directory of the current source code file, and then search the system path.

#include <iostream> // Include the standard I/O library
#include "my_utility.h" // Include custom header file

2. Macro Definition and Substitution

Literal replacement macro used to define symbolic constants or code fragments.

• #define: Define macro.
• #undef: Undefine macro.

#define MAX_SIZE 100 //Define a symbolic constant
#define SUM(a, b) ((a) + (b)) // Define a macro with parameters

int main() {
    int size = MAX_SIZE; // After preprocessing, it becomes int size = 100;
    int total = SUM(5, 3); // After preprocessing, it becomes int total = ((5) + (3));
    return 0;
}

3. Conditional Compilation

Allows you to decide whether a specific block of code should be compiled based on the value of a preprocessor macro.

• #ifdef: If the macro is defined, compile subsequent blocks.
• #ifndef: If the macro is not defined, compile subsequent blocks. (commonly used in Include Guard)
• #if: Determine whether to compile based on the value of the given constant integer expression.
• #else / #elif: Substitution or subsequent judgment of conditional compilation.
• #endif: End the conditional compilation block.

#define DEBUG_MODE 1

#if DEBUG_MODE
    std::cout << "Debug mode enabled" << std::endl;
#else
    std::cout << "Production mode running" << std::endl;
#endif

#ifndef MY_HEADER_H
#defineMY_HEADER_H
    // ... header file content (Include Guard example)
#endif

4. Error and Warning

• #error: Forces the preprocessor to stop compilation when encountering this directive and display the specified error message.
• #warning: Output the specified warning message without stopping compilation (non-standard directive, but supported by many compilers).

#ifndef VERSION_DEFINED
#error "Must define version number macro VERSION_DEFINED!"
#endif

5. Other Directives

• #pragma: Used to send special instructions to the compiler, the behavior is highly dependent on the compiler (e.g.#pragma onceand#pragma warning)。
• #line: Used to change the current line number and file name displayed when the compiler reports errors or warnings.

#pragma once

Definition and function

#pragma onceIs one of the C and C++ languagesPreprocessor Directive. Its core function is to ensure that the header file containing this directive will only be processed by the compiler during a single compilation process.once。

The purpose of this directive is to solve the problem of repeated inclusion of header files and prevent the same piece of code (such as class definitions, function prototypes or constant declarations) from being seen multiple times by the compiler due to multiple source code files or multi-level header file inclusion relationships, thereby causingRedefinitioncompilation error.

Comparison with Include Guard

In the C/C++ standard, traditionally by usingInclude GuardTo achieve the purpose of preventing repeated inclusion.#pragma onceProvides a cleaner alternative.

Include Guard (standard method)

This is the standard and portable approach, using conditional compilation directives:

#ifndef MY_HEADER_H
#defineMY_HEADER_H

// Header file content

#endif // MY_HEADER_H

It relies on a unique macro name (e.g.MY_HEADER_H) to control whether content is included.

#pragma once (non-standard method)

Use a single line directive:

#pragma once

// Header file content

The compiler will automatically track whether it has already processed this file in the current compilation session, and if so, skip the remaining contents of the file.

Advantages and Disadvantages

Advantages of #pragma once:

• Simplicity:Just one line of code eliminates the need to invent and maintain unique macro names.
• Compilation efficiency:For large projects, since the compiler can directly determine whether to skip based on the file path, it can theoretically provide a slight improvement in compilation speed.

Disadvantages of #pragma once:

• Non-standard:It is not part of the C++ language standard. Although it is supported by all major compilers (including MSVC, GCC, Clang), it may not be recognized on rare or older compilers.
• Path dependency:It relies on the compiler being able to correctly identify the uniqueness of the archive path. In some complex project structures or symbolic link environments, problems may occur (although they are extremely rare).

in conclusion

In modern development environments, especially when using Visual Studio or mainstream compilers,#pragma onceBecause of its simplicity and convenience, it prevents repeated inclusion of header files.Common and recommendedpractices.

#pragma warning

Definition and purpose

#pragma warningare preprocessor directives in the C and C++ languages ​​that are used within specific blocks of code.Selectively suppress, restore, or modify the severity of compiler warnings。

The main purpose of this directive is to provide finer control than project-level settings. Use this command, for example, when you need to include an older version of a library that generates a lot of warnings, but don't want to turn off the warnings globally.

Main operation syntax (MSVC format)

Although#pragma warningThe specific implementation may vary slightly between different compilers (especially MSVC is the most widely used), but its core actions are divided into the following types:

1. Disable warning (Disable)

From this point on, the compiler ignores the specified warning code.

#pragma warning( disable : 4996 ) // Disable C4996 (for example: warning about using unsafe functions)
// Contain or write code that will generate C4996

2. Recovery warning (Default)

Restores the specified warning to the default behavior set by the project or command line.

#pragma warning(default: 4996) //Restore C4996 to default state

3. Save and restore status (Push and Pop)

This is the safest and most recommended mode. It allows the warning settings to be modified while processing a specific block of code, and then ensures that the original warning state is restored immediately after the block ends to avoid side effects.

• #pragma warning( push ): Store the current stacking status of warning settings.
• #pragma warning( pop ):Restore to the latestpushThe stored state.

#pragma warning( push ) // 1. Save the current warning settings
#pragma warning( disable : 4996 4244 ) // 2. Disable multiple specific warnings
// Contains third-party header files or legacy code
#pragma warning( pop ) // 3. Restore to original settings
// Subsequent code will use the restored settings

4. Increase warning level (Error)

Promote a specific warning code to a compilation error. If the warning occurs, compilation will fail.

#pragma warning( error : 4005 ) // Treat warning 4005 as an error

Things to note

• Compiler specificities:warning number (e.g.4996)yescompiler specific. You must look up the correct number for the compiler you are using (such as MSVC).
• GCC/Clang differences:Although GCC and Clang support partial#pragma warningfunctions, but they generally prefer to use-W...command line flags or special#pragma GCC diagnosticStructure to control warnings.

C++ standard namespace

Overview

In C++ programming,std::Refers to the Standard Namespace. It is a container that contains all standard entities in the C++ language core and its Standard Library (Standard Library), including functions, classes, templates, macros, and objects.

usestd::The main purpose is to avoidName Collisions. If a standard function (e.g.coutorvector) are not placed in a separate namespace. When the user defines an entity with the same name, the compiler will not know which one to use.

The main content of std::

Most of the functionality of the C++ standard library is located instd::within the namespace. Main categories and features include:

1. Input/output streams (I/O Streams)

• std::cout: Standard output streaming object, used to print data to the console.
• std::cin: Standard input stream object, used to read data from the console.
• std::endl: Output newlines and flush the buffer.
• std::ifstream / std::ofstream: File input/output streaming.

2. Standard Containers (Containers)

Collection categories used to store data:

• std::vector:A dynamically sized array.
• std::list: Two-way connection sequence.
• std::map: A sorted collection of key-value pairs (red-black tree implementation).
• std::unordered_map: An unordered collection of key-value pairs (hash table implementation).
• std::set: A sorted collection of unique key values.
• std::string: Category used for processing strings.

3. Algorithms

A set of common functions for container and range operations:

• std::sort: Sort elements within the range.
• std::find: Find the specified value within the range.
• std::copy: Copies elements from one range to another.

4. Other core tools (Utilities)

• std::shared_ptr / std::unique_ptr: Intelligent indicators for automated memory management.
• std::thread: Used for multi-thread programming.
• std::function: Universal function wrapper.
• std::pair / std::tuple: Template used to store a fixed number of values ​​of different types.

Using std::

to accessstd::There are two main methods for entities in a namespace:

1. Fully Qualified Names

Write it clearly every time you use itstd::prefix. This is the safest and recommended approach, especially in header files, to avoid polluting the global namespace.

int main() {
    std::cout << "Hello World" << std::endl;
    std::vector<int> numbers;
    return 0;
}

2. using declarations (using Directives)

useusing namespace std;The entirestd::The contents of the namespace are introduced into the current scope, and the name can be used directly withoutstd::prefix.

#include <iostream>
using namespace std; // Introduce std:: namespace

int main() {
    cout << "Hello World" << endl; // No std:: prefix required
    vector<int> numbers;
    return 0;
}

althoughusing namespace std;Convenient, but should be avoided in large projects or header files as it increases the risk of naming conflicts.

std::string

1. Declaration and initial value

In C++,std::stringIs a category (Class). When you declare astd::stringWhen a variable is not given an initial value, it calls the Default Constructor.

• Initial value:Default isempty string(Right now"")。
• length:The length is0。
• Memory:It is not equivalent to the C languageNULLindicator, but a valid object.

#include <string>
#include <iostream>

void initialization_example() {
    std::string s; // declared but no value given
    
    std::cout << "Length: " << s.length() << std::endl; // Output: 0
    std::cout << "Content: '" << s << "'" << std::endl; // Output: ''
}

2. How to check if it is empty (Empty)

To check if a string does not contain any characters, the most recommended way is to useempty()member functions, which are better than checkinglength() == 0More semantic and more efficient in some containers.

void check_empty(std::string s) {
    if (s.empty()) {
        std::cout << "This is an empty string." << std::endl;
    }
}

3. How to check if it is Null?

This is a common misunderstanding:The std::string object itself is never null.

In C++, only "pointers" can benullptr. If you are referring to checking whether a string has not been assigned a value, usually checking whether itempty(). If you are dealing with C-style string pointers (char*), you need to check for null.

4. Method to clear string

If you want to restore an existing string to its original empty state, you can use the following method:

void clear_example() {
    std::string s = "Hello";
    
    s.clear(); // Method 1: Most commonly used
    s = ""; // Method 2: Reassign
    s = std::string(); // Method 3: Assign a default object
}

float to std::string

1. Use std::to_string (the simplest method)

This is the most straightforward method introduced in C++11, converting a numeric value into a string. By default, it retains 6 decimal places.

#include <string>
#include <iostream>

void simple_convert() {
    float val = 3.14159f;
    std::string s = std::to_string(val);
    
    std::cout << s << std::endl; // Output example: "3.141590"
}

2. Use std::stringstream (custom precision)

If you need to control the number of decimal places or a specific format,stringstreamIt is the most flexible option. it combines<iomanip>library to precisely control the output.

#include <sstream>
#include <iomanip>

void precision_convert() {
    float val = 3.1415926f;
    std::stringstream ss;
    
    // Set fixed point counting method (fixed) and retain 2 decimal places
    ss << std::fixed << std::setprecision(2) << val;
    
    std::string s = ss.str();
    std::cout << s << std::endl; // Output: "3.14"
}

3. Use std::format (C++20 efficient method)

In C++20, you can use format strings to handle type conversion and precision control at the same time, with the most concise syntax.

#include <format>

void modern_format() {
    float val = 123.456f;
    
    // {:.2f} represents float type and retains two decimal places
    std::string s = std::format("{:.2f}", val);
    
    std::cout << s << std::endl; // Output: "123.46" (will be automatically rounded)
}

4. Related processing usage: remove extra zeros at the end

std::to_stringOften there are extra zeros like "3.140000". you can combinefind_last_not_ofanderaseto clean up the string.

std::string remove_trailing_zeros(float val) {
    std::string s = std::to_string(val);
    
    // Remove trailing '0' until non-zero character is encountered
    s.erase(s.find_last_not_of('0') + 1, std::string::npos);
    
    // If the last character is a decimal point, also remove it
    if (s.back() == '.') {
        s.pop_back();
    }
    return s;
}

Conversion method comparison

std::string delimiter parsing

1. Use std::string::find loop analysis

This is the most standard approach in C++ that does not rely on external libraries. throughfindFind the position of the delimiter and usesubstrRetrieve substring.

#include <iostream>
#include <vector>
#include <string>

void split_by_three_chars() {
    std::string text = "Apple---Banana---Cherry---Date";
    std::string delimiter = "---";
    std::vector<std::string> tokens;
    
    size_t pos = 0;
    size_t last_pos = 0;

    //Loop to find delimiter
    while ((pos = text.find(delimiter, last_pos)) != std::string::npos) {
        tokens.push_back(text.substr(last_pos, pos - last_pos));
        last_pos = pos + delimiter.length();
    }
    
    //Put in the last fragment
    tokens.push_back(text.substr(last_pos));

    //output result
    for (const auto& t : tokens) std::cout << "[" << t << "]" << std::endl;
}

2. Use std::string_view (C++17 performance optimization)

If your string is very large, usestd::string_viewIt can avoid generating a large number of string copies during the splitting process, thereby greatly improving performance.

#include <string_view>
#include <vector>

std::vector<std::string_view> split_sv(std::string_view str, std::string_view del) {
    std::vector<std::string_view> output;
    size_t first = 0;

    while (first < str.size()) {
        const auto second = str.find(del, first);
        if (second == std::string_view::npos) {
            output.emplace_back(str.substr(first));
            break;
        }
        output.emplace_back(str.substr(first, second - first));
        first = second + del.size();
    }
    return output;
}

3. Use Regex regular expressions

If you want your code to look more concise, or if the separators may change, you can usestd::regex. But note that the execution performance of regular expressions is usually better than that of manualfindslow.

#include <regex>

void regex_split() {
    std::string text = "One###Two###Three";
    std::regex ws_re("###"); // Define 3 character separator
    
    std::copy(std::sregex_token_iterator(text.begin(), text.end(), ws_re, -1),
              std::sregex_token_iterator(),
              std::ostream_iterator<std::string>(std::cout, "\n"));
}

Comparison of analytical methods

Check if std::vector is empty

Overview

In C++, to checkstd::vectorIf the container contains no elements (i.e. has size zero), the most standard and recommended way is to use its member functionsempty(). This is becauseempty()Functions are usually better than direct checkssize()Is equal to zero more efficient, especially in some container implementations.

Check method

1. Use empty() function (recommended)

This is the preferred way to check if a vector is empty. It returns a Boolean value: if vector has no elements, it returnstrue; Otherwise returnfalse。

#include <vector>
#include <iostream>
#include <string>

void check_empty(const std::vector<std::string>& vec)
{
    if (vec.empty()) {
        std::cout << "Vector is empty (empty() == true)." << std::endl;
    } else {
        std::cout << "Vector is not empty (empty() == false). Number of elements: " << vec.size() << std::endl;
    }
}

int main()
{
    std::vector<std::string> empty_vec;
    std::vector<std::string> non_empty_vec = {"apple", "banana"};

    check_empty(empty_vec);
    check_empty(non_empty_vec);

    return 0;
}

2. Check the size() function

While valid, this is not the most idiomatic or potentially most efficient way of checking for empty state. It directly checks whether the number of elements in the vector is zero.

void check_size(const std::vector<int>& vec)
{
    if (vec.size() == 0) {
        std::cout << "Vector is empty (size() == 0)." << std::endl;
    } else {
        std::cout << "Vector is not empty (size() != 0)." << std::endl;
    }
}

Efficiency considerations

forstd::vectorIn terms ofempty()andsize() == 0The time complexity is $O(1)$ because vector stores its size as a member variable. However, for some other C++ standard containers (e.g.std::listorstd::forward_list), the standard recommendation is always to useempty(), because it is generally the most common and fastest idiom in the C++ standard container library for checking for empty status.

C++ parsing string into string vector

Overview

In C++, a character containing a space (Space) or Tab character (\t) separated string is parsed into individual words or tokens and stored instd::vector<std::string>, the most common method is to use **std::stringstream**. File streaming categorystd::stringstreamLike an in-memory stream, it defaults to using whitespace characters (including spaces, tabs, and newlines) as delimiters for reading operations.

Code example

The following is usedstd::stringstreamExample of C++ code that accomplishes this task:

#include <iostream>
#include <sstream> // include stringstream
#include <string>
#include <vector>

using namespace std;

/**
 * Separate the input string by any whitespace characters (space or Tab) and store it in the vector.
 * @param input_str The string to be parsed.
 * @return A vector containing all separated words.
 */
vector<string> split_string_by_whitespace(const string& input_str)
{
    vector<string> tokens;
    
    // 1. Create a stringstream object and initialize it with the input string
    stringstream ss(input_str);
    
    string token;
    
    // 2. Loop reading
    // The stream extraction operator (>>) will automatically read the next token using whitespace characters (space, Tab, etc.) as delimiters.
    // The operator returns true when a token is successfully read, otherwise (the end of the string is reached) it returns false.
    while (ss >> token)
    {
        tokens.push_back(token);
    }
    
    return tokens;
}

int main()
{
    //The input string contains multiple spaces and tab symbols (\t)
    string test_string = "Hello \tWorld this is \ta test string";
    
    cout << "Original string: " << test_string << endl;
    
    vector<string> result = split_string_by_whitespace(test_string);
    
    cout << "--- separate results ---" << endl;
    
    for (size_t i = 0; i < result.size(); ++i)
    {
        cout << "Token " << i + 1 << ": [" << result[i] << "]" << endl;
    }
    
    return 0;
}

Key mechanism explanation

• std::stringstream ss(input_str): This category treats the input string as an input stream (Input Stream), and its function is similar to that used when reading from a file.std::ifstream。
• Stream extraction operator (>>)：
  • It is the core mechanism of C++ input streaming.
  • It automatically ignores any leading whitespace characters (including spaces, tabs, newlines, etc.).
  • It reads all consecutive non-whitespace characters until the next whitespace character is encountered or the end of the stream.
  • This perfectly meets the requirement of "separated by any whitespace character" and automatically handles the case of multiple consecutive separators.
• while (ss >> token): This is an idiomatic C++ stream read loop. As long as the stream successfully extracts and stores the data totokenvariable, the loop will continue to execute.

.NET and std str transcoding garbled characters

reason

In C++/CLI,gcnew System::String(stdStr.c_str())willstd::string(usuallyANSI / UTF-8encoding) directly into .NETSystem::String,andSystem::StringThe expectation isUTF-16。
ifstdStrIf it contains non-ASCII characters (such as Chinese), garbled characters will appear.

Solution 1: Usemarshal_as(suggestion)

Namespace needs to be quoted:

#include <msclr/marshal_cppstd.h>
using namespace msclr::interop;

std::string stdStr = "Chinese test";
System::String^ netStr = marshal_as<System::String^>(stdStr);

✅ This method can automatically convert UTF-8 / ANSI to .NET Unicode correctly.

Solution 2: If you are sure that std::string is UTF-8, you can manually convert it

#include <codecvt>
#include <locale>

std::string utf8Str = u8"Chinese test";
std::wstring wideStr = std::wstring_convert<std::codecvt_utf8_utf16<wchar_t>>{}.from_bytes(utf8Str);
System::String^ netStr = gcnew System::String(wideStr.c_str());

Solution 3: If the source isstd::wstring

If the C++ side originally uses wide character strings, just use it directly:

std::wstring wstr = L"Chinese test";
System::String^ netStr = gcnew System::String(wstr.c_str());

suggestion

• likestd::stringThe source is UTF-8 (common in modern systems) → usemarshal_asorcodecvt。
• If it is an old ANSI character encoding (such as local encoding Big5, Shift-JIS) → it is recommended to convert it to UTF-8 in C++ first.
• Do not use it directlygcnew String(stdStr.c_str()), unless the string is pure English.

C++ multidimensional array initialized to 0

Initialized using std::array

In C++, you can usestd::arrayInitialize a multidimensional array to 0:

#include <iostream>
#include <array>
using namespace std;

int main() {
    array<array<int, 4>, 3> arr = {0}; // Initialize all elements to 0

    for (int i = 0; i < 3; i++) {
        for (int j = 0; j < 4; j++) {
            cout << arr[i][j] << " ";
        }
        cout << endl;
    }
    return 0;
}

Things to note

• std::arrayIs a container class in the C++ standard library, suitable for situations where the size is fixed at compile time.
• When initializing, you can directly set the element to 0 to ensure that all values ​​are cleared.
• Compared with traditional arrays,std::arrayProvides more functionality and security.

C++ std array loop method

Overview

In C++, looping is the basic operation of iterating over all elements in a container or array.std::vector(dynamically sized container) andstd::array(fixed-size containers) are containers in the C++ Standard Library (STL) whose elements can be iterated over in a variety of standard and safe ways.

1. Range-based for loop

This is the most modern, safest, and most concise traversal method introduced in C++11. It works with all standard containers (includingstd::vectorandstd::array）。

applicability:

• std::vector: **Preferred**
• std::array: **Preferred**

#include <iostream>
#include <vector>
#include <array>

void range_based_loop_example()
{
    std::vector<int> vec = {10, 20, 30};
    std::array<int, 4> arr = {1, 2, 3, 4};
    
    // Use auto& to traverse by reference, which is efficient and allows modification of elements
    std::cout << "--- Vector (modifiable) ---" << std::endl;
    for (auto& element : vec) {
        element += 1; // Modify element
        std::cout << element << " ";
    }
    std::cout << std::endl;
    
    // Use const auto& to traverse by constant reference, which is safe and does not allow modification of elements
    std::cout << "--- Array (read only) ---" << std::endl;
    for (const auto& element : arr) {
        std::cout << element << " ";
    }
    std::cout << std::endl;
}

2. Index-based for loop

This is a traditional and universal method. Suitable for situations where you need to access element indexes (for example: you need to operate two containers or arrays at the same time, or terminate at a specific index).

applicability:

• std::vector：applicable, usesize()Get the boundaries.
• std::array：applicable, usesize()Get the boundaries.

#include <iostream>
#include <vector>
#include <array>

void index_loop_example()
{
    std::vector<std::string> names = {"Alice", "Bob", "Charlie"};
    std::array<double, 3> prices = {10.5, 20.99, 5.0};
    
    // Use size_t or auto with size() to ensure the index type is correct and the bounds are safe
    for (size_t i = 0; i < names.size(); ++i) {
        // Use operator[] to access elements
        std::cout << "Name: " << names[i] << ", Price: " << prices[i] << std::endl;
    }
}

NOTE: Use[]C++ does not perform bounds checking when operators access elements. If bounds checking is required, useat()function (it will throw when out of boundsstd::out_of_rangeabnormal).

3. Iterator loop

This is the most flexible traversal method in C++ and works with all standard containers. Iterators are necessary when interacting with C++ Standard Library (STL) algorithms.

applicability:

• std::vector: **Recommended** (when STL algorithm or specific control is required)
• std::array: **Recommended** (when STL algorithm is required)

#include <iostream>
#include <vector>

void iterator_loop_example()
{
    std::vector<double> values = {100.0, 50.0, 10.0};
    
    // Simplify the iterator type using auto and use begin() and end() to get the range
    for (auto it = values.begin();
         it != values.end();
         ++it)
    {
        *it /= 10.0; // *it dereferences the iterator and accesses elements
        std::cout << *it << " ";
    }
    std::cout << std::endl;
    
    // Use cbegin() and cend() to ensure the iterator is const for read-only traversal
    for (auto cit = values.cbegin(); cit != values.cend(); ++cit)
    {
        // Trying *cit = 5; will result in a compilation error
        std::cout << *cit << " ";
    }
    std::cout << std::endl;
}

Summary and recommendations

Method for copying the contents of a multidimensional std::array

problem situation

std::array<std::array<std::array<float, 2>, 2>, 2> twoLines;
std::array<std::array<std::array<float, 2>, 2>, 2> twoLinesCopy;

Solution 1: Directly specify (= operator)

std::arraySupports complete copying from shallow to deep levels, so it can be used directly=：

twoLinesCopy = twoLines;

Solution 2: Use std::copy

std::copy(twoLines.begin(), twoLines.end(), twoLinesCopy.begin());

Solution 3: Use std::memcpy (suitable for POD type, float is applicable here)

std::memcpy(&twoLinesCopy, &twoLines, sizeof(twoLines));

suggestion

• The simplest and safest way:twoLinesCopy = twoLines;
• std::copySuitable for applications requiring STL generic methods.
• memcpyOnly applicable to POD (Plain Old Data), and no constructor/deconstructor logic.

C++ capture indicator memory access error

Overview

In C++, this happens when you access an invalid or out-of-bounds memory location when you manipulate pointers or use indexes to access arrays/memory.Undefined Behavior. This behavior usually manifests itself as a programCrash, such as a Segmentation Fault or Access Violation error.

C++ standardtry-catchThe mechanism is mainly used to catch C++ exceptions (Exceptions) that are explicitly thrown in the program code, but cannot directly catch hardware or memory access errors issued by the operating system.

Why can't try-catch catch memory access errors?

When a C++ program attempts to access invalid memory, for example:

1. Out-of-Bounds Access:Access memory after the end of the array.
2. Dereferencing a Null Pointer:An attempt was made to read or write at addressnullptrof memory.
3. Use After Free:Access passeddeleteorfreeFreed memory.

These operations will trigger the protection mechanism of the underlying operating system (such as Access Violation on Windows or SIGSEGV signal on Unix/Linux), and the program will be terminated by the operating system. These errors are not C++ language-level exception objects (such asstd::exception), so the standardtry-catchThere is no way to capture them.

Error example (standard try-catch does not work)

#include <iostream>
#include <vector>

using namespace std;

void crash_function()
{
    // Array out-of-bounds access example
    vector<int> vec = {10, 20};
    cout << "Attempt to access out-of-bounds index..." << endl;
    
    // This is undefined behavior and is likely to lead to an access violation crash
    int value = vec[100]; // Error: Accessing invalid index
    cout << "value: " << value << endl;
}

int main()
{
    try
    {
        crash_function();
        //The program will crash at vec[100] above and will not be executed here.
        // And the catch block will not be executed.
    }
    catch (const exception& e)
    {
        // Operating system level errors cannot be captured here
        cerr << "Caught C++ exception: " << e.what() << endl;
    }
    catch(...)
    {
        // Still unable to catch memory access errors
        cerr << "Unknown exception caught" << endl;
    }

    return 0;
}

How to deal with this kind of error?

1. Use safe C++ mechanisms (preferred)

The best approach is to prevent such errors from happening during the design phase of your code rather than trying to catch them:

• Container access:forstd::vectorandstd::string,useat()member function instead[]operator. whenat()When out of bounds, it throws a C++ exceptionstd::out_of_range, can be standardizedtry-catchcapture.
• Indicator management:Use smart indicators (std::unique_ptr, std::shared_ptr) to automate the memory life cycle. Never dereferencenullptr, and checks whether the pointer is valid before dereferencing it.

void safe_function()
{
    vector<int> vec = {10, 20};
    try
    {
        int value = vec.at(100); // throw std::out_of_range exception
        cout << "value: " << value << endl;
    }
    catch (const std::out_of_range& e)
    {
        // Successfully caught C++ exception
        cerr << "Safe capture: " << e.what() << endl;
    }
}

2. Use platform-specific Structured Exception Handling (SEH)

In a Windows environment, you can use Microsoft's extended Structured Exception Handling (SEH) to catch operating system-level errors, including Access Violations. This is usually done using__tryand__exceptKeywords.

Note: SEH is non-standard and not recommended for use in portable C++ code. In C++/CLI projects, you can use .NET'stry/catchCatch some SEH exceptions, but this requires specific compiler settings (e.g./EHa)。

C++ catching array index memory access error

Overview

In C++, when you use a raw array (C-style array) or pointer for indexing operations, it occurs if a memory location that is outside the range of the array or is invalid is accessed.Undefined Behavior. This behavior often results in programs running at the operating system levelCrash, such as Access Violation or Segmentation Fault.

The key points are:C++ standardtry-catchThe mechanism is designed to catch C++ exceptions (Exceptions) that are explicitly thrown in the program code. It cannot catch hardware or memory protection errors issued by the operating system.

Why doesn't standard try-catch work?

When a C++ program attempts to access invalid memory (for example, accessing a static arrayarr[100], but the array only has 10 elements), which violates the operating system's memory protection rules. The operating system steps in and sends a signal or exception to terminate the program, preventing it from damaging the system or other programs. These errors are not C++ language levelstd::exceptionobject and therefore cannot be used by the standardtry-catchBlock capture.

Error demonstration (standard try-catch cannot catch crashes)

#include <iostream>
#include <stdexcept> // Include standard exceptions

using namespace std;

void crash_function()
{
    // C-style primitive array
    int raw_array[5] = {1, 2, 3, 4, 5};
    cout << "Attempt to access out-of-bounds index..." << endl;
    
    // Array out-of-bounds access: This is a typical undefined behavior
    // If a memory that the program does not have permission to access is accessed, an access violation will be triggered immediately, causing the program to crash.
    int value = raw_array[100]; // Error: Accessing invalid index
    cout << "Value (if reachable): " << value << endl;
}

int main()
{
    try
    {
        crash_function();
        // The program will terminate at raw_array[100] and will not be executed here.
    }
    catch (const exception& e)
    {
        // Operating system level memory access errors cannot be captured here
        cerr << "Caught C++ exception: " << e.what() << endl;
    }

    return 0;
}

Correct handling and prevention methods (preferred)

The best practice is to avoid out-of-bounds accesses entirely, or to use the safe containers and methods provided by C++, which throw catchable C++ exceptions when an error occurs:

1. Use std::vector::at()

forstd::vectorcontainer, useat()member function instead of original[]operator. When access goes out of bounds,at()will throw **std::out_of_range**Exception, this is a standard C++ exception and can betry-catchCaught and disposed of safely.

#include <vector>
// ... (other include)

void safe_vector_access()
{
    std::vector<int> vec = {10, 20};
    try
    {
        // Use at() for bounds checking
        int value = vec.at(100);
        std::cout << "value: " << value << std::endl;
    }
    catch (const std::out_of_range& e)
    {
        // Successfully caught C++ exception
        std::cerr << "Safely caught out-of-bounds exception: " << e.what() << std::endl;
    }
}

2. Avoid using raw arrays and use std::array instead

Introduced using C++11std::arrayReplaces C-style arrays. Althoughstd::arrayof[]The operator is not checked, but itsat()The function will also throwstd::out_of_rangeabnormal.

#include <array>
// ...

void safe_array_access()
{
    std::array<int, 5> arr = {1, 2, 3, 4, 5};
    try
    {
        int value = arr.at(5); // Access out of bounds, throw exception
    }
    catch (const std::out_of_range& e)
    {
        std::cerr << "std::array out-of-bounds capture: " << e.what() << std::endl;
    }
}

Platform-specific structured exception handling (advanced/non-standard)

On certain platforms such as Windows, you can use **Structured Exception Handling (SEH)** to catch hardware exceptions (such as access violations) issued by the operating system. This usually involves using a Microsoft extension__tryand__exceptKeywords. This method is platform non-standard, sacrifices code portability, and is generally not recommended as a routine error handling mechanism.

.NET List

Basic introduction

In .NET (languages ​​like C++/CLI, C#, VB.NET, etc.),List<T>is a generic collection category, Can store any type of data (T) and provide dynamic resizing function. it belongs toSystem::Collections::Genericnamespace.

Declaration and initialization

//Import namespace
using namespace System;
using namespace System::Collections::Generic;

int main()
{
    // Declare a List to store int
    List<int>^ numbers = gcnew List<int>();

    //Add elements directly during initialization
    List<String^>^ names = gcnew List<String^>({ "Alice", "Bob", "Charlie" });
}

Add, remove and access elements

List<int>^ nums = gcnew List<int>();

//Add new element
nums->Add(10);
nums->Add(20);
nums->Add(30);

//Insert at specified position
nums->Insert(1, 15); // Insert 15 at index 1

//Remove specified value
nums->Remove(20);

//Remove the specified index
nums->RemoveAt(0);

// Get elements
int value = nums[1]; // Get the element at index 1

// Check if it contains
if (nums->Contains(30))
    Console::WriteLine("Found 30");

// Clear all elements
nums->Clear();

Traverse List

// for loop
for (int i = 0; i < nums->Count; i++)
{
    Console::WriteLine("{0}th element: {1}", i, nums[i]);
}

// for each loop
for each (int n in nums)
{
    Console::WriteLine(n);
}

Common properties and methods

Example output

10
15
30

C++/CLI List access out of range

1. What happens when an index goes out of range?

in C++/CLISystem::Collections::Generic::List<T>, if you try to access an index that doesn't exist (e.g.sizeis 1 but accessindex1), programWon'tGiven a default value (such as 0 or null), an exception will be thrown immediately.

• Exception thrown: System::ArgumentOutOfRangeException
• result:If exception handling is not performed, the program will crash.
• reason:.NET collection classes have strict bounds checking to prevent buffer overflows or unexpected memory accesses.

2. Safe access practices

Before making an index access, you should always checkCountattribute, or usetry-catchBlocks capture potential errors.

//Method A: Pre-inspection (most recommended, best performance)
if (nums->Count > 1) {
    int value = nums[1];
    // Processing logic
} else {
    // Handle insufficient index situation
}

//Method B: Exception catching
try {
    int value = nums[1];
} catch (System::ArgumentOutOfRangeException^ ex) {
    System::Diagnostics::Debug::WriteLine("Index out of range!");
}

3. Performance in different situations

4. Summary of preventive measures

• Check length:Always accessingnums[i]Confirm beforei < nums->Count。
• Initialize default values:If you want to get a default value if it cannot be found, you can use the ternary operator:

  int value = (nums->Count > 1) ? nums[1] : 0;

• Use Enumerable:If using LINQ (requires citationSystem::Linq), you can useElementAtOrDefault(1), which returns the type's default value (0 for int) if the index does not exist.

C++ std maximum and minimum values

1. C++ standard library functions: std::min and std::max

In C++, the most commonly used tools are defined in<algorithm>instd::minandstd::max. They support basic types, objects, and even initialization lists.

#include <algorithm>
#include <iostream>
#include <vector>

void basic_usage() {
    int a = 10, b = 20;

    // 1. Compare two numbers
    int small = std::min(a, b);
    int large = std::max(a, b);

    // 2. Compare initialization lists (C++11)
    int lowest = std::min({5, 1, 9, 3}); // return 1

    // 3. std::minmax (C++11) - get the maximum and minimum values at the same time
    auto result = std::minmax({10, 20, 30, 40});
    std::cout << "Min: " << result.first << ", Max: " << result.second;
}

2. Numeric limits: FLT_MAX and std::numeric_limits

When you need to initialize a "minimum" variable, you usually set it to the largest positive number that the type can express, to ensure that it will be updated on the first comparison.

C-style limits (from <cfloat> and <climits>)

• FLT_MAX: floatThe maximum value (about $3.4 \times 10^{38}$).
• DBL_MAX: doublethe maximum value.
• INT_MAX: intthe maximum value.

C++ style limits (recommended to use <limits>)

std::numeric_limitsProvides a consistent and templated way to obtain numeric properties, which is very useful when writing generic code.

#include <limits>

void limits_example() {
    // Get the maximum value
    float max_f = std::numeric_limits<float>::max();
    int max_i = std::numeric_limits<int>::max();

    // Get the "minimum positive number" (note: for floating point numbers, min() returns the minimum positive number instead of the maximum negative number)
    float min_positive_f = std::numeric_limits<float>::min();

    // Get the "lowest value" (the real smallest negative number)
    float lowest_f = std::numeric_limits<float>::lowest();
}

3. The maximum and minimum values ​​in the collection: std::min_element

If you are dealing withstd::vectoror array, you need to usestd::min_elementorstd::max_element, what they return isIterator。

#include <vector>
#include <algorithm>

void collection_example() {
    std::vector<float> scores = {88.5f, 92.0f, 79.5f, 100.0f};

    // Get the iterator of the maximum value
    auto it = std::max_element(scores.begin(), scores.end());

    if (it != scores.end()) {
        std::cout << "Highest score: " << *it;
    }
}

4. Key Differences and Pitfalls

C++ std gets pi

Overview

In the C++ standard math library, although the constant definition of $\pi$ was not officially standardized until C++20, there are several commonly used methods to obtain approximate values ​​of pi in different C++ versions and compiler environments.

To ensure code portability and accuracy, the most recommended approach is to use thestd::numbers::pi。

1. Use C++20 standard constants (recommended)

Starting with C++20, the standard library is<numbers>Precise and type-safe mathematical constants are provided in the header file.

Header file:

• #include <numbers>

How to use:

#include <iostream>
#include <numbers> // introduced in C++20

void use_cpp20_pi()
{
    // std::numbers::pi_v<T> is a template variable that provides an exact value based on type T
    // std::numbers::pi (without <T>) is equivalent to std::numbers::pi_v<double>
    
    double pi_double = std::numbers::pi;
    float pi_float = std::numbers::pi_v<float>;
    long double pi_long_double = std::numbers::pi_v<long double>;

    std::cout.precision(16); // Set output precision
    std::cout << "C++20 (double): " << pi_double << std::endl;
    std::cout << "C++20 (float): " << pi_float << std::endl;
}

2. Use macros of traditional C language (C++17 and before)

Before C++20, many developers relied on the C math library (<cmath>or<math.h>) provides a non-standard macro. Although these macros exist on most modern systems, they are not part of the C++ standard and are not guaranteed to work in all environments.

Header file:

• #include <cmath>(or<math.h>)

How to use:

#include <iostream>
#include <cmath> // Traditional C math library

void use_c_macro_pi()
{
    // M_PI is the most common PI macro, usually defined as double type.
    // Note: To enable this macro, the _USE_MATH_DEFINES macro may need to be defined on some systems.
    
    #ifdef M_PI
        double pi_value = M_PI;
        std::cout.precision(16);
        std::cout << "C Macro (M_PI): " << pi_value << std::endl;
    #else
        std::cout << "The M_PI macro is undefined. You may need to define _USE_MATH_DEFINES." << std::endl;
    #endif
}

3. Calculate or define by yourself

If you cannot use C++20 or C macros, you can define $\pi$ yourself as a constant, or use a mathematical function to calculate it (e.g. $\arccos(-1)$).

How to use:

#include <iostream>
#include <cmath>

void define_or_calculate_pi()
{
    // Define through mathematical operations:
    const double PI_CUSTOM_CALC = std::acos(-1.0);
    
    // Define it directly as a constant:
    const double PI_CUSTOM_DEFINE = 3.14159265358979323846;

    std::cout.precision(16);
    std::cout << "Custom Calculate: " << PI_CUSTOM_CALC << std::endl;
    std::cout << "Custom Define: " << PI_CUSTOM_DEFINE << std::endl;
}

Recommended summary

• **C++20 or newer:** Always usestd::numbers::pi。
• **C++17 or earlier:**
  • If the project allows, you can define it yourselfconst double PI = 3.14159...;
  • or try usingM_PIMacro, but be aware of its non-standard nature.

Calculate the standard deviation of a List

Sample program

#include <cmath>
#include <iostream>
using namespace System;
using namespace System::Collections::Generic;

int main()
{
    //Create test data
    Dictionary<int, List<float>^>^ data = gcnew Dictionary<int, List<float>^>();
    data->Add(1, gcnew List<float>({ 1.2f, 2.3f, 3.4f }));
    data->Add(2, gcnew List<float>({ 4.5f, 5.5f, 6.5f, 7.5f }));
    data->Add(3, gcnew List<float>()); // Test the empty list

    // Calculate the standard deviation of each key
    for each (KeyValuePair<int, List<float>^> entry in data)
    {
        int key = entry.Key;
        List<float>^ values = entry.Value;

        if (values == nullptr || values->Count == 0)
        {
            Console::WriteLine("Key {0}: No data", key);
            continue;
        }

        // Calculate the average
        double sum = 0.0;
        for each (float v in values)
        {
            sum += v;
        }
        double mean = sum / values->Count;

        // Calculate the number of mutations (number of maternal mutations)
        double variance = 0.0;
        for each (float v in values)
        {
            double diff = v - mean;
            variance += diff * diff;
        }
        variance /= values->Count; // If you want the sample variation, change to (values->Count - 1)

        // standard deviation
        double stddev = Math::Sqrt(variance);

        Console::WriteLine("Key {0}: standard deviation = {1:F4}", key, stddev);
    }

    return 0;
}

illustrate

• data: Dictionary that stores each group of floating point number lists, the key is an integer.
• mean: The average of all elements.
• variance: The squared mean of the difference between each element and the mean.
• stddev: The standard deviation is the square root of the variation.
• If you need the sample standard deviation, change the denominator tovalues->Count - 1。

Output example

Key 1: Standard deviation = 0.8981
Key 2: Standard deviation = 1.1180
Key 3: No information

.NET Dictionary

Basic introduction

In .NET (languages ​​like C++/CLI, C#, VB.NET, etc.),Dictionary<TKey, TValue>is a generic collection, Used to store "key-value pairs". Each key must be unique, and the value must be repeatable. it belongs toSystem::Collections::Genericnamespace.

Declaration and initialization

//Import namespace
using namespace System;
using namespace System::Collections::Generic;

int main()
{
    //Create a Dictionary with Key as int and Value as String
    Dictionary<int, String^>^ users = gcnew Dictionary<int, String^>();

    // Direct initialization
    users->Add(1, "Alice");
    users->Add(2, "Bob");
    users->Add(3, "Charlie");

    return 0;
}

Add, modify, delete and access elements

Dictionary<String^, int>^ ages = gcnew Dictionary<String^, int>();

//Add new element
ages->Add("Tom", 25);
ages->Add("Jerry", 30);

//Modify value (via indexer)
ages["Tom"] = 26;

// Add or modify (if key does not exist, it will be added automatically)
ages["Spike"] = 40;

// delete element
ages->Remove("Jerry");

//Access elements
if (ages->ContainsKey("Tom"))
{
    Console::WriteLine("Tom's age is {0}", ages["Tom"]);
}

// Try to get the value (avoid exceptions)
int age;
if (ages->TryGetValue("Spike", age))
{
    Console::WriteLine("Spike's age is {0}", age);
}

Traverse Dictionary

Dictionary<int, String^>^ users = gcnew Dictionary<int, String^>();
users->Add(1, "Alice");
users->Add(2, "Bob");
users->Add(3, "Charlie");

// Use KeyValuePair loop
for each (KeyValuePair<int, String^> entry in users)
{
    Console::WriteLine("Key = {0}, Value = {1}", entry.Key, entry.Value);
}

//Get only Keys
for each (int key in users->Keys)
{
    Console::WriteLine("Key: {0}", key);
}

//Get only Values
for each (String^ name in users->Values)
{
    Console::WriteLine("Name: {0}", name);
}

Common properties and methods

Example output

Key = 1, Value = Alice
Key = 2, Value = Bob
Key = 3, Value = Charlie
Tom's age is 26
Spike's age is 40

std::sort

1. Basic usage

std::sortIs the most commonly used sorting algorithm in the C++ Standard Template Library (STL). it is located<algorithm>In the header file, by default, elements within the range will beAscendingto sort.

#include <algorithm>
#include <vector>
#include <iostream>

void basic_sort() {
    std::vector<int> nums = {5, 2, 9, 1, 5, 6};
    
    // Sorting range: from start iterator to end iterator
    std::sort(nums.begin(), nums.end());
    
    // Result: 1, 2, 5, 5, 6, 9
}

2. Custom sorting criteria (descending power or specific conditions)

You can change the sorting logic by passing in the third parameter (comparison function or lambda expression).

// 1. Use predefined std::greater for exponentiation
std::sort(nums.begin(), nums.end(), std::greater<int>());

// 2. Use Lambda expressions to customize logic
std::sort(nums.begin(), nums.end(), [](int a, int b) {
    return a > b; // When true is returned, a will be ranked in front of b
});

3. Structure or category sorting

For custom types, comparison logic must be provided, otherwise the compiler will report an error because it does not know how to compare sizes.

struct Student {
    std::string name;
    int score;
};

void sort_students() {
    std::vector<Student> students = {{"Alice", 90}, {"Bob", 85}, {"Charlie", 95}};
    
    // Sort by score from high to low
    std::sort(students.begin(), students.end(), [](const Student& a, const Student& b) {
        return a.score > b.score;
    });
}

4. Underlying principle: Introspective sorting (Introsort)

std::sortNot a single sorting algorithm, but a hybrid algorithm designed to combine the advantages of each algorithm and avoid the worst-case scenario. Its operating logic is as follows:

• Quick Sort: Quick sort is used initially, which is the fastest.
• Heap Sort: When the recursion depth is too deep (which may cause quick sort to degenerate to $O(n^2)$), it will automatically switch to stacking sort, ensuring $O(n \log n)$.
• Insertion Sort: When the data interval is reduced to a certain extent (usually within 16 elements), switch to insertion sort to take advantage of its extremely low constant term performance under small data amounts.

Summary of performance and features

C++ template

1. What is a template?

In C++, template is a generic programming tool that allows us to write functions or categories without specifying specific data types, thereby making the code more reusable. Templates help handle different types of data in a single code, avoiding duplicate function or class definitions.

2. Function Template

Function templates allow us to write functions that can handle different types. The syntax of function template is as follows:

template <typename T>
T add(T a, T b) {
    return a + b;
}

In this example,addThe function can handle any type that supports addition operations, such asint、floatordouble。

When using it, you can call it like this:

int result = add(3, 4); // use int type
double result2 = add(3.5, 2.7); // use double type

3. Class Template

Category templates allow us to create categories that can be applied to multiple types. The syntax of a category template is as follows:

template <typename T>
class MyClass {
private:
    T data;
public:
    MyClass(T data) : data(data) {}
    T getData() { return data; }
};

In this exampleMyClassCategories can use any type of data asdata。

You can use it like this:

MyClass<int> obj1(5); // int type
MyClass<double> obj2(3.14); // double type

4. Multiple template parameters

Templates can accept multiple parameters, for example:

template <typename T, typename U>
class Pair {
private:
    T first;
    U second;
public:
    Pair(T first, U second) : first(first), second(second) {}
    T getFirst() { return first; }
    U getSecond() { return second; }
};

Such a template class can store two different types of data:

Pair<int, double> pair1(1, 3.14);

5. Template Specialization

Template specialization allows us to specifically define templates of specific types. For example:

template <>
class MyClass<int> {
public:
    MyClass(int data) { /* Specialized behavior */ }
};

This code is specializedMyClassrightintThe behavior of a type that makes it different from other types.

6. Non-Type Template Parameters

Templates can also accept non-type parameters, such as constant values:

template <typename T, int Size>
class Array {
private:
    T data[Size];
public:
    int getSize() const { return Size; }
};

here,SizeIs a non-type parameter indicating the size of the array.

Usage example:

Array<int, 10> arr; // Create an int type array with a size of 10

7. Conclusion

C++ templates are powerful, making code more versatile and reducing duplication. Understanding how to use technology such as function templates, category templates, and template specialization will greatly enhance the flexibility and performance of programming.

C++ Class cross-reference

// ClassA.h
#ifndef CLASSA_H
#define CLASSA_H

// Forward declaration ClassB
class ClassB;

class ClassA {
public:
    ClassA();
    void setB(ClassB* b); // Set the pointer to ClassB
    void showBData(); // Display ClassB data

private:
    ClassB* b; // pointer to ClassB
};

#endif

// ClassB.h
#ifndef CLASSB_H
#define CLASSB_H

// Forward declaration ClassA
class ClassA;

class ClassB {
public:
    ClassB(int data);
    int getData(); // Get data
    void setA(ClassA* a); // Set the pointer to ClassA
    void showAInfo(); // Display ClassA information

private:
    int data;
    ClassA* a; // pointer to ClassA
};

#endif

#include "ClassA.h"
#include "ClassB.h"
#include <iostream>

ClassA::ClassA() : b(nullptr) {}

void ClassA::setB(ClassB* b) {
    this->b = b;
}

void ClassA::showBData() {
    if (b != nullptr) {
        std::cout << "ClassB data: " << b->getData() << std::endl;
    }
}
        

#include "ClassB.h"
#include "ClassA.h"
#include <iostream>

ClassB::ClassB(int data) : data(data), a(nullptr) {}

int ClassB::getData() {
    return data;
}

void ClassB::setA(ClassA* a) {
    this->a = a;
}

void ClassB::showAInfo() {
    if (a != nullptr) {
        a->showBData();
    }
}
        

C++ Friend

In C++,friendKeywords can be used on functions or classes to allow other functions or classes to access private and protected members of the class. Such a design allows external functions or classes to operate without violating the encapsulation principle.

A Friend function is an external function that is granted access to the private and protected members of another class. When declared inside a category, end withfriendJust modify the keywords.

Examples are as follows:

#include <iostream>
using namespace std;

class Box {
private:
    double width;

public:
    Box(double w) : width(w) {}

    //Declare friend function
    friend void showWidth(Box &b);
};

//Friend function definition, can access private members of Box class
void showWidth(Box &b) {
    cout << "Box width: " << b.width << endl;
}

int main() {
    Box box(10.5);
    showWidth(box); // Access private member width
    return 0;
}

In this example,showWidthAlthough the function is an ordinary function outside the Box category, it can still access the private members of the Box category because it is declared as a friend function.width。

The Friend class allows one class to access all members of another class. Such a setup is useful when categories need to work closely together, but should be used with caution to avoid exposing too many internal details.

Examples are as follows:

#include <iostream>
using namespace std;

class Square; // forward declaration

class Rectangle {
private:
    double width, height;

public:
    Rectangle(double w, double h) : width(w), height(h) {}

    // Declare the Square category as friend
    friend class Square;
};

class Square {
public:
    double areaOfRectangle(Rectangle &rect) {
        return rect.width * rect.height;
    }
};

int main() {
    Rectangle rect(5.0, 3.0);
    Square square;
    cout << "Area of rectangle: " << square.areaOfRectangle(rect) << endl;
    return 0;
}

In this example, the Square class is declared as a friend class of the Rectangle class, so the member functions in the Square class can directly access the private members of the Rectangle class.widthandheight。

• Data encapsulation: Allows certain functions or categories to perform direct operations without destroying the internal encapsulation of the category. It is suitable for inter-category communication that requires high coupling.
• Operator overloading: You can use friend functions to overload operators to enable access to private or protected members inside a class.
• Specific access needs: Allows limited access to external functions or classes without breaking encapsulation, suitable for specific designs in library development.

Be careful when using friend functions and friend classes in C++. Excessive use will destroy the encapsulation of the class. Therefore, the friend keyword is usually only used when designing classes or functions that require close cooperation.

C++ read file line by line

Overview

In C++, the most standard and recommended way to read a text file line-by-line is to use the file streaming class from the standard librarystd::ifstreamCooperatestd::getlinefunction.

std::ifstreamResponsible for opening and managing file streaming, andstd::getlineis responsible for reading an entire line of text from the stream (until a newline character is encountered\n) and store it instd::stringin the object.

Code example

The following is a complete C++ example that demonstrates how to read line by line a file namedexample.txtfile contents.

#include <iostream>
#include <fstream> // Include file streaming library
#include <string> // Include string category

using namespace std;

void read_file_line_by_line(const string& filename)
{
    // 1. Create std::ifstream object
    //Try to open the specified file.
    ifstream input_file(filename);
    
    // 2. Check whether the file is opened successfully
    if (!input_file.is_open())
    {
        cerr << "Error: Unable to open file" << filename << endl;
        return;
    }
    
    string line;
    int line_number = 1;
    
    // 3. Use std::getline function to read line by line
    // The loop condition (getline(stream, string)) returns true each time the read is successful.
    // When end of file (EOF) is reached or an error occurs, return false and the loop ends.
    while (getline(input_file, line))
    {
        cout << "Line " << line_number << ": " << line << endl;
        line_number++;
    }
    
    // 4. Close file streaming
    // When the input_file object exceeds its scope, the destructor will automatically close the file.
    // But calling close() explicitly is also acceptable.
    input_file.close();
    
    cout << "File reading completed." << endl;
}

int main()
{
    read_file_line_by_line("example.txt");
    return 0;
}

Key concepts

• std::ifstream: inherited fromstd::istream, specifically used to process file input streams. at build time (or usingopen()function) will try to open the file.
• input_file.is_open(): This is standard error checking. If the file does not exist or the program does not have read permission, this function will returnfalse。
• std::getline(stream, string)：
  • It starts from the first parameter (input stream, e.g.input_file) to read data.
  • It stores all characters read (including whitespace) into the second parameter (std::stringobjects, such asline)middle.
  • it will read andthrow awaynewline character\n,thereforelineThe variable does not contain newline characters.
  • When the read is successful, it will return the input stream object itself, which will be converted totrue。
• Automatic resource management:In C++, whenstd::ifstreamobject (input_file) is destroyed, it will automatically call its destructor to close the underlying file control code. This is a reliable resource management method (RAII, Resource Acquisition Is Initialization).

C++ Write file line by line and align fields

Overview

In C++, converting a data structure (e.g. fromstd::vector<std::string>To write the file line by line (tokens read in) and ensure that each column (token) maintains a fixed width alignment in the line, you need to use **std::ofstream** Use **I/O Stream Manipulators** to control the output format.

The main format control tools are:

• std::setw(width): Set the minimum width of the next output field.
• std::left / std::right: Set the alignment of text in the column (left or right).

Code example

The following is a C++ example of how to convert a two-dimensional data structure containing multiple fields (usingstd::vector<std::vector<std::string>>Simulate) and write the file, making sure that each token field has a fixed width of 15 characters and is left aligned.

#include <iostream>
#include <fstream> // Include file output stream
#include <string>
#include <vector>
#include <iomanip> // Include I/O stream manipulator (setw, left/right)

using namespace std;

// Simulate a two-dimensional data structure, each row contains multiple fields (tokens)
using TableData = vector<vector<string>>;

void write_aligned_data(const string& filename, const TableData& data, int column_width)
{
    // 1. Create std::ofstream object and open the file
    ofstream output_file(filename);
    
    // 2. Check whether the file is opened successfully
    if (!output_file.is_open())
    {
        cerr << "Error: Unable to open file " << filename << " for writing." << endl;
        return;
    }

    //Set global alignment: left-aligned (L-Justified)
    output_file << left;
    
    // 3. Write data line by line
    for (const auto& row : data)
    {
        // 4. Write token by token and set the width
        for (const auto& token : row)
        {
            // setw(width) only affects the next output item immediately following it
            output_file << setw(column_width) << token;
        }
        
        // 5. Insert a newline character after the end of the line to start a new line
        output_file << "\n";
    }
    
    // 6. Close file streaming
    output_file.close();
    
    cout << "Data has been successfully written to the file: " << filename << endl;
}

int main()
{
    //Sample data: contains four rows, each row has three fields
    TableData data = {
        {"Name", "Item", "Price"},
        {"Alice", "Book", "19.99"},
        {"BobJohnson", "PenSet", "4.50"},
        {"Charlie", "Notebook", "800.75"}
    };
    
    const int COLUMN_WIDTH = 15; // Define the width of each column
    
    write_aligned_data("output_aligned.txt", data, COLUMN_WIDTH);
    
    return 0;
}

Analysis of key technologies

• std::ofstream: File output streaming category. Used to write data into files.
• #include <iomanip>: This header file must be included to use the I/O stream manipulator.
• output_file << left;：
  • Set the alignment mode to left. Once this setting is applied to a stream, it will remain in effect untilrightorinternalcover.
• output_file << setw(column_width) << token;：
  • setw(width): Set the minimum width of the data to be output next. If the length of the output string is less than the set width, it will be padded with blank characters.
  • important: setwOnly valid for the next output operation immediately following it, so it must be called once before each output token.
• Character width:Note that the C++ standard stream's width calculations for multi-byte character sets (such as UTF-8) may be inaccurate when dealing with non-ASCII characters (such as Chinese). When dealing with internationalized text, more complex libraries may be needed to correctly calculate visual width. The above example works for ASCII or single-width characters.

.NET C++ reads HTML table

Overview

In .NET C++ (or more commonly C++/CLI) projects, reading and parsing HTML tables in files usually requires the use of an external HTML parsing library, because the standard .NET framework and C++ themselves do not have built-in powerful HTML DOM parsing capabilities. A common and efficient option is to use a library in C# or another .NET language and then reference it in C++/CLI.

For C++/CLI projects, the most practical and recommended approach is to useHtml Agility Pack(HAP), a very popular .NET HTML parser. Although HAP is written in C#, it can be used seamlessly as a reference in any .NET language, including C++/CLI.

Steps to use Html Agility Pack

1. Create a C++/CLI project:Make sure your project is a C++/CLI project (for example: CLR Console Application or Windows Forms App).
2. Get Html Agility Pack:
   • Search and install through Visual Studio's NuGet package managerHtmlAgilityPack。
3. Program code implementation (C++/CLI):

Code example

The following is a C++/CLI code snippet that demonstrates how to read a local HTML file and parse the first table in it (<table>) content:

#using <System.dll>
#using <System.Xml.dll>
#using <HtmlAgilityPack.dll> // Make sure the reference is included

using namespace System;
using namespace System::IO;
using namespace HtmlAgilityPack;

void ParseHtmlTable(String^ filePath)
{
    // Check if the file exists
    if (!File::Exists(filePath))
    {
        Console::WriteLine("Error: File does not exist.");
        return;
    }

    // 1. Load HTML file
    HtmlDocument^ doc = gcnew HtmlDocument();
    try
    {
        //Load HTML from file
        doc->Load(filePath);
    }
    catch (Exception^ ex)
    {
        Console::WriteLine("An error occurred while loading the file: " + ex->Message);
        return;
    }

    // 2. Select the first <table> node
    HtmlNode^ table = doc->DocumentNode->SelectSingleNode("//table");

    if (table != nullptr)
    {
        Console::WriteLine("Find the HTML table and start parsing...");

        // 3. Select all <tr> (table column) nodes
        HtmlNodeCollection^ rows = table->SelectNodes(".//tr");

        if (rows != nullptr)
        {
            for each (HtmlNode^ row in rows)
            {
                // 4. Select the <td> or <th> (cell) node
                // Use the | operator to select <td> or <th>
                HtmlNodeCollection^ cells = row->SelectNodes("td | th");

                if (cells != nullptr)
                {
                    String^ rowData = "";
                    for each (HtmlNode^ cell in cells)
                    {
                        // Get the internal text of the cell and remove the leading and trailing blanks
                        rowData += cell->InnerText->Trim() + "\t";
                    }
                    Console::WriteLine(rowData);
                }
            }
        }
        else
        {
            Console::WriteLine("<tr> tag not found in table.");
        }
    }
    else
    {
        Console::WriteLine("<table> tag not found in file.");
    }
}

int main(array<String^>^ args)
{
    // Please replace "your_html_file.html" with your actual file path
    String^ htmlFilePath = "C:\\path\\to\\your_html_file.html";
    ParseHtmlTable(htmlFilePath);
    return 0;
}

Explanation of key technical points

• C++/CLI：Allows C++ code to interoperate with the .NET Framework (CLR), which is the basis for using C# libraries such as HAP.
• Html Agility Pack：Provide powerfulXPathsupport.
  • doc->Load(filePath): Read local HTML files into HAP's DOM structure.
  • SelectSingleNode("//table"): Use XPath expression to select the first one in the file<table>element.
  • SelectNodes(".//tr"): Select the current node (<table>) all under<tr>element.
  • SelectNodes("td | th"):Select<tr>All under the node<td>or<th>element.
  • cell->InnerText: Get the plain text of the cell content, and HTML tags will be automatically removed.
• gcnewand^：This is the syntax used by C++/CLI to create and manage .NET objects (managed by the CLR garbage collection mechanism).

.NET C++ HTTP API Server

Using the HttpListener class

HttpListenerClass allows you to create a simple, self-hosted HTTP server in a C++/CLI application.

Project requirements and settings

Create a C++/CLI project in Visual Studio (e.g. CLR Console Application) and make sure it is referencedSystem.dll。

Code example

The following is usedHttpListenerC++/CLI code to create a simple API server and handle GET requests.

#using <System.dll>

using namespace System;
using namespace System::Net;
using namespace System::Threading::Tasks;
using namespace System::Text;

// Function to handle incoming HTTP requests
void HandleRequest(HttpListenerContext^ context)
{
    HttpListenerRequest^ request = context->Request;
    HttpListenerResponse^ response = context->Response;

    //Default response settings
    response->ContentType = "application/json; charset=utf-8";
    String^ responseString = "";
    response->StatusCode = 200; // Default is OK

    // Check the requested path and method
    String^ url = request->RawUrl->ToLower();
    
    if (url == "/api/status" && request->HttpMethod == "GET")
    {
        // Handle GET /api/status request
        responseString = "{\"status\": \"Server Running\", \"language\": \"C++/CLI\"}";
    }
    else
    {
        // Handle other undefined requests
        response->StatusCode = 404; // Not Found
        responseString = "{\"error\": \"404 Not Found\", \"path\": \"" + url + "\"}";
    }

    // Convert string response to bytes and write to output stream
    array<Byte>^ buffer = Encoding::UTF8->GetBytes(responseString);
    response->ContentLength64 = buffer->Length;
    
    try
    {
        response->OutputStream->Write(buffer, 0, buffer->Length);
        response->OutputStream->Close();
    }
    catch (Exception^)
    {
        // Ignore write errors, usually client disconnects
    }
}

// Function to start HttpListener
void StartListener(String^ prefix)
{
    HttpListener^ listener = gcnew HttpListener();
    
    try
    {
        listener->Prefixes->Add(prefix);
        listener->Start();
        Console::WriteLine("C++/CLI API Server started, listening: {0}", prefix);
    }
    catch (Exception^ ex)
    {
        Console::WriteLine("An error occurred while starting HttpListener. Please confirm whether the permission or address is occupied.");
        Console::WriteLine("Error message: {0}", ex->Message);
        return;
    }

    // Asynchronous loop, continue to receive requests
    while (listener->IsListening)
    {
        try
        {
            // Wait for the next request synchronously
            HttpListenerContext^ context = listener->GetContext();
            
            // Use Task to process requests asynchronously to avoid blocking listening loops
            Task::Factory->StartNew(gcnew Action<HttpListenerContext^>(&HandleRequest), context);
        }
        catch (Exception^)
        {
            // An exception will be thrown when the listener is stopped
            break;
        }
    }
    
    if (listener->IsListening)
    {
        listener->Close();
    }
}

int main(array<String^>^ args)
{
    //Set the URL prefix for monitoring
    // NOTE: On Windows, administrator rights or using netsh to register the URL namespace may be required.
    String^ listeningPrefix = "http://localhost:8080/";
    StartListener(listeningPrefix);
    
    Console::WriteLine("Press any key to end the server...");
    Console::ReadKey(true);
    
    return 0;
}

Note on execution permissions

On Windows systems, if the code is not run with administrator privileges,HttpListenerAn access-denied exception may be thrown when listening on a specific port. You can register a URL namespace to allow non-administrator accounts to run the server with the following command:

netsh http add urlacl url=http://+:8080/ user=Everyone

inhttp://+:8080/should be replaced with the actual prefix used in your code.

ASP.NET Core C# HTTP API Server

Overview

In ASP.NET Core, the standard project type for creating an HTTP API server is Web API. Latest .NET version (.NET 6 and above) recommendedMinimal APIsto quickly and lightweightly create API endpoints.

Minimal APIs combine startup logic (Startup) and routing definitions (Routing) into a single file using minimal files and lines of codeProgram.csmiddle.

Project setup (using CLI)

To create a new Minimal API project, you can use the .NET CLI (Command Line Interface):

dotnet new web -n MinimalApiServer
cd MinimalApiServer

Program code example: Program.cs

The following is the Minimal API server core fileProgram.csfull content. It defines server startup, middleware, and two API endpoints (one GET, one POST).

// 1. Create a WebApplication instance (replacing the traditional Startup category)
var builder = WebApplication.CreateBuilder(args);

// 2. Register service (Service Configuration)
//Here you can add database connection, verification services, etc.
builder.Services.AddEndpointsApiExplorer();
builder.Services.AddSwaggerGen(); // Enable Swagger/OpenAPI support

var app = builder.Build();

// 3. Configure HTTP request pipeline (Middleware Configuration)

// In the development environment, enable Swagger UI for testing
if (app.Environment.IsDevelopment())
{
    app.UseSwagger();
    app.UseSwaggerUI();
}

// Enable HTTPS redirection (recommended for production environments)
// app.UseHttpsRedirection();

// 4. Define API endpoint (Endpoint Definition)
// Entity class used for POST requests
public record Product(int Id, string Name, decimal Price);

// GET request endpoint: /api/hello
app.MapGet("/api/hello", () =>
{
    return Results.Ok(new { message = "Hello from ASP.NET Core Minimal API!", timestamp = DateTime.Now });
});

// GET request endpoint: /api/products/{id}
app.MapGet("/api/products/{id}", (int id) =>
{
    // Assume querying products from the database
    if (id == 1)
    {
        return Results.Ok(new Product(1, "laptop", 1200.00m));
    }
    return Results.NotFound(new { error = $"Product ID {id} not found" });
});

// POST request endpoint: /api/products
app.MapPost("/api/products", (Product product) =>
{
    // ASP.NET Core will automatically deserialize the JSON request body into a Product entity
    Console.WriteLine($"Received new product: {product.Name} (ID: {product.Id})");
    
    //In actual application, the database addition operation will be performed here.
    //Return 201 Created status code
    return Results.Created($"/api/products/{product.Id}", product);
});

// 5. Start the application and start listening for HTTP requests
app.Run();

Execution and testing

1. Run the server:Execute in project directorydotnet run。
2. Test endpoint:Servers are usually defaulted tohttp://localhost:5000orhttps://localhost:7000run.
   • Access using a browser or tool such as Postman/api/hello (GET)。
   • access/swaggerPath to view Swagger UI for testing all defined API endpoints.

Key technical points

• WebApplication.CreateBuilder(args)：This is the starting point for ASP.NET Core launch. It creates Hosts and Builders for configuring the application's services and middleware.
• app.Services.AddSwaggerGen()：Register for services to supportSwagger/OpenAPIfiles, greatly simplifying API testing and file generation.
• Middleware: app.UseSwagger()、app.UseHttpsRedirection()Etc., these components constitute the request processing pipeline in sequence. Every request flows through these middlewares.
• app.MapGet / app.MapPost：This is the core method used in Minimal APIs to define routing and processing logic. They bind HTTP verbs to specific URL paths and Lambda handler functions.
• Automatic model binding:existapp.MapPostIn the example, when the client sends JSON data, ASP.NET Core will automatically convert (deserialize) it into C#ProductRecord.

vcpkg C++ package manager

1. Download vcpkg source code

vcpkg is not installed through the installation file, but directly copies the source code from GitHub to local. It is recommended to place it in the root directory of the disk (such asC:\vcpkg) to avoid paths that are too long or contain spaces. Please open the terminal (CMD or PowerShell) and execute:

git clone https://github.com/microsoft/vcpkg.git
cd vcpkg

2. Perform initial compilation (Bootstrap)

After downloading, you need to execute the script to compile the executable file of vcpkg, which will generatevcpkg.exe：

• Windows: implement.\bootstrap-vcpkg.bat
• macOS / Linux: implement./bootstrap-vcpkg.sh

3. Integrate into development environment

This step allows development tools (such as Visual Studio) to automatically recognize the packages installed by vcpkg:

• Visual Studio (MSBuild) integration:

  implement.\vcpkg integrate install. After completion, the new project you create in VS can be directly#includeKit, no need to manually set the path.

• CMake integration:

  When executing CMake configuration, add the following parameters to link to the vcpkg toolchain file:

  -DCMAKE_TOOLCHAIN_FILE=[vcpkg path]/scripts/buildsystems/vcpkg.cmake

Common instruction comparison table

Set environment variables (recommended)

In order to facilitate the use of vcpkg in any path, it is recommended to add the vcpkg folder path to the systemPATHin environment variables. In addition, if you want to install the 64-bit package by default, you can add an environment variableVCPKG_DEFAULT_TRIPLETand set its value tox64-windows。

Things to note

Before using vcpkg on Windows, make sure it is installedVisual Studio 2015 or newer, and check the "Desktop development using C++" workload (including English language pack), otherwise errors may occur when compiling the package.

.NET MessageBox

1. Basic syntax and common overloading

In C++/CLI,MessageBox::ShowIt is a static method used to pop up a dialog box. The most complete call method includes: message content, title, button type, icon and default button.

using namespace System::Windows::Forms;

//The simplest display
MessageBox::Show("Operation completed");

//Full parameter version
DialogResult result = MessageBox::Show(
    "Are you sure you want to delete this file?", // Message content (Text)
    "Confirm deletion", // window title (Caption)
    MessageBoxButtons::YesNo, // Button combination
    MessageBoxIcon::Warning, // Prompt icon
    MessageBoxDefaultButton::Button2 // Default focus is on the second button (No)
);

2. Process the return value (DialogResult)

Must be checked when a button contains multiple options such as Yes/No or OK/CancelDialogResultto determine the subsequent logic.

if (result == DialogResult::Yes) {
    //Execute deletion logic
} else {
    //Cancel operation
}

3. Common enumeration value reference

4. Advanced technique: combine std::string or float

becauseMessageBox::ShowReceived is .NETSystem::String^, if you have native C++ data, you need to convert it.

#include <string>
#include <msclr/marshal_cppstd.h>

// Display std::string in MessageBox
std::string cpp_str = "Core Error";
MessageBox::Show(msclr::interop::marshal_as<System::String^>(cpp_str));

//Format the float and display it
float dist = 1.23456f;
MessageBox::Show("The distance is: " + dist.ToString("F2")); // Display "The distance is: 1.23"

Common errors and solutions

• Ambiguous call: Sometimes the compiler will confuse the overloaded version if only two strings are passed in. It is recommended to explicitly addMessageBoxButtons::OK。
• thread problem: MessageBox is forced to respond (Modal) by default and will block the current UI thread. Make sure to execute thread calls in the main UI to avoid screen freezes.
• Namespace conflict: If used simultaneouslyusing namespace System;andusing namespace System::Windows::Forms;, possibly with otherMessageBoxDefinition conflict, it is recommended to use the full nameSystem::Windows::Forms::MessageBox::Show(...)。

.NET custom input dialog box

1. Why does MessageBox not support Edit Box?

System::Windows::Forms::MessageBoxis designed to display "notifications" or "warnings" and obtain user decisions via standard buttons. It does not have the function of dynamically inserting control items. If you need to enter text, the standard approach is to inheritSystem::Windows::Forms::FormCreate a custom dialog box.

2. Implement custom InputBox

This is a simple, reusable category that contains labels, text boxes, and OK buttons.

using namespace System;
using namespace System::Windows::Forms;
using namespace System::Drawing;

public ref class CustomInputBox : public Form {
public:
    TextBox^ txtInput;
    Button^ btnOK;
    Label^ lblPrompt;

    CustomInputBox(String^ title, String^ prompt) {
        //Set basic properties of the form
        this->Text = title;
        this->Size = System::Drawing::Size(300, 150);
        this->FormBorderStyle = System::Windows::Forms::FormBorderStyle::FixedDialog;
        this->StartPosition = FormStartPosition::CenterParent;
        this->MaximizeBox = false;
        this->MinimizeBox = false;

        // Prompt text
        lblPrompt = gcnew Label();
        lblPrompt->Text = prompt;
        lblPrompt->Location = Point(15, 15);
        lblPrompt->Size = System::Drawing::Size(250, 20);

        // input box
        txtInput = gcnew TextBox();
        txtInput->Location = Point(15, 40);
        txtInput->Size = System::Drawing::Size(250, 20);

        // OK button
        btnOK = gcnew Button();
        btnOK->Text = "OK";
        btnOK->DialogResult = System::Windows::Forms::DialogResult::OK;
        btnOK->Location = Point(190, 75);
        
        // Add controls
        this->Controls->Add(lblPrompt);
        this->Controls->Add(txtInput);
        this->Controls->Add(btnOK);
        
        //Set the default button (pressing Enter is equivalent to clicking OK)
        this->AcceptButton = btnOK;
    }
};

3. How to call in the main program

useShowDialog()The method opens the window in forced response mode and checks whether the return value isOK。

void OnButtonClick() {
    CustomInputBox^ ibox = gcnew CustomInputBox("System input", "Please enter the file name:");
    
    if (ibox->ShowDialog() == System::Windows::Forms::DialogResult::OK) {
        String^ userInput = ibox->txtInput->Text;
        
        if (!String::IsNullOrWhiteSpace(userInput)) {
            // Process the string entered by the user
            MessageBox::Show("You entered: " + userInput);
        }
    }
}

4. Plan comparison

Common pitfalls

• Resource release: In C++/CLI, although there is GC, for the Form class it is recommended to usedelete ibox;Or make sure it releases resources correctly after the scope ends.
• Thread safety: All UI control operations (including displaying custom dialog boxes) must be inSTA (Single Thread Apartment)Execute in the execution thread, usually the main UI execution thread.
• Forced response (Modal):useShowDialog()rather thanShow(), this ensures that the user must close the input window before operating the main window, and can directly obtainDialogResult。

PictureBox red cross error

Phenomenon description

PictureBox displays a red cross, indicating that theOnPaintAn exception occurred during execution, However, .NET swallows this exception by default and will not show it in the general log or interruption points, making it difficult to debug.

Common causes

Error message comparison

Debug steps

Since the red cross problem may not occur often, it is recommended to proceed in order:

1. Intercepting OnPaint exceptions: Inherit PictureBox, overrideOnPaint, Use try/catch to catch exceptions and output them toDebug.WriteLine, execute after capturethis.Image = nullAvoid persistent red crosses.
2. Confirm exception message: Observe the message in the Visual Studio Output window, Check the table of error messages above to find the root cause.
3. Tracking Bitmap lifecycle: Each time assigned to Dispose Bitmap, recordEnvironment.StackTrace, find out who released the resource at what point in time.
4. Prevent Dispose of a Bitmap that is still in use: Check before Dispose Whether the PictureBox still holds the Bitmap, if so, firstImageSet to null and then Dispose.

Safe picture changing mode

All image changing operations are performed through this mode to ensure thread safety and correct Dispose order:

// C++/CLI
void SafeUpdateImage(Bitmap^ newBmp) {
    if (pictureBox1->InvokeRequired) {
        pictureBox1->Invoke(gcnew Action<Bitmap^>(
            this, &YourClass::SafeUpdateImage), newBmp);
        return;
    }
    Bitmap^ old = safe_cast<Bitmap^>(pictureBox1->Image);
    pictureBox1->Image = newBmp; // Change to the new picture first
    if (old != nullptr) old->Dispose(); // Dispose the old picture again
}

The correct way to create Bitmap from Stream

// Error: bitmap invalid after stream is closed
using (var stream = new FileStream(path, FileMode.Open)) {
    pictureBox1.Image = new Bitmap(stream);
}

// Correct: deep copy, does not rely on stream
Bitmap loaded;
using (var stream = new FileStream(path, FileMode.Open))
using (var tmp = new Bitmap(stream)) {
    loaded = new Bitmap(tmp);
}
pictureBox1.Image = loaded;

Troubleshooting checklist

Peripheral control program development

definition

Peripheral control programs are applications used to communicate and operate external hardware devices (such as printers, scanners, motors, PLCs, sensors, etc.) connected to the computer or host.

Common communication interfaces

• USB
• Serial（RS-232 / RS-485）
• Bluetooth
• Wi-Fi / TCP Socket
• I2C/SPI (mostly used in embedded)

Common development languages ​​and environments

Example 1: Using Python to control the serial port

import serial

ser = serial.Serial('COM3', 9600, timeout=1)
ser.write(b'ON\n') # Send control commands
response = ser.readline()
print("Device response:", response.decode())
ser.close()

Example 2: C# controls COM Port device

SerialPort port = new SerialPort("COM4", 9600);
port.Open();
port.WriteLine("MOVE 100");
string response = port.ReadLine();
Console.WriteLine("Response: " + response);
port.Close();

Example 3: Control USB HID device

• Availablelibusb（C）、pyusb（Python）、HidSharp(C#) and other function libraries
• Need to master USB Protocol and device Descriptor structure

Example 4: TCP communication control peripherals

import socket

s = socket.socket()
s.connect(('192.168.1.100', 5000))
s.sendall(b'START\n')
data = s.recv(1024)
print("Response:", data.decode())
s.close()

Application cases

• PLC automation equipment control
• 3D printer/laser engraving machine
• Production line scanner and sensor collection
• POS system peripheral integration (invoice machine, barcode machine)

Things to note

• Need to understand the protocol and command format of the device
• Some devices require drivers or SDKs to communicate
• Communication delays, synchronization processing and error detection must be properly designed

in conclusion

Peripheral control program development requires selecting the appropriate language and function library based on the device's interface and protocol, and combining serial, USB, network and other communication methods, which can be widely used in various automation and industrial control fields.

Barcode Reader program development

Common platform support

• Android / iOS
• Windows desktop application
• Web front-end (HTML5 + JavaScript)
• Embedded devices (such as Raspberry Pi)

Common barcode types

• 1D barcode: EAN-13, Code 39, Code 128, UPC
• 2D barcode: QR Code, Data Matrix, PDF417

Common development kits and tools

• ZXing（Java、Android、C++、JavaScript）
• ZBar（C / Python / Linux / Android）
• Dynamsoft Barcode Reader SDK(Commercial use, supports multiple languages)
• ML Kit Barcode Scanning（Android / Firebase）
• jsQR、QuaggaJS(Web application)

Android development example (using ZXing)

// build.gradle
implementation 'com.journeyapps:zxing-android-embedded:4.3.0'

//Java call scan screen
IntentIntegrator integrator = new IntentIntegrator(this);
integrator.setPrompt("Please scan the barcode");
integrator.setBeepEnabled(true);
integrator.setOrientationLocked(false);
integrator.initiateScan();

//onActivityResult receives the result
@Override
protected void onActivityResult(int requestCode, int resultCode, Intent data) {
    IntentResult result = IntentIntegrator.parseActivityResult(requestCode, resultCode, data);
    if(result != null) {
        if(result.getContents() != null) {
            String barcode = result.getContents();
            Log.d("Barcode content", barcode);
        }
    }
}

Web development examples (using QuaggaJS)

<script src="https://unpkg.com/[email protected]/dist/quagga.min.js"></script>

<script>
Quagga.init({
    inputStream: {
        name: "Live",
        type: "LiveStream",
        target: document.querySelector('#scanner')
    },
    decoder: {
        readers: ["code_128_reader", "ean_reader", "upc_reader"]
    }
}, function(err) {
    if (!err) {
        Quagga.start();
    }
});

Quagga.onDetected(function(result) {
    console.log("Barcode content: ", result.codeResult.code);
});
</script>

Platform vs. Suite Comparison

in conclusion

To develop Barcode Reader, you can choose the appropriate open source package according to the platform. ZXing is the most widely supported choice. QuaggaJS is available for the Web. If you need commercial-level support, you can consider Dynamsoft Barcode SDK.

Barcode with control characters

Basic principles

Barcode readers are essentially devices that simulate keyboard input. When a barcode is scanned, it "outputs" the content in the barcode into a character stream, just like keyboard input.

Therefore, the barcodeCan contain control code (Control Code),butRequires barcode reader and barcode format supportJust fine.

Common control code types in barcodes

• Ctrl+A ~ Ctrl+Z（ASCII 0x01 ~ 0x1A）
• Enter（CR，ASCII 0x0D）
• Tab（ASCII 0x09）
• ESC（ASCII 0x1B）

Condition 1: The barcode format must support control codes

For example, the following format can encode control codes:

• Code 128: Supports full ASCII 128 characters (including Ctrl characters)
• QR Code: Supports all Unicode/Binary data

Condition 2: The barcode encoding tool must support embedding control codes

Some barcode generators allow embedding special characters, such as:

• \x0Dmeans Enter
• \x03means Ctrl-C
• ^C、[CTRL+C]Special syntax depends on the tool

Condition 3: The barcode scanner must support output control codes

Most professional barcode printers (e.g. Zebra, Honeywell) are preset at the factoryDisable control code output, needs to be enabled through the "Settings barcode" provided by the scanner:

• Enable control character transmission
• Enable function key emulation
• Interpret Code 128 FNC1 as ASCII

Condition 4: The operating system and application must receive it correctly

For example in a Windows application:

• If barcode output Ctrl+O, the application needs to supportkeydownandCtrlMatching shortcut key processing.
• If the barcode output is a special key combination (such as Alt-C), most applications need to process it from the low-level input API.

Example: Encoding Ctrl-C barcode content

Encoding ASCII 3 with Code128:

Input:\x03Hello World

After the barcode is scanned, Ctrl+C will be triggered and "Hello World" will be output.

in conclusion

• Barcodes can contain control codes (such as Ctrl, Alt), but they must meet the following conditions:
• Barcode formats (such as Code128) supported
• Generator correctly encoded
• Scanner settings correctly
• The application can receive correctly

Android development

Android development platform

Platform introduction

development tools

Application life cycle

Advantages

learning resources

Conclusion

Chromebook using Android Studio

Installation steps

1. turn onSettings → Developer → Linux Development Environment (Crostini), allocate at least 20GB of disk space and 8GB of memory.
2. Update kit:

   sudo apt update && sudo apt upgrade -y
   sudo apt install -y wget curl unzip zip git ca-certificates

3. Install JDK (version 17 recommended):

   sudo apt install -y openjdk-17-jdk

4. Download and install Android Studio:

   sudo dpkg -i ~/Downloads/android-studio-*.deb || sudo apt -f install -y
   android-studio

   or decompress.tar.gz：

   tar -xzf ~/Downloads/android-studio-*.tar.gz -C ~
   ~/android-studio/bin/studio.sh

5. Follow the installation wizard to complete the SDK and tool settings.

Test with a physical device

1. On ChromeOS:Settings → Developer → Linux → USB Device, check the box to share the phone with Linux.
2. Turn on the phoneDeveloper Options → USB Debugging, allowing RSA fingerprinting when connected.
3. Linux environment confirmation device:

   sudo apt install -y android-sdk-platform-tools
   adb kill-server
   adb start-server
   adb devices

Performance and emulator recommendations

• Most Chromebook emulators are underpowered, so it is recommended to use a physical phone first.
• If you must use an emulator, choose the x86_64 version and lower the resolution and RAM configuration.
• Avoid running too many applications at the same time and ensure that the Crostini container has sufficient resources.

alternatives

• VS Code + Android SDK: Relatively lightweight, suitable for Flutter / React Native development.
• Cloud development environment: Such as GitHub Codespaces and Gitpod, which reduce the burden on the local machine.

Chromebook uses VS Code + Android SDK

Install VS Code

1. Turn on ChromeOSSettings → Developer → Linux Development Environment (Crostini)。
2. Download the VS Code Linux version (.deb file) and install it in the Linux container:

   sudo dpkg -i ~/Downloads/code_*.deb || sudo apt -f install -y

3. Start VS Code and install commonly used extensions (such asKotlin、Flutter、Java Extension Pack）。

Install Android SDK and tools

1. Install necessary packages:

   sudo apt update && sudo apt install -y openjdk-17-jdk unzip git

2. Download Android SDK Command Line Tools:

   wget https://dl.google.com/android/repository/commandlinetools-linux-11076708_latest.zip
   unzip commandlinetools-linux-*_latest.zip -d ~/android-sdk

3. Set environment variables (can be added~/.bashrcor~/.zshrc）：

   export ANDROID_SDK_ROOT=$HOME/android-sdk
   export PATH=$ANDROID_SDK_ROOT/cmdline-tools/bin:$ANDROID_SDK_ROOT/platform-tools:$PATH

4. Use sdkmanager to install platform tools and necessary packages:

   sdkmanager "platform-tools" "platforms;android-34" "build-tools;34.0.0"

Connect to physical mobile phone test

1. On ChromeOS:Settings → Developer → Linux → USB Device, check phone sharing.
2. Turn on the phoneDeveloper Options → USB Debugging。
3. Linux container confirmation device:

   adb devices

   SeedeviceThe test is ready to be deployed.

Develop in VS Code

• Create a Kotlin/Java project, or useFlutter / React Nativeframe.
• Install the corresponding extension package in VS Code to support syntax highlighting and debugging.
• byadb installor framework CLI (such asflutter run) Deploy the App to the mobile phone.

Suitable for the situation

• Chromebooks have limited hardware resources and are not suitable for running Android Studio.
• A lightweight IDE is required and tested on a physical phone.
• Cross-platform development (Flutter, React Native) is more efficient.

Android App Example

Create new project

Create a new project in Android Studio, select the "Empty Activity" template, and set the project name and other basic information.

Writing interface (activity_main.xml)

existres/layout/activity_main.xmlA simple user interface designed in the file, for example, containing a button and a text display area:

<LinearLayout xmlns:android="http://schemas.android.com/apk/res/android"
            android:layout_width="match_parent"
            android:layout_height="match_parent"
            android:orientation="vertical"
            android:gravity="center">
        
            <Button
                android:id="@+id/button"
                android:layout_width="wrap_content"
                android:layout_height="wrap_content"
                android:text="Click me" />
        
            <TextView
                android:id="@+id/textView"
                android:layout_width="wrap_content"
                android:layout_height="wrap_content"
                android:text="Hello, World!"
                android:layout_marginTop="20dp" />
        
        </LinearLayout>

Write program logic (MainActivity.java)

existMainActivity.java, set the click event of the button to change the content of the text display area:

package com.example.simpleapp;

        import android.os.Bundle;
        import android.view.View;
        import android.widget.Button;
        import android.widget.TextView;
        import androidx.appcompat.app.AppCompatActivity;

        public class MainActivity extends AppCompatActivity {
            @Override
            protected void onCreate(Bundle savedInstanceState) {
                super.onCreate(savedInstanceState);
                setContentView(R.layout.activity_main);

                Button button = findViewById(R.id.button);
                TextView textView = findViewById(R.id.textView);

                button.setOnClickListener(new View.OnClickListener() {
                    @Override
                    public void onClick(View v) {
                        textView.setText("You clicked the button!");
                    }
                });
            }
        }

Execute application

Click the Run button in Android Studio to test your app on the emulator or a connected physical device. After clicking the button, the text will change to "You clicked the button!".

Android gets current GPS location

Permission settings

existAndroidManifest.xmlAdd the following permissions:

<uses-permission android:name="android.permission.ACCESS_FINE_LOCATION" />
<uses-permission android:name="android.permission.ACCESS_COARSE_LOCATION" />

Checking and requesting permissions (Android 6.0 and above)

if (ContextCompat.checkSelfPermission(this, Manifest.permission.ACCESS_FINE_LOCATION)
    != PackageManager.PERMISSION_GRANTED) {
    ActivityCompat.requestPermissions(this,
        new String[]{Manifest.permission.ACCESS_FINE_LOCATION}, 1);
}

Get LocationManager

LocationManager locationManager = (LocationManager) getSystemService(Context.LOCATION_SERVICE);

Get current location

LocationListener locationListener = new LocationListener() {
    @Override
    public void onLocationChanged(@NonNull Location location) {
        double latitude = location.getLatitude();
        double longitude = location.getLongitude();
        Log.d("GPS", "Latitude: " + latitude + ", Longitude: " + longitude);
    }
};

if (ActivityCompat.checkSelfPermission(this, Manifest.permission.ACCESS_FINE_LOCATION)
    == PackageManager.PERMISSION_GRANTED) {
    locationManager.requestLocationUpdates(LocationManager.GPS_PROVIDER,
        1000, // millisecond interval
        1, // Minimum distance (meters)
        locationListener);
}

Use FusedLocationProviderClient (recommended)

FusedLocationProviderClient fusedLocationClient = LocationServices.getFusedLocationProviderClient(this);

if (ActivityCompat.checkSelfPermission(this, Manifest.permission.ACCESS_FINE_LOCATION)
    == PackageManager.PERMISSION_GRANTED) {
    fusedLocationClient.getLastLocation()
        .addOnSuccessListener(this, location -> {
            if (location != null) {
                double lat = location.getLatitude();
                double lng = location.getLongitude();
                Log.d("GPS", "Lat: " + lat + ", Lng: " + lng);
            }
        });
}

Sound Control Android App

Basic concepts

To implement a voice assistant function like Siri or Hey Google, you need to combine the following components:

• Speech Recognition
• Voice wake-up (Voice Trigger, optional)
• Text To Speech (TTS)
• Foreground Service

Add necessary permissions

<uses-permission android:name="android.permission.RECORD_AUDIO"/>
<uses-permission android:name="android.permission.INTERNET"/>

Voice recognition initialization

SpeechRecognizer recognizer = SpeechRecognizer.createSpeechRecognizer(this);
Intent intent = new Intent(RecognizerIntent.ACTION_RECOGNIZE_SPEECH);
intent.putExtra(RecognizerIntent.EXTRA_LANGUAGE_MODEL,
                RecognizerIntent.LANGUAGE_MODEL_FREE_FORM);
intent.putExtra(RecognizerIntent.EXTRA_LANGUAGE, Locale.getDefault());

recognizer.setRecognitionListener(new RecognitionListener() {
    @Override
    public void onResults(Bundle results) {
        ArrayList<String> matches = results.getStringArrayList(SpeechRecognizer.RESULTS_RECOGNITION);
        if (matches != null && !matches.isEmpty()) {
            String command = matches.get(0).toLowerCase();
            if (command.contains("Open camera")) {
                // perform operations
            }
        }
    }
    //Other necessary overwriting methods are omitted
});

recognizer.startListening(intent);

Speech synthesis (response to user)

TextToSpeech tts = new TextToSpeech(this, status -> {
    if (status == TextToSpeech.SUCCESS) {
        tts.setLanguage(Locale.TAIWAN);
        tts.speak("Hello, I'm here.", TextToSpeech.QUEUE_FLUSH, null, null);
    }
});

Permanent background monitoring (optional)

If you want to keep the background and wake it up by voice, you need to use:

• Foreground Service
• Third-party wake libraries, such as Snowboy (discontinued), Porcupine (Picovoice）

Things to note

• SpeechRecognizer cannot continuously listen in the background.
• It is recommended to use Hotword Engine or use an external speech SDK.
• Dynamic requests required starting from Android 10RECORD_AUDIOpermissions.

Permanent background monitoring

Purpose

The goal of permanent background monitoring is to allow the App to detect voice wake-up words (such as "Hey Assistant") and activate corresponding functions even when the screen is not open.

challenge

• Android does not allow SpeechRecognizerLong background running.
• Official voice recognitionCan only receive audio for a short period of timeAnd need to be actively aroused.
• Voice wake-up requires the use of special algorithms or external SDKs.

solution architecture

1. useProspect services(Foreground Service) remains operational.
2. Voice wake-up adoptsPorcupineWait for offline Hotword Engine.
3. After waking up successfully, startSpeechRecognizerRecognize complete voice commands.

Step 1: Prospect service establishment

public class VoiceService extends Service {
    @Override
    public int onStartCommand(Intent intent, int flags, int startId) {
        Notification notification = new NotificationCompat.Builder(this, "voice_channel")
            .setContentTitle("Voice assistant is operating")
            .setSmallIcon(R.drawable.ic_mic)
            .build();
        startForeground(1, notification);

        //Initialize Hotword detection
        startHotwordDetection();
        return START_STICKY;
    }

    @Override
    public IBinder onBind(Intent intent) {
        return null;
    }
}

Step 2: Create notification channel (Android 8+)

NotificationChannel channel = new NotificationChannel("voice_channel",
    "Voice Assistant", NotificationManager.IMPORTANCE_LOW);
NotificationManager manager = getSystemService(NotificationManager.class);
manager.createNotificationChannel(channel);

Step 3: Use Porcupine for voice wake-up

Porcupine provides an Android SDK that recognizes custom keywords and operates completely offline.

PorcupineManager porcupineManager = new PorcupineManager.Builder()
    .setAccessKey("your key")
    .setKeywordPath("hey_assistant.ppn")
    .setSensitivity(0.7f)
    .build((keywordIndex) -> {
        //Call SpeechRecognizer for speech recognition when woken up
        startSpeechRecognition();
    });

porcupineManager.start();

Step 4: Start the service

Intent serviceIntent = new Intent(this, VoiceService.class);
ContextCompat.startForegroundService(this, serviceIntent);

Things to note

• Prospect servicesRequired after Android 8, otherwise the system will terminate the app.
• Porcupine supports offline operation, but you need to register an account to obtain itAccessKey。
• Remember to handle user authorizationRECORD_AUDIOpermissions.

Optimization of Hey Google Voice Assistant

User-side optimization suggestions

• Make sure the speech model training is completed:Conduct "Voice Match" training in "Google Assistant Settings" to make the device more accurate in recognizing personal voices.
• Avoid background noise:"Hey Google" works better when used in a quiet environment.
• With the device close to the mouth:The sound quality of the device's microphone will significantly affect the wake-up effect.
• Update Google App:Latest versions often improve speech models and wake sensitivity.
• Turn on the "Always on" setting:Make voice wake-up available when the screen is off (limited on some devices and versions).

Developer-side optimization methods

• Using the Google Assistant SDK:
  Integrates native voice assistant experiences on embedded devices or Android. Suitable for IoT devices or special application development.
  Google Assistant SDK official website
• Enable App Actions:
  Integrate Google Assistant commands through App Actions, allowing users to speak voice commands to open apps and perform specific functions.
  For example: "Hey Google, open my shopping list" will trigger your app behavior.
  App Actions file
• Deep Link + Shortcut provides semantic links:
  Allow your voice assistant to correctly trigger specific pages or actions in your app.

Optimization of voice wake-up trigger conditions

• System level wake-up is only allowed for native Google Assistant.
• Non-system assistants (third-party apps) cannot completely replace "Hey Google".
• If you need to customize the wake word, it is recommended to use tools such as Porcupine and Snowboy (discontinued) to build your own voice assistant.

Be aware of restrictions

• "Hey Google" voice triggering is a system-level permission.Cannot be intercepted by third-party apps。
• Google has not opened the SDK authorization for "Hey Google" voice triggering.
• Android 11+ further restricts microphone permissions and background access.

Apple product program development

development tools

• Xcode:Apple's official development environment (IDE) supports macOS, iOS, watchOS and tvOS application development.
• Swift:A modern programming language from Apple that is safe and efficient.
• SwiftUI:Declarative UI framework to simplify UI development.
• UIKit:A traditional UI framework that provides complete UI controls and custom functions.

Development environment settings

1. Download and install the latest version ofXcode。

2. Open Xcode and go toPreferences > Accounts, sign in with your Apple ID to enable developer features.

3. Install via XcodeCommand Line Toolsto use Swift command-line tools.

Development platform

• iOS:Develop iPhone and iPad apps.
• macOS:Develop desktop applications.
• watchOS:Develop apps for Apple Watch.
• tvOS:Develop Apple TV apps.

Create project

1. Open Xcode and select "Create a new Xcode project".

2. Select a suitable application template, e.g.App (iOS/macOS)。

3. Set the project name, Bundle Identifier and language (Swift or Objective-C).

4. Select a UI framework (SwiftUI or UIKit).

Simulation and testing

• useiOS SimulatorTest the application on different devices.
• throughTestFlightRelease a beta test version to testers.
• useXCTestConduct unit testing and UI testing.

Application publishing

1. RegisterApple Developer Program(Annual fee USD $99 required).

2. Set via XcodeApp Store Connectand submit the app.

3. Follow Apple’sApp Store Review GuidelinesMake sure your app meets listing specifications.

iOS development

development tools

iOS development mainly usesXcode, the official integrated development environment (IDE) provided by Apple.

• Programming languages: Swift and Objective-C
• Support simulator for testing
• Provides Interface Builder design visual interface

development process

1. Design application architecture and interface
2. Write code to implement functions
3. Test using simulator or real machine
4. Perform error troubleshooting and performance optimization
5. Submit and publish via the App Store

Essential knowledge

Learning iOS development requires mastering the following basics:

• Swift syntax and basics
• Use of UIKit or SwiftUI
• Basic design patterns such as MVC, MVVM
• Handling network requests and JSON
• Application life cycle and background operations

Develop resources

Here are some helpful learning and development resources:

• Apple developer official website
• Swift official website
• Ray Wenderlich Teaching Website
• Stack Overflowtechnical community

Xcode

Features

Xcode is an integrated development environment (IDE) provided by Apple for the development of macOS, iOS, watchOS, and tvOS applications.

• Supports Swift and Objective-C languages
• Built-in simulator can test multiple devices and system versions
• Provides Interface Builder design visual interface
• Integrate Git version control
• Powerful compiler and debugging tools

Main components

1. Code Editor：Provides syntax highlighting, code completion and error prompts
2. Interface Builder：Supports intuitive drag-and-drop design interface
3. Simulator：Used to test and simulate applications running on different devices
4. Debugging tools:Supports breakpoints, memory detection and performance analysis

Installation and updates

You can download the latest version of Xcode from the Mac App Store or Apple's official developer website.

• System requirements: macOS latest version
• Regularly updated to get the latest features and bugfixes

Tips

Practical tips to improve development efficiency:

• Use shortcut keys to quickly switch between files and views
• Configure custom templates to reuse code
• Enable static analysis tools to check for potential issues
• Learn how to use LLDB commands for efficient debugging

resource

Related learning and reference resources:

• Xcode official website
• Apple Developer Documentation
• Ray Wenderlich Teaching Website

Swift

language features

Swift is Apple’s modern programming language for developing iOS, macOS, watchOS, and tvOS apps.

• Fast and efficient performance
• Concise and easy-to-read syntax
• Supports safety checks, such as optional types to avoid null pointer errors
• Supports functional programming and object-oriented programming
• Powerful error handling mechanism

Grammar basics

• Variables and constants:usevarandlet
• Optional types:use?and!Handles the presence or absence of a value
• Control process:includeif、switch、for、while
• function:usefuncDefinition, supports parameter labels and multiple return values
• Categories and Structure:useclassandstruct

core concepts

1. Use Protocol to implement interfaces and protocols
2. Extend category functionality using Extension
3. Generics support improves code reusability
4. Use closures to implement higher-order functions
5. Supports custom operators and tuples

Application scenarios

Swift is not just for the Apple ecosystem, it can also be used for server-side development and cross-platform tools.

• iOS and macOS application development
• Server-side applications using Swift on Server
• Academic research and rapid prototyping

resource

Related learning and reference resources:

• Swift official website
• Apple Swift developer resources
• Hacking with Swift tutorial website
• Stack Overflowtechnical community

Objective-C

Features

Objective-C is a C-based object-oriented programming language originally developed by NeXT Corporation and later widely used by Apple for macOS and iOS application development.

• Based on C language, with high performance and flexibility
• Support dynamic runtime (Runtime)
• Use Smalltalk-style messaging

grammatical structure

Objective-C's syntax combines the features of C and Smalltalk, using the @ symbol to indicate language extensions.

• Class definition:use@interfaceand@implementation
• Messaging:Use square bracket syntax[object method]
• property:use@propertyand@synthesize

core concepts

1. Object-oriented programming, including classes and inheritance
2. Protocols are used to define collections of methods
3. Categories are used to extend category functionality
4. Block syntax for closures and callbacks

development tools

Objective-C development is primarily done using Apple’s Xcode.

• Provides editing, compiling and debugging functions
• Integrate Interface Builder to design visual interface
• Supports iOS and macOS application development

resource

Here are some resources for learning and reference:

• Apple official documentation
• objc.ioTechnology Blog
• Stack Overflowtechnical community
• GitHubSave and share code

GNews

What is GNews?

GNews is a news aggregation platform developed by Google, designed to help users obtain the latest global news. It integrates content from various news sources and uses artificial intelligence technology to personalize recommendations of news of interest to users.

Main functions of GNews

• News aggregation:GNews collects the latest news reports from major news websites, blogs and social media, providing a diverse range of news sources.
• Personalized recommendations:Based on the user's reading habits and interests, GNews will intelligently recommend relevant news content, allowing users to quickly find topics of interest.
• Live updates:The platform will continue to update news information so that users can obtain the latest domestic and international news at any time.
• Multi-language support:GNews supports multiple languages ​​to facilitate users in different regions to obtain news information.

How to use GNews?

Users can use this platform by visiting GNews’ website or downloading its app. On the platform, users can select topics of interest, follow specific news sources, and customize news feeds to suit their needs.

Advantages of GNews

• Convenience:Users do not need to visit each news website one by one, GNews concentrates all relevant information on one platform.
• Get information quickly:The real-time update function allows users to quickly obtain the latest news, whether it is political, economic, technology or entertainment news.
• Diverse news sources:GNews integrates news from different sources to enable users to obtain more comprehensive information.

in conclusion

GNews is a powerful news aggregation tool that uses artificial intelligence technology to provide users with a personalized news experience. As news information changes rapidly, GNews helps users quickly keep up with world trends and obtain the information they need.

Automated process tools

AI development program

Bolt

Bolt is a fast, lightweight AI development framework that focuses on providing developers with simple and efficient tools to build applications. Features include:

• Quick Start: Suitable for novice developers to quickly build prototypes.
• Modular design: highly scalable and suitable for a variety of application scenarios.
• Lightweight: takes up less resources and runs efficiently.

Cursor

Cursor is an editor tool designed specifically for AI program development, providing intelligent auxiliary code writing and debugging functions. Features include:

• Code generation: Leverage AI intelligent completion and hints.
• Syntax error checking: Reduce development error rates.
• Built-in version control: Improve team collaboration efficiency.

v0

v0 is an AI platform based on visual development that allows users to build models and applications through a drag-and-drop interface. Features include:

• No coding required: suitable for developers with no technical background.
• Intuitive interface: Visual operation, low learning curve.
• Instant preview: quickly test the model effect.

Codeium

Codeium is a program editor combined with AI intelligent assistance, focusing on improving the efficiency and accuracy of program development. Features include:

• Smart completion: Generate high-quality code based on context.
• Multi-language support: Suitable for development needs in multiple programming languages.
• Integrated environment: compatible with multiple IDEs and seamlessly integrated.

software engineering

definition

Software engineering is an engineering discipline that develops, operates, and maintains software in a systematic and planned manner. The goal is to build high-quality, maintainable, reliable, and demanding software systems.

core goals

• Improve software quality
• Reduce development costs
• Shorten development time
• Enhance maintenance and scalability

Software Development Life Cycle (SDLC)

1. Need analysis:Collect and analyze user requirements and prepare requirements specifications
2. System design:Design system architecture, modules, databases and interfaces
3. Implementation:Write code based on design
4. test:Conduct unit testing, integration testing, system testing and acceptance testing
5. Deployment and launch:Deploy the system to production environment
6. Maintenance and updates:Fix bugs, update features, optimize performance

Common development methodologies

• Waterfall:Phased completion, clear process, high change cost
• Agile:Iterative development, rapid feedback, and emphasis on teamwork
• Scrum：An agile method, divided into sprints, daily meetings, retrospectives and other rituals
• DevOps：Integrate development and operation and maintenance, emphasizing automation and continuous delivery

Common tools and techniques

• Version control: Git, GitHub, GitLab
• Project management: Jira, Trello, Asana
• CI/CD：Jenkins、GitHub Actions、GitLab CI
• Testing tools: JUnit, Selenium, Postman

Quality attributes

• Availability
• Maintainability
• Scalability
• Security
• Efficiency

Division of roles

• Systems Analyst:Responsible for demand interviews and analysis
• Software designer:Design system architecture and modules
• programmer:Responsible for writing and testing programs
• Test Engineer:Execute tests and report defects
• project Manager:Planning schedule and coordinating team

Agile development

core value

• People and interactions trump processes and tools
• Working software is better than detailed documentation
• Customer collaboration trumps contract negotiation
• Dealing with change is better than following a plan

main principles

1. Address customer needs with the highest priority, providing value through early and consistent delivery.
2. Welcome changes in requirements and use them to create competitive advantages for customers even in late stages of development.
3. Frequent delivery of working software, often with lead times measured in weeks or months.
4. Development and business teams should work together every day to promote communication and understanding.
5. Center around motivated individuals, provide support and trust them to get the job done.
6. Face-to-face communication is the most effective way of communication.
7. The primary measure of progress is working software.
8. Promote a sustainable development pace, and all participants should be able to maintain a stable pace.
9. Pursue technical excellence and design simplicity, and improve flexibility and quality.
10. Keep it simple and focus on getting the necessary work done.
11. Self-organizing teams produce optimal designs and architectures.
12. Reflect regularly and adjust behaviors to improve effectiveness.

Commonly used frameworks

• Scrum: Emphasizes short sprints and fixed-paced development.
• Kanban: Focus on continuous flow and visualization work.
• Extreme Programming (XP): Focus on technical practice and high-frequency delivery.
• Lean: Focus on reducing waste and improving value stream efficiency.

Applicable scenarios

• Projects with frequently changing requirements.
• Situations where a minimum viable product (MVP) needs to be delivered quickly.
• The development process requires close customer involvement.
• Cross-functional team collaboration projects.

Software development requirements document template

Project name

Please enter the official name of this project.

version record

• Version:1.0
• Date of writing:YYYY-MM-DD
• Written by:Name
• Reviewer:Name

Project background and goals

Explain the reason for this project, background issues and core issues to be solved, and define clear project goals.

System overview

Outline the system functions and overall architecture, which can be paired with the system architecture diagram.

Functional requirements

• Function number:For example: F-001
• Function name:User registration
• Function description:Users can register via email and password.
• Prerequisites:Users without registered accounts
• enter:Email, password
• deal with:Verify format and store in database
• Output:Registration success message or error message

non-functional requirements

• The system should have 99.9% availability
• Supports at least 100 concurrent requests per second
• Response time is less than 2 seconds
• Support HTTPS encryption

User roles and permissions

• General users:Register, log in, and use basic functions
• administrator:User management, system settings

Screen and interface requirements

You can attach main screen sketches or wireframes to describe the elements and interactive processes of each screen.

Data requirements

List the main data table structure, fields, correlations, etc.

Integration and interdependence

List external systems, APIs, or other software components that need to be integrated.

Timeline and Milestones

• Requirement confirmation: YYYY-MM-DD
• The first version of development is completed: YYYY-MM-DD
• Test completed: YYYY-MM-DD
• Online time: YYYY-MM-DD

Risks and Limitations

List potential risks, constraints (such as budget, manpower, technology, etc.).

appendix

Related document links, reference materials, term definitions, etc.

design pattern

definition

Design Patterns are a set of proven software design solutions that are mainly used in object-oriented programming to solve recurring design problems in specific situations.

Three major categories

• Establishment type:(Creational) focuses on how to create objects
• Structural type:(Structural) Focus on the combination of categories and objects
• Behavioral type:(Behavioral) Handles the interaction and distribution of responsibilities between objects

Common establishment patterns

• Singleton mode:(Singleton) Ensure that a category will only have one instance in the system
• Factory method:(Factory Method) Use subcategories to decide which object to create
• Abstract factory:(Abstract Factory) provides an interface for creating a set of related objects
• Builder mode:(Builder) Separates the creation process of objects from their presentation
• Prototype mode:(Prototype) Create new objects by copying existing instances

Common structural patterns

• Adapter mode:(Adapter) Converts one interface to another interface expected by the user
• Decorator pattern:(Decorator) Dynamically add new functions without modifying the original category
• Appearance mode:(Facade) Provides a unified interface to access complex subsystems
• Combination mode:(Composite) Combine objects into a tree structure to represent the whole/part hierarchy
• Agent mode:(Proxy) Provides a mechanism to control access to target objects

Common behavioral patterns

• Observer mode:(Observer) When an object changes state, automatically notify all dependents
• Strategy mode:(Strategy) defines a series of algorithms and can be replaced during execution
• Command mode:(Command) Encapsulate operations as objects to support operations such as restoration and scheduling
• Status mode:(State) allows an object to change its behavior when its internal state changes
• Iterator pattern:(Iterator) provides a way to traverse the elements of a collection

Advantages of design patterns

• Improve the reusability and maintainability of program structures
• Make the system more flexible and scalable
• Strengthen communication and consensus in team development

Usage suggestions

• Choose the appropriate model based on actual needs to avoid over-design
• Better results when combined with SOLID principles
• Importing design patterns during refactoring can improve readability and stability

version control

definition

Version Control is a system that manages file change history. It is widely used in software development, allowing multiple developers to collaborate at the same time and retaining a complete record of each change.

Main functions

• Track the content and reasons for each file change
• Revert to any historical version
• Branch management to support multi-version parallel development
• Integrate changes from multiple developers

Version control system type

• Local version control:Records are stored on a single computer
• Centralized version control:(Centralized VCS) such as SVN, all developers collaborate through a central server
• Decentralized version control:(Distributed VCS) such as Git, each developer has a complete version library

Common tools

• Git：The most widely used decentralized version control system
• GitHub：Git's cloud hosting platform supports Pull Request, Issue, CI/CD and other functions
• GitLab：Similar to GitHub, with built-in CI/CD management capabilities
• Bitbucket：Supports Git and Mercurial, suitable for integration with Jira

Git basic instructions

• git init:Initialize the version library
• git clone [url]:Copy remote project
• git add [file]:Add changes to the staging area
• git commit -m "message": Submit changes
• git push: Push changes to the remote repository
• git pull: Get and merge remote changes
• git branch: View or create branches
• git merge: merge branches

branching strategy

• Main / Master：Stable and deployable version
• Develop：Main development branch
• Feature Branch：For new function development
• Release Branch：Prepare for release
• Hotfix Branch：Urgent fix for production issues

benefit

• Improve team collaboration efficiency
• Reduce version conflicts and risks
• Ensure the development process is traceable
• Supports quick switching and testing of different versions

Git

Introduction

Git is a decentralized version control system developed by Linus Torvalds (the father of Linux) in 2005 for tracking code changes, collaborative development, and version management. It is the most widely used version control tool today and is used in personal development, team projects and open source communities.

core concepts

• Repository:The folder that stores all version history and files can be divided into local and remote.
• Commit：A change record represents the modification status of a group of files.
• Branch:Independent development lines allow multiple people to develop different functions at the same time without interfering with each other.
• Merge:Integrate changes from different branches into the master branch.
• Clone：Copy the remote repository to local.
• Push / Pull：Upload or download changes to a remote repository.

Install Git

1. Go toGit official websiteDownload the appropriate version.
2. After installation, open Command Prompt (CMD) or PowerShell.
3. entergit --versionto confirm the installation was successful.

Initial settings

When using Git for the first time, you need to set developer information:

git config --global user.name "your name"
git config --global user.email "Your [email protected]"

basic instructions

git init # Initialize local repository
git clone URL # Copy the remote repository
git add . # Add changes to the staging area
git commit -m "description" # Create a commit
git status # View current status
git log # View commit records
git branch # View branch list
git checkout branch name # switch branches
git merge branch name # Merge the specified branch
git push # Push changes to the remote end
git pull # Pull updates from the remote end

Common workflow

1. Create or copy a repository (git initorgit clone）
2. Edit or add files
3. Add the modifications to the staging area (git add）
4. Commit changes (git commit）
5. Sync with remote end (git push / git pull）

Integrate with Visual Studio

• Visual Studio has built-in Git support and can directly operate branches, Commit, and Push.
• Can connect to GitHub, Azure DevOps, GitLab and other services.
• Version changes can be viewed and managed through the "Git Changes" panel.

advantage

• Decentralized architecture, each developer has a complete version history.
• It is fast, efficient and supports offline operations.
• Powerful branching and merging functions.
• Cross-platform support (Windows, Linux, macOS).

Common remote services

• GitHub— The world’s largest open source community platform
• Azure DevOps— Microsoft enterprise-level integration platform
• GitLab— Self-hosted and cloud version control services
• Bitbucket— Project platform supporting Git and Mercurial

Git reset sync remote

Core command: git reset

If you want to completely discard all local changes that have not been committed (Uncommitted) and have been committed but not yet pushed (Local Commits), and reset the branch to be completely consistent with the remote end, please follow the steps below.

1. Get the latest status of the remote end

First, make sure your local index is updated with the latest progress on the remote end:

git fetch --all

2. Force reset to remote branch

use--hardThe parameter points the current branch to the corresponding remote branch (assuming your branch is main):

git reset --hard origin/main

Handle untracked files (Untracked Files)

git reset --hardOnly files already tracked by Git will be processed. If you have newly created files or folders locally but have not yet git added them, they will remain. To delete them all, usecleaninstruction:

Delete untracked files

# Execute the test first to see which files will be deleted
git clean -n

# Officially delete files (-f) and directories (-d)
git clean -fd

Common scenario comparison table

Things to note

• Irreversibility: git reset --hardis a dangerous operation. After execution, all local uncommitted code willUnable to retrieve。
• Branch name:Please make sure toorigin/mainReplace with the actual remote branch name you are currently using (e.g.origin/masterororigin/dev）。

Git conflict file comparison

When a Git merge branch conflicts, the file will be marked as "Unmerged". At this time, Git will insert a conflict mark in the file, and we need to use comparison commands to locate the problem and fix it manually.

1. Find the conflicting file list

Before resolving the conflict, first identify which files are in conflict:

Conflicting files will be listed inUnmerged pathsunder section.

2. Compare conflicting content

When the file contains conflicting tags, you can use the following methods to view the differences:

• Check out the standard differences:
  git diff
  This will show the difference between the current working directory and the staging area. In case of conflict it will show<<<<<<<、=======and>>>>>>>mark.
• Compare versions on a specific end:
  Compare the current branch (Ours) with the conflict file:
  git diff --ours <file>
  Compare the "branch to be merged in (Theirs)" with the conflict file:
  git diff --theirs <file>
• Three-way comparison (Base, Ours, Theirs):
  Show the common ancestor and the differences between branches at both ends:
  git diff -p <file>

3. Interpret conflict markers

Git will directly modify the conflict file, the format is as follows:

<<<<<<< HEAD (or your branch name)
This is the content of your current branch (Ours)
=======
This is the content of the other party's branch (Theirs)
>>>>>>> branch_name

4. Use graphical tools for comparison (recommended)

For complex conflicts, the text interface is difficult to read. It is recommended to call a dedicated merge tool:

This will start something likeMeld, P4Merge, or VS CodeUsing other tools, compare three versions side-by-side: base version (Base), local version (Local), and remote version (Remote).

5. Actions after resolution

After you manually edit the files and decide which code to keep, you must perform the following steps to complete the merge:

1. Remove from file<<<<, ====, >>>>mark.
2. git add <file>(Add the resolved file to the temporary storage area).
3. git commit(Complete merge commit).

GitHub

GitHub is a cloud-based version control and collaboration platform mainly used for software development. it usesGitImplement version control, allowing developers to manage a project's source code, track changes, and collaborate with others.

Main functions

• Version control:Track file changes through Git, keep historical records, and be able to go back to previous versions.
• Collaboration tools:Developers can collaborate on development through pull requests, code reviews and other functions.
• project management:Provides issues, projects and other functions to track tasks and plans.
• GitHub Actions：Support CI/CD (Continuous Integration and Continuous Delivery) automated workflows.

Benefits of using GitHub

• Centrally manage code to ensure code transparency and traceability.
• Provides complete version control tools to facilitate rolling back or merging different versions.
• Developers can conduct code reviews through the platform to improve code quality.
• Support automated processes and improve development efficiency.

Applicable objects

GitHub is suitable for all types of developers, whether they are solo developers, open source communities, or enterprise teams. It can meet the version control and collaboration needs of small projects to large software projects.

GitHub usesgit pull --rebase

git pull --rebaseIt pulls the latest changes from the remote branch and re-applies the local changes to the latest remote commit, instead of using traditional merge.

Usage

git pull --rebase

Workflow

1. Get the latest changes from the remote branch.
2. Temporarily remove local commits.
3. Apply remote commits to local branch.
4. Reapply local commits to keep history linear.

Why use--rebase？

• Keep linear history: --rebaseYou can keep your commit history concise, with no redundant merge commits.
• Avoid merge commits:Prevents excessive merge records from being generated using traditional merges.

Usage examples

Suppose you commit some changes locally and there are new commits on the remote end, usegit pull --rebasemeeting:

• Download and apply the remote's latest commit.
• Reapply your local commits after these latest remote commits.

resolve conflicts

If a conflict occurs while replaying a local commit, Git will ask for manual conflict resolution:

1. resolve conflicts and usegit addStage resolved files.
2. usegit rebase --continueContinue replaying the commit.
3. If you want to cancel the operation, usegit rebase --abortBack to status quo.

Summarize

git pull --rebaseIt is an important tool for keeping Git submission history clean. It is especially suitable for multi-person collaboration. It can avoid redundant merge submissions and keep the submission record of the code base linear.

GitHub User Access

To allow other developers to participate in your project, you need to add them as collaborators through the repository's settings page. The following is the operation process:

Operation steps

1. Enter the repository: Open the homepage of the project you want to authorize on GitHub.
2. Open settings: Click on the rightmost side of the menu bar aboveSettings(Gear icon).
3. Manage access: In the left sidebar, clickAccessunder categoryCollaborators。
4. Authentication: You may be asked to re-enter your GitHub password or perform two-step verification to confirm your identity.
5. Add new member:ClickAdd peoplebutton.
6. Search for users: Enter the other party’s name in the pop-up windowGitHub account name、full nameore-mail。
7. Send invitation: After selecting the correct user, clickAdd [name] to this repository。

Permission Levels

For personal account repositories, invitees usually have read and write permissions by default; if it is an organization account, more detailed settings can be made:

Things to note

• accept invitation: After adding, the other party will not automatically join. You must check the mailbox bound to its GitHub and clickAccept Invitationlink, or accept the invitation in the GitHub notification center.
• Invitation time limit: Invitation links usually expire after 7 days.
• Private repository restrictions: If you have a free GitHub account, there is no limit to the number of collaborators you can have on a private repository, but some advanced features, such as protected branches in GitHub Actions, may be limited.

Delete a GitHub repository

Deleting a GitHub repository via web interface

This is the most common and intuitive way to delete. Please note that once a repository is deleted, it cannot be restored unless you have a backup.

1. Go to GitHub and go to the file you want to deleteRepository main page。
2. Click on the rightmost side of the paging tab aboveSettings(settings).
3. Scroll the page to the bottom and find the one marked in redDanger Zone(danger area).
4. ClickDelete this repositorybutton.
5. A confirmation window will pop up asking you to enter the full name of the repository (usually in the formatUsername/repository name）。
6. Click the button below to confirm deletion.

Delete using GitHub CLI

If you have GitHub's command line tool (gh) installed, you can quickly delete it through the terminal:

gh repo delete <username>/<repository name> --confirm

Notice:--confirmThe parameter will skip the confirmation step directly, please use it with caution.

Delete local repository folder

After the remote repository on GitHub is deleted, the local folder on your computer will still exist. To remove completely, delete the folder manually:

Linux / macOS / ChromeOS

rm -rf <folder name>

Windows (CMD)

rd /s /q <folder name>

Delete Behavior Checklist

Things to note

• Permission restrictions:You must be the owner of the repository or a member of the organization with administrator permissions to perform a delete operation.
• Public repositories:Deleting a public repository will not delete other people's forks, but it will break the link to the original project.
• Two-stage verification:When you delete an important repository, GitHub may ask you for a password or two-step authentication (2FA) again.

Visual Studio uses Git

Overview

Visual Studio has built-in Git integration function, but you still need to install the Git executable file in the system to perform version control operations (such as Commit, Push, Pull, Merge, etc.).

Step 1: Confirm whether Git is installed

1. Open Command Prompt (CMD) or PowerShell.
2. Enter the following command:

   git --version

3. If it displays something likegit version 2.x.x, means it has been installed; if "Unrecognized git command" is displayed, you need to install Git.

Step 2: Download and install Git

1. Go toGit official download page。
2. Select the corresponding operating system version (Windows, macOS, or Linux).
3. After downloading, execute the installation program. It is recommended to use the default settings and click "Next" all the way.
4. After the installation is complete, restart Visual Studio.

Step 3: Set up Git user information

Turn on the command prompt and set your username and email:

git config --global user.name "your name"
git config --global user.email "Your [email protected]"

This information will be appended to each commit and used to identify the developer.

Step 4: Enable Git in Visual Studio

1. Start Visual Studio.
2. Click on the menu aboveView → Team Explorer。
3. chooseConnect → Clone a repositoryto copy the remote repository.
4. Or click in an existing projectGit → Create Git repository, making the project part of version control.

Step 5: Verify integration status

• You can see the current branch name in the status column of Visual Studio.
• You can use the "Git Changes" window to perform Commit, Pull, and Push operations.
• If Visual Studio displays "Git not detected", please confirm that the Git installation path has been added to the system environment variables.

Supplement: Use Visual Studio’s built-in Git installation

If using Visual Studio 2022 or newer, you can check during the installation process「Git for Windows」options. In this way, Visual Studio will automatically install and integrate Git without manual configuration.

Effect after completion

Once the settings are complete, you can complete the following operations directly in Visual Studio:

• Create branches, commit, push and merge
• View version history and compare differences
• Integrate with GitHub, Azure Repos or other remote repositories

Azure DevOps

Platform introduction

Azure DevOps is an integrated development and collaboration platform provided by Microsoft that supports the complete DevOps process from code management, construction, automated testing to deployment. It is suitable for a variety of programming languages ​​and frameworks and can be used for personal projects or large enterprise team development.

Main service components

• Azure Repos：Provides Git or TFVC version control and supports Pull Request, branch management and source code review.
• Azure Boards：Used for work project tracking, Kanban, Scrum planning and progress management.
• Azure Pipelines：Supports CI/CD (continuous integration and continuous deployment), which can automatically build, test and deploy applications to multiple platforms.
• Azure Test Plans：Provides manual and automated test management tools to ensure software quality.
• Azure Artifacts：Manage packages and dependent libraries, and create private NuGet, npm, Maven and other packaging sources.

Main features

• Cross-platform support, runs on Windows, Linux, macOS.
• Integrates with GitHub, Docker, Kubernetes, Jenkins, and more.
• Provides REST API and YAML-based Pipeline configuration.
• Supports cloud and on-premises deployment (Azure DevOps Server).

Common uses

• Create Git repositories and manage code versions.
• Establish an automated construction process (CI).
• Set up an automated deployment process (CD) to deploy applications to Azure, AWS or on-premises servers.
• Manage product backlog and development progress through Boards.
• Perform quality verification and test reporting in Test Plans.

Usage

1. Go toAzure DevOps official website。
2. Sign in or sign up for a Microsoft account.
3. Create new organizations and projects.
4. Set up modules such as Repos, Pipelines, and Boards to start the development and deployment process.

advantage

• Centralized development management platform integrates the entire process from planning to deployment.
• Clear authority and role management, suitable for large teams.
• It is highly flexible and supports multi-cloud and multi-language development environments.

Applicable scenarios

• CI/CD process management for enterprise-level projects.
• Cross-department collaboration and task tracking.
• Open source and private code repository management.

Azure DevOps access tiers

Step instructions

1. log inAzure DevOps。
2. Click on the lower left cornerOrganization Settings。
3. Select from the menu on the leftUsers。
4. The system lists all users in the organization and their correspondingAccess Level。

Common Access Level Types

• Basic：Most development features such as Boards, Repos, Pipelines are available.
• Stakeholder：Only able to view and participate in basic work projects (free).
• Basic + Test Plans：Includes Basic permissions and access to test management features.
• Visual Studio Subscriber：If the account is bound to a Visual Studio subscription, it will automatically have the corresponding level of permissions.

Advanced operation

• If you are an organization administrator, click on the right side of the userThree-dot menu (⋯)Modify its level.
• If the permissions are insufficient, please contact the organization owner or system administrator to increase the access level.

hint

If you encounter permission restriction issues when using Azure DevOps Pipelines, Repos or Boards, please first confirm whether your Access Level isBasicor above.

CMake cross-platform build system

CMake definition

CMake is an open source, cross-platform "Meta-build System". It does not directly compile the program code itself, but automatically generates project files suitable for the current operating system and development environment by reading the configuration file (CMakeLists.txt). For example, on Windows it generates a Visual Studio solution, while on Linux it generates a Makefile or Ninja configuration file.

Install CMake

How to install CMake on different operating systems is as follows:

• Windows: Go to the official website cmake.org to download the MSI installer. Be sure to check "Add CMake to the system PATH" during installation so that characters can be used directly in the command prompt.
• macOS: It is recommended to use Homebrew and enter in the terminalbrew install cmake. You can also download the .dmg file and install it manually.
• Linux: Install using a package manager, such as Ubuntu inputsudo apt update && sudo apt install cmake。

Core functions

CMake is designed to simplify the management of complex software projects. It supports multi-level directory structures, handles dependencies between libraries, and automatically searches for installed libraries on the system. In addition, it also provides testing (CTest) and packaging (CPack) functions, covering the complete life cycle from development to release.

Mainstream construction process

x64 and ARM64 support

CMake has excellent cross-architecture support capabilities. In Windows environment, developers can specify the target architecture through parameters, such as using-A x64or-A ARM64. For modern Apple Silicon, CMake supports generating Universal Binaries, allowing programs to run on both Intel and Apple M-series processors.

Why choose CMake

Traditional Makefiles are difficult to maintain and do not have cross-platform versatility. CMake allows developers to simply write a set of configuration files, and developers around the world can easily open and compile projects in their preferred IDE (such as VS Code, CLion, Xcode). This is why most mainstream open source C++ projects currently use CMake as a standard tool.

Program deployment

definition

Application Deployment is the process of moving a developed application from a development environment to a testing or production environment so that users can actually use it. It ensures that the system can perform stably and reproducibly in different environments.

Deployment process

1. Build:Compile and package source code into executable files or image files.
2. Test:Verify functionality, performance and compatibility.
3. Deploy:Deploy applications and configuration files to the server.
4. Monitor:Continuously monitor operating status, record logs and performance indicators.

Common deployment methods

• Manual deployment:Manually upload files and modify settings, suitable for small projects.
• Automated deployment:Use tools to automate processes and reduce human error.
• Containerized deployment:Use technologies such as Docker to encapsulate applications and make the environment consistent.
• Cloud deployment:Deploy applications to cloud platforms such as AWS, Azure, Google Cloud, and more.

Common deployment architecture

• Standalone deployment:Both the application and the database run on the same server.
• Separation of front and back ends:Front-end (such as React, Vue) and back-end (such as Node.js, Python) are deployed independently.
• Microservice architecture:Split the application into multiple small services and deploy and manage them independently.
• Cluster deployment:Utilize multiple servers to distribute load and ensure high availability.

Automated deployment tools

• Jenkins：Continuous integration and automated deployment tools.
• GitHub Actions：CI/CD platform integrated into GitHub.
• GitLab CI/CD：GitLab's built-in automated process management.
• Ansible：Used to automate server configuration and application deployment.
• Docker Compose / Kubernetes：Support the deployment and management of containerized applications.

Deployment example

# Deploy the example using GitHub Actions
name: Deploy Web App
on:
  push:
    branches: [ "main" ]
jobs:
  deploy:
    runs-on: ubuntu-latest
    steps:
      - uses: actions/checkout@v4
      - name: Build Docker image
        run: docker build -t myapp .
      - name: Deploy to server
        run: docker run -d -p 80:80 myapp

Deployment considerations

• Environment variables and configuration files are separated
• Version control and rollback mechanism
• Automated testing and verification
• Monitoring and log collection
• Security (e.g. SSL, credentials, permission settings)

Deployment strategy

• Blue-green deployment:Use two sets of environments (Blue/Green), and switch traffic after the new version is online to ensure that services are not interrupted.
• Rolling update:Gradually replace old version containers to maintain service continuity.
• Canary deployment:Let a small number of users use the new version first, and then fully promote it after it is stable.

best practices

• Automate deployment processes to ensure consistency
• Record deployment changes using version control
• Establish a continuous integration and continuous delivery (CI/CD) pipeline
• Roll back to previous version at any time
• Monitor performance and error logs to continuously optimize the deployment process

Containerization technology

definition

Docker (container technology) is an open source platform for automating the deployment, scaling and management of applications. It uses "Containers" to package applications and their dependencies together to ensure consistent execution in different environments.

core concepts

• Image:An executable environment snapshot containing the application and its dependencies
• Container:An executable instance created from an image file, running in an isolated environment
• Dockerfile：A configuration file that describes how to construct an image file
• Docker Hub：Official image file hosting platform, similar to "GitHub for Docker"

Advantages of Docker

• Cross-platform consistency: ensure that "the development environment and the production environment are consistent"
• Lightweight and fast: faster than virtual machine startup and less resource usage
• High portability: the image file can be executed on any platform that supports Docker
• Facilitates version control and rollback

basic instructions

• docker pull [image name]: Download image file
• docker run [image name]: Execution container
• docker ps: View executing containers
• docker stop [container ID]: Stop the container
• docker rm [container ID]: Delete container
• docker build -t [name] .: Create an image file based on Dockerfile
• docker images: View image file list

Dockerfile example

# Use Python image file
FROM python:3.10

# Set working directory
WORKDIR/app

# Copy program files
COPY ./app

#Install dependency packages
RUN pip install -r requirements.txt

#Specify the instructions to be executed when the container starts
CMD ["python", "app.py"]

Docker Compose

Docker Compose is a tool that can define multiple container applications. Usedocker-compose.ymlFiles define each service.

version: '3'
services:
  web:
    build: .
    ports:
      - "8000:8000"
  db:
    image: mysql:8
    environment:
      MYSQL_ROOT_PASSWORD: example

Application scenarios

• Establish consistent development and testing environments
• Quickly deploy microservice architecture
• Continuous integration and deployment (CI/CD) process
• Isolated running of multi-version applications

Related technologies

• Kubernetes：Container orchestration and automation management platform
• Podman：Daemon-free container tools
• Docker Swarm：Docker's native container cluster management system

Deploying .NET C++ programs to Ubuntu

Preface

Deployment of .NET C++ (C++/CLI or .NET-based C++ applications) on Ubuntu systems needs to be handled separately according to the project type. If you use pure C++/CLI (relying on .NET Framework), you will not be able to execute it directly on Linux and you need to useCross-platform technology for .NET 6/7/8(Such as C++/CLR → C# or using C++/Native with .NET interoperability).

Overview of deployment methods

Depending on the nature of the application, deployment can be divided into three scenarios:

1. Pure C++ application:Compiled with g++ and executed directly on Ubuntu.
2. .NET Managed C++ (C++/CLI)：Needs to be redesigned for C# or C++/CLI for Core supported by .NET 6 and above (supported by Windows only).
3. Hybrid application:Mainly C#, C++ provides performance modules in the form of dynamic function libraries (.so).

Environment installation

Install the .NET SDK and necessary tools in Ubuntu:

sudo apt update
sudo apt install -y dotnet-sdk-8.0
sudo apt install -y build-essential

(Can be replaced according to versiondotnet-sdk-8.0fordotnet-sdk-7.0or other versions)

Create project

If you use .NET C# as the main program and call the C++ library:

# Create the main application
dotnet new console -n MyApp
cdMyApp

# Create native function library
mkdir native
cd native
nano add.cpp

add.cpp example:

extern "C" int add(int a, int b) {
    return a + b;
}

Compile as a shared function library:

g++ -fPIC -shared -o libadd.so add.cpp

Calling C++ libraries in C# (.NET)

AtProgram.csjoin in:

using System;
using System.Runtime.InteropServices;

class Program
{
    [DllImport("libadd.so", EntryPoint="add")]
    public static extern int Add(int a, int b);

    static void Main()
    {
        Console.WriteLine("3 + 5 = " + Add(3, 5));
    }
}

Execution and testing

Return to the project root directory and execute:

dotnet run

If the output result is3 + 5 = 8, indicating successful deployment.

Release deployment version

Programs can be packaged into independently executable deployment files:

dotnet publish -c Release -r linux-x64 --self-contained true

The generated executable file is located inbin/Release/net8.0/linux-x64/publish/。

Docker container deployment (optional)

Applications can be containerized using Docker to run on any Ubuntu system:

# Dockerfile example
FROM mcr.microsoft.com/dotnet/runtime:8.0
WORKDIR/app
COPY ./publish .
COPY ./native/libadd.so /usr/lib/
ENTRYPOINT ["./MyApp"]

Build and execute containers:

docker build -t myapp .
docker run --rm myapp

Things to note

• C++/CLI only supports Windows and cannot be executed directly on Ubuntu.
• If cross-platform interoperability is required, it is recommended to use P/Invoke (C# calls C++).
• Linux use.soFunction library, used by Windows.dll。
• Make sure the library path is included inLD_LIBRARY_PATH。

Summarize

A feasible way to deploy .NET C++ programs to Ubuntu is usually toC# + C++ native library” structure. If the original program relies on C++/CLI, it needs to be redesigned to support cross-platform .NET. The most recommended way is to use .NET 8 + native C++ modules with Docker to achieve a stable, portable and consistent deployment process.

Module dependency analysis

Module dependency analysis is a software engineering technique used to identify and track external files (such as DLLs, libraries, frameworks) that an application depends on when executing. This is crucial to ensure that the software performs correctly in different environments.

core analysis dimensions

• Static Dependencies: Dependencies that have been determined during the program compilation and linking stages. When the operating system starts the program, it will first load all modules marked as static dependencies.
• Dynamic Dependencies: During program execution, modules are loaded as needed through APIs such as LoadLibrary or GetProcAddress.
• Delay-load Dependencies: A special dynamic loading method. The module will only be loaded when the relevant function is called for the first time to optimize the program startup speed.
• Transitive Dependencies: When a program depends on A.dll, and A.dll depends on B.dll, B.dll is a transitive dependency.

main purpose of analysis

Common analysis tools

• Dependencies (Modern Dependency Walker): Currently the most mainstream open source analysis tool on the Windows platform, supporting Win10/11 API Sets.
• ldd (Linux): In Unix-like systems, used to list the shared library dependencies required by the executable file.
• Process Explorer: Provided by Sysinternals, you can instantly view all DLL modules loaded by the running program.
• objdump / nm: The underlying command line tool is used to view the export and import tables of binary files.

Dependency Walker

Dependency Walker (commonly known as Depends.exe) is a free tool for analyzing the file structure of Windows applications (such as .exe, .dll, .sys, etc.). It is primarily used to scan any 32-bit or 64-bit Windows module and build a hierarchical tree of all related dependent modules.

Core functions

• Module dependency analysis: Lists all DLL files required for program execution and displays their hierarchical relationships.
• View export and import functions: Lists all functions exported by the module, as well as functions being called by other modules.
• Detect missing files: Mark the necessary components that cannot be found when the program is executed. This is the key to solving the problem of the program not being able to start.
• Invalid module detection: Check for broken or non-Windows mods.
• Schema check: Identify whether the module is 32-bit (x86) or 64-bit (x64) to avoid loading errors caused by mixed architectures.
• Execution phase analysis: Through the Runtime Profile function, monitor dynamically loaded modules and errors that occur when the program is running.

Common application scenarios

Usage restrictions and alternatives

Because Dependency Walker has not been updated for many years, it has poor support for some features of modern Windows 10/11 (such as API Sets or lazy-loaded DLLs), and false red warnings often appear. In this case, the following open source alternative tools are recommended:

• Dependencies: Developed by lucasg, it is a modern rewrite of Dependency Walker and fully supports Windows 10/11.
• Dependency Walker official version: Only recommended for maintenance of traditional legacy systems (Windows XP/7).

Dependencies

Dependencies is an open source, modern dependency analysis tool designed to replace Dependency Walker (Depends.exe), which has been discontinued for many years. It is fully optimized for the system architecture of Windows 10 and Windows 11, and can accurately analyze module associations in modern software environments.

Why choose Dependencies

• Support API Sets parsing: Can correctly handleapi-ms-win-*Virtual DLLs that begin with will not generate false circular dependency warnings or missing file errors.
• Supports modern Windows features: Fully supports tracking of WinSxS (Side-by-Side Isolation) and delay-load (Delay-load) modules.
• Greatly improved performance: Compared to the slow speed of older tools when handling large projects, Dependencies uses multi-thread scanning and loads very quickly.
• Intuitive interface: Retains the classic tree view, and enhances the search, filtering and module path positioning functions.

Function comparison table

Common shortcut functions

• reverse dependency query: Check which modules a specific DLL is referenced by.
• Export function verification: Quickly compare whether the function serial number or name between the import table (Import Table) and the export table (Export Table) matches.
• path tracing: Clearly indicate which folder (System32, SysWOW64 or application directory) the system finally loads the module from.

Obtain and install

Dependencies is a green installation-free software that can usually be downloaded from GitHublucasg/DependenciesThe repository downloads the compiled Release version. When using it, just drag the .exe or .dll file you want to analyze into the window to start analysis.

Execution phase detection function differences

The two tools have different design philosophies: Dependency Walker (Depends.exe) includes a dynamic analyzer (Profiler), while modern Dependencies focus on static analysis and API Sets parsing.

Why don't Dependencies execute the program?

• design positioning: Dependencies is mainly designed as a "static analysis tool" to solve parsing errors caused by Windows API Sets. It does not include a debugger engine.
• Security considerations: Executing unknown code is risky, and modern analysis tools tend to leave dynamic monitoring to professional debugging tools.
• alternatives: If you need to dynamically monitor the DLL loading process, developers usually use it together withProcess Monitor (ProcMon)。

How to use Dependency Walker for motion detection (Profiling)

If you need to use the "Executor" to troubleshoot DLL errors that only occur when running (such as delayed loading failures or path search problems), you can continue to use Dependency Walker's Profile function:

1. Start Dependency Walker and load your .exe.
2. Press on the keyboardF7Or click on the menuProfile -> Start Profiling。
3. Set parameters (such as working directory, command line parameters) in the pop-up dialog box, and then clickOK。
4. Look at the Log window at the bottom: When the program executes and crashes, the log at the bottom will indicate in red text which DLL the error occurred while loading.

Comparison of advantages and disadvantages of motion detection

Advanced debugging suggestions

If you are having too many false errors in Dependency Walker's Profiler and can't see the point, it is recommended to use the following combination instead:

• Use Dependencies (static): First confirm whether the basic entity DLL (such as VC++ Runtime, Qt, OpenSSL) exists.
• Using Process Monitor (dynamic): When executing the program, filterProcess NameandPath(ending in .dll). See if there are anyNAME NOT FOUNDrecords, which can tell you exactly where the program failed to find the DLL.

Program development AI tools

GitHub Copilot

Developed by GitHub in partnership with OpenAI, it is currently the most widely used AI coding assistant in the world. It can be embedded directly inVisual Studio Code、JetBrainsIn other popular development environments, it analyzes comments and context to instantly suggest lines of code or entire functions.

Cursor

This is an independent software developed based on VS Code architecture.AI code editor. Unlike a simple plug-in, Cursor deeply integrates AI into the bottom layer of the editor. It understands the structure (Context) of the entire project and allows developers to talk directly to the entire library to perform global refactoring or bug fixes.

Claude 3.5 Sonnet

Although Claude is a general-purpose AI, its3.5 Sonnet modelIt is recognized as one of the top choices in terms of program logic, architectural design and debugging capabilities. matchArtifactsFunction, developers can directly preview the generated web page front-end screen or interactive components in the dialog box.

Tabnine

Tabnine attaches great importance toPrivacy and security, suitable for enterprises with high requirements for source code confidentiality. It supports fully localized deployment (Self-hosting), ensures that the code will not be uploaded to the cloud for training, and provides accurate auto-completion functionality.

Amazon Q (formerly CodeWhisperer)

Development Assistant provided by Amazon Web Services (AWS). In addition to basic code generation, it is especially enhanced withAWS cloud servicesThe integration can help developers quickly write infrastructure as code (IaC) or handle security scans and updates.

Replit Agent

AI agent built into the Replit online development environment. It is characterized byAutomated deploymentWith environment construction, users only need to describe the functions of the application, and Agent can automatically create files, install dependent packages, write code, and complete online operation.

v0.dev

Presented by VercelAI for front-end development. Users describe interface requirements through natural language, and v0 will directly generate front-end component code based on React, Tailwind CSS and Shadcn UI, greatly reducing the time of UI stereotypes.

Agent-based AI programming assistant

Agentic AI Coding Assistants. The characteristics of this type of tool are its ability to directly access the terminal, read and write files, execute tests, and have the ability to plan tasks independently.

1. Terminal and CLI tools

This type of tool is closest to Claude Code and runs directly in your terminal, making it suitable for developers who are accustomed to command line operations.

• Aider：The most popular open source CLI programming assistant available today. It supports a variety of LLMs (such as GPT-4o, Claude 3.5 Sonnet), can automatically handle Git Commit, and has powerful program code reconstruction capabilities.
• Open Interpreter：Let AI execute code on your local computer. It is not limited to programming, but can also control system settings and process data files. It is a broader computer agent.
• Mentat：An open source tool focused on navigating large code bases that automatically scans the entire project and makes modifications based on natural language instructions.

2. IDE and editor tools

Such tools integrate agent capabilities into the editor to provide a more intuitive visual development experience.

• Cursor：Currently recognized as the most powerful AI code editor (based on VS Code). Its "Composer" mode is very similar to Claude Code, which can automatically write code across multiple files and execute terminal commands.
• Windsurf：The new generation IDE launched by Codeium emphasizes the concept of "Flow". The AI ​​assistant can sense your movements and perform tasks autonomously in the background, with extremely high environmental awareness.
• Cline (formerly known as Devinsist/Claude Dev):A VS Code extension that allows Claude to read and write files, execute Bash commands, and operate the browser. It is currently the closest open source alternative to Claude Code on VS Code.

3. Fully automated AI engineer (completely autonomous)

Such tools usually have more independent operating capabilities and can complete the process from demand analysis to deployment on the web or in a stand-alone environment.

• Devin：Known as the world's first "AI software engineer", he can independently solve issues on GitHub and has his own browser, editor and terminal.
• Replit Agent：Integrated into the Replit cloud development environment, users only need to describe the desired functions, and it can build and deploy complete web applications from scratch.

Core function comparison table

How to choose

• if you preferMinimalism and speed:chooseClaude CodeorAider。
• if you preferVisualization and project management:chooseCursororWindsurf。
• if you needCompletely open source and highly customizable:chooseClineorAider。

Claude Code

Claude Code is an agent-based command line interface (CLI) tool developed by Anthropic. It runs directly in the developer's terminal and can understand the entire program code library, execute terminal commands, edit files, and assist in processing Git workflow, allowing developers to complete complex programming tasks through natural language.

Core functions

• Automatic planning and execution:Understand user needs and develop a multi-step execution plan, and execute compilation, testing or repair instructions directly on the terminal.
• Code editing and refactoring:Automatically analyze existing code and perform optimizations, bug fixes, or large-scale feature refactorings.
• Git integration:Assists in generating Commit messages, handling branch switches, and resolving merge conflicts.
• Project understanding:Instead of manually providing context, it automatically scans archives, analyzes the architecture, and answers questions about project logic.

Installation method

Claude Code requires Node.js 18+ environment, which can be installed through the following methods:

1. Install using npm:
   npm install -g @anthropic-ai/claude-code
2. Start the tool:
   Enter in the project directoryclaudeNow you can start interactive mode.

Operation modes and shortcut keys

Commonly used built-in commands

• /help: Shows all available commands and instructions.
• /clear: Clear the current conversation history and reset the context space.
• /stats: Display recent Token usage and expense statistics.
• /doctor: Diagnose whether the installation status is normal.
• /compact: Manually compress conversation content to save space.

Hardware and system requirements

• operating system:macOS 13+, Windows 10+, or a major Linux distribution.
• Memory:4 GB RAM and above recommended.
• Account permissions:Requires Claude Pro, Max, Teams or Enterprise subscription plan, or a valid API Key for billing.

For frequent permission confirmation pop-ups and the need for cross-directory access, optimization settings can be made through startup parameters and built-in commands.

1. Reduce permission confirmation pop-ups

By default, Claude Code will confirm dangerous commands (such as deleting files and executing system commands) one by one. To enter "automation mode" to reduce interference, use the following parameters:

• Automatically accept mode:
  Added at startup--auto-accept, which causes the tool to automatically execute the suggested instructions without stopping to ask. Note, however, that this is not possible when executing destructive instructions such asrm).
  claude --auto-accept
• Trust mode:
  Enter in conversation/approve-all, which grants temporary full trust for subsequent operations in the current session.

2. Access directories other than the current directory

For safety reasons, Claude Code will be limited to the project directory where it is started by default. To access external paths, there are three methods:

• Specify multiple paths when starting:
  You can list all directories that need to be indexed after the startup command:
  claude . /path/to/another/project
• Join manually using the @ symbol:
  During a conversation, you can type directly@And paste the absolute path to the external file, adding its contents into context:
  Please help me refer to the settings of @/etc/nginx/nginx.conf to modify the current project.
• Create a symbolic link:
  Create a link to an external directory in the current project directory. Claude Code will usually read along the link:
  ln -s /path/to/external_dir ./external_dir

3. Set up a permanent alias (Alias)

In order to facilitate quick startup and bring in these settings in the future, it is recommended that you.bashrcor.zshrc(Cygwin/Linux) Add alias:

Things to note

• Security:turn on--auto-acceptLater, Claude may execute shell commands that you have not reviewed. It is recommended to only use it in a trusted project environment.
• Token consumption:When you add a large number of external directories, the number of scanned files will increase, which may increase API Token consumption and processing time.

For the needs of fully automated startup (no confirmation) and access to external directories, Claude Code provides special CLI parameters to achieve this. These features are often referred to as "YOLO mode" and are suitable for use in a trusted environment or sandbox.

1. Fully automatic execution: skip all permission confirmations

If you don't want to executels、mkdirOr if you keep clicking OK when editing a file, you can use the following dangerous parameters:

• Global skip mode:
  --dangerously-skip-permissions
  This is the most powerful parameter. It will simply bypass all permission prompts, giving Claude full autonomy to execute.
• Only enable the skip option (do not start immediately):
  --allow-dangerously-skip-permissions
  This makes "skip permissions" an optional mode, usually paired with--permission-mode bypassPermissionsuse.

2. Access external directories: add permission path

By default, Claude can only read and write the directory in which it is started. To access other paths, use--add-dir：

• Instruction format:
  claude --add-dir <path>
• Multiple directory example:
  claude --add-dir ../another-project --add-dir /var/log
  This way Claude can work on both the current project and the specified external directory.

3. Combination skills: The most powerful automatic start command

If you want to get it right at once and develop directly between multiple directories without hindrance, you can use a combination of:

Startup parameter comparison table

Safety reminder

• Use caution to skip permissions: --dangerously-skip-permissionsDestructive. If Claude misjudges the command (such as accidentally deleting a file), the system will not intercept it.
• Environmental restrictions:In some systems, ifrootorsudoExecuting this parameter with an identity may be blocked due to security mechanisms. It is recommended to execute it under general user permissions.

Cursor

Cursor is an AI code editor developed based on VS Code. It is not just a plug-in, but it embeds AI capabilities deep into the bottom layer of the editor, allowing it to understand the context of the entire project and have the ability to automatically write, reconstruct, and repair code. Since it inherits VS Code, users can directly import all original settings and extended functions.

Basic installation

1. Download software:Go toCursor official website, the system will automatically detect your operating system (Windows, macOS, Linux) and provide download.
2. Import settings:After installation and startup, Cursor will ask if you want to import extensions, shortcuts, and themes from VS Code. It is recommended to choose Import to seamlessly integrate development habits.
3. Login account:Register and log in to your Cursor account. The free version (Hobby) provides a limited number of queries on higher-order models (such as GPT-4o or Claude 3.5 Sonnet), after which it switches to a smaller model.

Core function usage

1. Cmd + K (Inline Edit)

Call the AI ​​directly in the code editor:

• Select a piece of code and clickCmd + K(Windows forCtrl + K)。
• Enter instructions (for example: change this function to asynchronous mode, add error handling).
• The AI ​​will make changes in place, and you can click "Accept" or "Reject".

2. Cmd + L (Chat Pane)

Sidebar chat dialog:

• pressCmd + LOpen the chat window.
• You can ask questions about the project as a whole (eg: Where is the login logic for this project?).
• use@Files、@Codebaseor@Webto specify the range of the AI ​​reference.

3. Cmd + I (Composer mode)

This is Cursor’s most powerful “agent” mode:

• pressCmd + IStart Composer.
• Enter a complex functional requirement (for example: create a complete member registration page, including front-end verification and back-end API).
• AI will automatically create or modify multiple files at the same time and automatically execute terminal commands.

Advanced setup techniques

Comparison of commonly used shortcut keys

• Cmd + K: Edit the selected code or generate new code.
• Cmd + L: Turn on the AI ​​chat sidebar.
• Cmd + I: Start full project generation mode (Composer).
• Cmd + Enter: Forces the AI ​​to scan the entire Codebase in Chat mode.
• @: Enter @ in the input box to mark files, folders, Git records or files.

Things to note

• Privacy mode:"Privacy Mode" can be turned on in the settings. After turning it on, your code will not be used to train the model (only for Pro users or above or specific settings).
• Terminal integration:You can press directly on the terminalCmd + K, let AI help you write complex Shell instructions.

Aider

Install

Aider is a terminal-based AI pairing programming tool that supports direct modification of local source code. Before installation, please ensure that the system hasPython 3.8The above versions are the same asGit。

python -m pip install aider-chat

API key configuration

Before starting, the key provided by the AI ​​service provider needs to be set as an environment variable. Aider supports various models such as Anthropic and OpenAI by default:

• macOS/Linux: export ANTHROPIC_API_KEY=your-key
• Windows: setx ANTHROPIC_API_KEY your-key

After the setting is completed, enter in the terminalaiderThis will open the conversation interface.

File operations

Aider will only read and modify files that you specify to join the context to save token costs and improve accuracy:

• /add <file>: Add the file to the conversation so that the AI ​​can see and edit the content.
• /drop <file>: Remove the profile from the conversation and no longer let it be tracked by the AI.
• /ls: Display the list of files included in the current conversation.

Code editing

You can directly use natural language to ask Aider to perform tasks, such as: "Help me split this class into two files" or "Fix existing bugs." After performing the modification, Aider will automaticallyGit Commit, and automatically compose a submission message based on the modified content.

instruction

Aider uses local model

Feasibility and principles

Aider supports multiple ways to interface with locally executed free and open source models. The most common method is to useOllama、LM StudioorLocalAIWait for the tool to create a local server that is compatible with the OpenAI API specification, and then let Aider point to the server address.

Implemented via Ollama

This is currently the simplest and most mainstream localization solution:

• Download model:Execute in terminalollama pull deepseek-coder-v2(Or choose a model optimized for programs such as Qwen2.5-Coder).
• Start Aider:use--modelThe parameter specifies the Ollama model path.

aider --model ollama/deepseek-coder-v2

core setting parameters

Limitations and Challenges of Local Models

• Logical reasoning ability:Although local models (such as Llama 3 or DeepSeek) perform well on basic code writing, they are still slightly less accurate than Claude 3.5 Sonnet when dealing with complex cross-archive reconstructions.
• Context window:Local models are limited by VRAM size and the context window is usually smaller. If there are too many project files, the memory limit may be exceeded and the reply may be interrupted.
• Hardware requirements:It is recommended to have a GPU with at least 12GB of VRAM to smoothly execute 7B or 14B level code-specific models.

Recommended local models

Not all models are suitable for use with Aider. The following are local open source models currently recognized by the community as performing better in program development tasks:

• DeepSeek-Coder-V2：One of the strongest open source code models currently, the logic is very strict.
• Qwen2.5-Coder：The model developed by Alibaba is excellent at understanding and generating multi-language code.
• CodeQwen：For AiderRepository MapThe function has better compatibility.

Laws related to software development

Copyright law

Software codes, design documents, user manuals, etc. are all covered by copyright protection. Copyright protection is automatically established upon completion of creation, no registration is required, and unauthorized copying, modification, distribution or commercial use is prohibited.

patent law

If the software involves technological innovation or unique algorithms, you can apply for patent protection. Patents are subject to review and approval and protect the inventor's exclusive rights to technical solutions, but pure program logic is usually not protected by patents.

trademark law

Software names, logos, and brand identities can be registered as trademarks to protect the brand image and market differentiation and prevent others from using it to cause confusion.

Business Secrets Protection Act

Undisclosed programming details, algorithms, database structures, etc. are business secrets. The company should protect it through confidentiality agreements, information management systems, etc.

Information security related regulations

For example, the Personal Information Law and the Information Security Management Law require users to properly handle personal data and ensure information security during the software development process. Violations may face administrative or criminal liability.

Contract law and authorization clauses

Software development often involves contract contracts, cooperative development contracts, software licensing terms, etc., which clearly define the rights and obligations of both parties, program code ownership, maintenance responsibilities, confidentiality obligations, etc. to avoid legal disputes.

Internet and e-commerce related regulations

The online operation of software involves network services, electronic payments, online protection of intellectual property rights, etc., and must comply with regulations such as the Electronic Signature Law and the Electronic Transaction Law.

Intellectual Property Rights in Software Development

Copyright

The software code falls within the scope of copyright protection, and developers automatically own the copyright without additional registration. Copyright protection covers source code, design files, user manuals, etc.

patent

If the software involves novelty and technological innovation, it may be possible to apply for patent protection. Examples include unique algorithms or methods of solving technical problems, but pure code logic is usually outside the scope of patent protection.

Trademark rights

Software names and logos can be registered as trademarks to distinguish products from different sources and avoid market confusion.

business secrets

Undisclosed programming details, database structures, and algorithms can be protected through trade secrets. It needs to rely on the company's internal confidentiality agreement and management measures to maintain it.

Authorization mode

Software can be released through different licensing models, such as proprietary software, free software, and open source licenses (such as GPL, MIT). Different licensing models affect users' rights to modify and redistribute.

Infringement risk

Unauthorized use of another person's code, images or algorithms may constitute copyright infringement. The development process should ensure that the source of the program code is legal to avoid legal disputes.

international protection

Intellectual property rights are territorial, but can obtain a certain degree of protection in many countries through international treaties such as the Berne Convention and the TRIPS Agreement.

Software mall

Software sales platform

In 2026, choosing the right software sales platform will depend on your software type and target market. The following are the classifications and characteristics of mainstream platforms:

Universal digital products and SaaS platforms

• Paddle / Lemon Squeezy：The financial custody platform (Merchant of Record) specially designed for software developers can automatically handle complex digital consumption tax and compliance issues in countries around the world.
• Gumroad：The interface is minimalist, suitable for individual creators to sell gadgets, plug-ins or scripts, and the threshold for entry is extremely low.
• AppSumo：The world's leading software promotion platform is suitable for quickly acquiring a large number of users through the Lifetime License Plan (LTD) in the early stages of new product launches.
• Shopify：Through the powerful expansion ecosystem, it can also be transformed into an official brand website that sells software serial numbers and digital licenses.

Game and desktop application distribution platform

• Steam：The world's largest PC game mall, with the strongest organic traffic, takes a 30% sales share.
• Epic Games Store：Provides a more competitive 12% commission rate, with additional discounts for works developed using Unreal Engine.
• itch.io：The first choice for independent developers, it supports free pricing and optional commission ratios, and is suitable for experimental or small and medium-sized projects.

Enterprise and mobile application mall

• App Store / Google Play：The only mainstream channel for mobile applications, providing a complete subscription and in-app purchase mechanism.
• AWS / Azure Marketplace：Suitable for B2B enterprise software, it allows enterprise customers to purchase and deploy directly through cloud service billing.

Taiwanese local e-commerce platform

• SHOPLINE / CYBERBIZ：Provides complete localized cash flow (such as Line Pay, Jiekou Payment), suitable for operating software official websites of local brands in Taiwan.
• Shopee shopping:It has huge traffic and is suitable for selling low-price digital tools, serial numbers or teaching files.

Platform selection comparison table

email: [email protected]

T:0000

資訊與搜尋 | 回dev首頁
email: Yan Sa [email protected] Line: 阿央
電話: 02-27566655 ,03-5924828

阿央
泱泱科技
捷昱科技泱泱企業