一种快速编程的方法，推荐给大伙。

go 有 gc 可能还行，c 大概率不行，我不觉得 LLM 会帮你加上有用的生命周期管理

依赖怎么办呢

我经常用 julia 先写然后自己翻译成 cpp 也感觉差不多，而且以后需要翻译的场景可能越来越少 https://www.reddit.com/r/Julia/comments/1g1gzlw/does_julia_still_produce_large_executables/

@alexhx 现代大语言模型可以自动 用目标语言合适的模块 替换。这是真正厉害的地方。

我自己写项目就这样，先用 Python 然后翻译成 C++，不过是纯粹锻炼和好玩。公司的不用。

一个具体的例子,用 ruby 写一个一个 自动生成 图片压缩的服务, 然后 转换成 go

require 'sinatra'
require 'open-uri'
require 'rmagick'
require 'time'


# Initialize cache
$image_cache = {}

# Cleanup task to remove stale cache entries
Thread.new do
loop do
sleep 8 * 60 * 60 # Sleep for 24 hours
current_time = Time.now
$image_cache.delete_if do |url, entry|
current_time - entry[:last_accessed] > 24 * 60 * 60
end
end
end

get '/' do
# Get the URL parameter
url = params['url']

# Ensure the URL is provided
halt 400, 'URL parameter is missing' if url.nil?

# Check cache for the image
if $image_cache.key?(url)
entry = $image_cache[url]
entry[:last_accessed] = Time.now
content_type 'image/jpeg'
return entry[:image_data]
end

# Fetch the image from the URL
begin
image_data = URI.open(url).read
rescue => e
halt 500, "Failed to fetch image: #{e.message}"
end

# Read the image data using RMagick
begin
image = Magick::Image.from_blob(image_data).first
rescue => e
halt 500, "Failed to read image: #{e.message}"
end

# Resize the image while maintaining the aspect ratio
begin
resized_image = image.change_geometry("360") do |cols, rows, img|
img.resize(cols, rows)
end
rescue => e
halt 500, "Failed to resize image: #{e.message}"
end

# Convert the resized image to a blob
resized_image_data = resized_image.to_blob

begin
# Convert the resized image to a blob
resized_image_data = resized_image.to_blob
rescue Magick::ImageMagickError => e
puts "Error: #{e.message}"
puts "Backtrace: #{e.backtrace.join("\n")}"
end

# Store the resized image in the cache
$image_cache[url] = {
image_data: resized_image_data,
last_accessed: Time.now
}

# Send the resized image data to the user
content_type 'image/jpeg'
resized_image_data
end%

转换后的代码,

package main

import (
"bytes"
"fmt"
"image"
"image/jpeg"
"log"
"net/http"
"sync"
"time"

"github.com/nfnt/resize"
)

const (
port = 8300
cacheCleanupInterval = 8 * time.Hour
cacheExpiratiOnTime= 24 * time.Hour
)

type CacheEntry struct {
ImageData []byte
LastAccessed time.Time
}

var (
imageCache = make(map[string]*CacheEntry)
cacheMutex sync.RWMutex
)

func main() {
go cleanupCache()

http.HandleFunc("/", handleRequest)
log.Printf("Server starting on port %d", port)
log.Fatal( http.ListenAndServe(fmt.Sprintf(":%d", port), nil))
}

func cleanupCache() {
for {
time.Sleep(cacheCleanupInterval)
currentTime := time.Now()
cacheMutex.Lock()
for url, entry := range imageCache {
if currentTime.Sub(entry.LastAccessed) > cacheExpirationTime {
delete(imageCache, url)
}
}
cacheMutex.Unlock()
}
}

func handleRequest(w http.ResponseWriter, r *http.Request) {
url := r.URL.Query().Get("url")
if url == "" {
http.Error(w, "URL parameter is missing", http.StatusBadRequest)
return
}

cacheMutex.RLock()
if entry, ok := imageCache[url]; ok {
entry.LastAccessed = time.Now()
cacheMutex.RUnlock()
w.Header().Set("Content-Type", "image/jpeg")
w.Write(entry.ImageData)
return
}
cacheMutex.RUnlock()


log.Printf("Fetching image: %v\n", url)

resp, err := http.Get(url)
if err != nil {
http.Error(w, fmt.Sprintf("Failed to fetch image: %v", err), http.StatusInternalServerError)
return
}
defer resp.Body.Close()

img, _, err := image.Decode(resp.Body)
if err != nil {
http.Error(w, fmt.Sprintf("Failed to decode image: %v", err), http.StatusInternalServerError)
return
}

resizedImg := resize.Resize(360, 0, img, resize.Lanczos3)

var buf bytes.Buffer
if err := jpeg.Encode(&buf, resizedImg, nil); err != nil {
http.Error(w, fmt.Sprintf("Failed to encode resized image: %v", err), http.StatusInternalServerError)
return
}

cacheMutex.Lock()
imageCache[url] = &CacheEntry{
ImageData: buf.Bytes(),
LastAccessed: time.Now(),
}
cacheMutex.Unlock()

w.Header().Set("Content-Type", "image/jpeg")
w.Write(buf.Bytes())
}%

@mike163 不是，我贴的这个不是大模型，这个是人工设计解析转译的，语言转译是非常复杂的，小脚本可以，大一点就没办法了，以我的黄油为例，游戏引擎，渲染框架，UI 框架，到游戏本身，数据结构，AI 算法，游戏逻辑，拼 UI ，还有周边的所有工具链，我做完这一整套，也就四五兆的代码，一个很小的文字黄油，而 AI 要做这种小项目的转译，目前希望也很渺茫