golang学习之ngrok源码分析

栏目: Go · 发布时间: 5年前

内容简介:网上有两篇文章已对ngrok原理讲得很清晰了:
从去年开始就对 go 语言产生一点兴趣,总感觉 java 有时太过臃肿,是时候尝试一种新语言了。我看了几本关于go的书,不过看完就忘了,最近开发一个微信公众号项目,使用ngrok做内网穿透,顺道研究一下ngrok源码,巩固一下go语言。

网上有两篇文章已对ngrok原理讲得很清晰了:

https://blog.messyidea.com/archives/41/

https://tonybai.com/2015/05/14/ngrok-source-intro/

我试图记录一些细节,然后实现在个简单的ngrok。

  1. 源码/ngrok/cache/lru.go,这是随手点开的一个文件,从名字可以看出它的用处:实现一个lru(Least Recently Used)缓存,在以往的项目中,我使用的本地缓存是guava提供的工具,很久前我还用过WeakReferenceMap作为一个简单缓存。以下是ngrok中的源码
type LRUCache struct {
   mu sync.Mutex

   // list & table of *entry objects
   list  *list.List
   table map[string]*list.Element

   // Our current size, in bytes. Obviously a gross simplification and low-grade
   // approximation.
   size uint64

   // How many bytes we are limiting the cache to.
   capacity uint64
}

type Item struct {
   Key   string
   Value Value
}

虽然算法很简单,但有几个地方还是值得把玩的。list是一个链表,存储的是item,它可以看成是一个具有优先级的队列,每次访问一个元素时,都会对将相应元素移至队首。这样当list长度超过capacity时,就可以移除队尾的元素。而table是用作索引,通过一个键查询缓存时,通过table得到list中一个element的引用。而list保存的item有一个key属性对应用table中的key,这样很方便对table做移除操作。

以前我想在java中用PriorityBolckingQueue,和ConcurrentHashMap实现一个类似的缓存(不仅是LRU缓存,它们组合实现多种缓存算法),但只想到用map保存键值对,而queue用来做优先队列(这种双向索引的方式我没想到)。说不定用JDK中的HashLinkedMap更容易实现LRU,因为他已同时具备链表和索引的功能。

完整代码如下:

// Copyright 2012, Google Inc. All rights reserved.
// Use of this source code is governed by a BSD-style
// license that can be found in the LICENSE file.

// The implementation borrows heavily from SmallLRUCache (originally by Nathan
// Schrenk). The object maintains a doubly-linked list of elements in the
// When an element is accessed it is promoted to the head of the list, and when
// space is needed the element at the tail of the list (the least recently used
// element) is evicted.
package cache

import (
 "container/list"
 "encoding/gob"
 "fmt"
 "io"
 "os"
 "sync"
 "time"
)

type LRUCache struct {
 mu sync.Mutex

 // list & table of *entry objects
 list  *list.List
 table map[string]*list.Element

 // Our current size, in bytes. Obviously a gross simplification and low-grade
 // approximation.
 size uint64

 // How many bytes we are limiting the cache to.
 capacity uint64
}

// Values that go into LRUCache need to satisfy this interface.
type Value interface {
 Size() int
}

type Item struct {
 Key   string
 Value Value
}

type entry struct {
 key           string
 value         Value
 size          int
 time_accessed time.Time
}

func NewLRUCache(capacity uint64) *LRUCache {
 return &LRUCache{
     list:     list.New(),
     table:    make(map[string]*list.Element),
     capacity: capacity,
 }
}

func (lru *LRUCache) Get(key string) (v Value, ok bool) {
 lru.mu.Lock()
 defer lru.mu.Unlock()

 element := lru.table[key]
 if element == nil {
     return nil, false
 }
 lru.moveToFront(element)
 return element.Value.(*entry).value, true
}

func (lru *LRUCache) Set(key string, value Value) {
 lru.mu.Lock()
 defer lru.mu.Unlock()

 if element := lru.table[key]; element != nil {
     lru.updateInplace(element, value)
 } else {
     lru.addNew(key, value)
 }
}

func (lru *LRUCache) SetIfAbsent(key string, value Value) {
 lru.mu.Lock()
 defer lru.mu.Unlock()

 if element := lru.table[key]; element != nil {
     lru.moveToFront(element)
 } else {
     lru.addNew(key, value)
 }
}

func (lru *LRUCache) Delete(key string) bool {
 lru.mu.Lock()
 defer lru.mu.Unlock()

 element := lru.table[key]
 if element == nil {
     return false
 }

 lru.list.Remove(element)
 delete(lru.table, key)
 lru.size -= uint64(element.Value.(*entry).size)
 return true
}

func (lru *LRUCache) Clear() {
 lru.mu.Lock()
 defer lru.mu.Unlock()

 lru.list.Init()
 lru.table = make(map[string]*list.Element)
 lru.size = 0
}

func (lru *LRUCache) SetCapacity(capacity uint64) {
 lru.mu.Lock()
 defer lru.mu.Unlock()

 lru.capacity = capacity
 lru.checkCapacity()
}

func (lru *LRUCache) Stats() (length, size, capacity uint64, oldest time.Time) {
 lru.mu.Lock()
 defer lru.mu.Unlock()
 if lastElem := lru.list.Back(); lastElem != nil {
     oldest = lastElem.Value.(*entry).time_accessed
 }
 return uint64(lru.list.Len()), lru.size, lru.capacity, oldest
}

func (lru *LRUCache) StatsJSON() string {
 if lru == nil {
     return "{}"
 }
 l, s, c, o := lru.Stats()
 return fmt.Sprintf("{\"Length\": %v, \"Size\": %v, \"Capacity\": %v, \"OldestAccess\": \"%v\"}", l, s, c, o)
}

func (lru *LRUCache) Keys() []string {
 lru.mu.Lock()
 defer lru.mu.Unlock()

 keys := make([]string, 0, lru.list.Len())
 for e := lru.list.Front(); e != nil; e = e.Next() {
     keys = append(keys, e.Value.(*entry).key)
 }
 return keys
}

func (lru *LRUCache) Items() []Item {
 lru.mu.Lock()
 defer lru.mu.Unlock()

 items := make([]Item, 0, lru.list.Len())
 for e := lru.list.Front(); e != nil; e = e.Next() {
     v := e.Value.(*entry)
     items = append(items, Item{Key: v.key, Value: v.value})
 }
 return items
}

func (lru *LRUCache) SaveItems(w io.Writer) error {
 items := lru.Items()
 encoder := gob.NewEncoder(w)
 return encoder.Encode(items)
}

func (lru *LRUCache) SaveItemsToFile(path string) error {
 if wr, err := os.OpenFile(path, os.O_CREATE|os.O_WRONLY|os.O_TRUNC, 0644); err != nil {
     return err
 } else {
     defer wr.Close()
     return lru.SaveItems(wr)
 }
}

func (lru *LRUCache) LoadItems(r io.Reader) error {
 items := make([]Item, 0)
 decoder := gob.NewDecoder(r)
 if err := decoder.Decode(&items); err != nil {
     return err
 }

 lru.mu.Lock()
 defer lru.mu.Unlock()
 for _, item := range items {
     // XXX: copied from Set()
     if element := lru.table[item.Key]; element != nil {
         lru.updateInplace(element, item.Value)
     } else {
         lru.addNew(item.Key, item.Value)
     }
 }

 return nil
}

func (lru *LRUCache) LoadItemsFromFile(path string) error {
 if rd, err := os.Open(path); err != nil {
     return err
 } else {
     defer rd.Close()
     return lru.LoadItems(rd)
 }
}

func (lru *LRUCache) updateInplace(element *list.Element, value Value) {
 valueSize := value.Size()
 sizeDiff := valueSize - element.Value.(*entry).size
 element.Value.(*entry).value = value
 element.Value.(*entry).size = valueSize
 lru.size += uint64(sizeDiff)
 lru.moveToFront(element)
 lru.checkCapacity()
}

func (lru *LRUCache) moveToFront(element *list.Element) {
 lru.list.MoveToFront(element)
 element.Value.(*entry).time_accessed = time.Now()
}

func (lru *LRUCache) addNew(key string, value Value) {
 newEntry := &entry{key, value, value.Size(), time.Now()}
 element := lru.list.PushFront(newEntry)
 lru.table[key] = element
 lru.size += uint64(newEntry.size)
 lru.checkCapacity()
}

func (lru *LRUCache) checkCapacity() {
 // Partially duplicated from Delete
 for lru.size > lru.capacity {
     delElem := lru.list.Back()
     delValue := delElem.Value.(*entry)
     lru.list.Remove(delElem)
     delete(lru.table, delValue.key)
     lru.size -= uint64(delValue.size)
 }
}

以上就是本文的全部内容,希望对大家的学习有所帮助,也希望大家多多支持 码农网

查看所有标签

猜你喜欢:

本站部分资源来源于网络,本站转载出于传递更多信息之目的,版权归原作者或者来源机构所有,如转载稿涉及版权问题,请联系我们

交互设计之路

交互设计之路

库帕 / Chris Ding / 电子工业出版社 / 2006-3 / 38.00元

本书是基于众多商务案例,讲述如何创建更好的、高客户忠诚度的软件产品和基于软件的高科技产品的书。本书列举了很多真实可信的实际例子,说明目前在软件产品和基于软件的高科技产品中,普遍存在着“难用”的问题。作者认为,“难用”问题是由这些产品中存在着的高度“认知摩擦”引起的,而产生这个问题的根源在于现今软件开发过程中欠缺了一个为用户利益着想的前期“交互设计”阶段。“难用”的产品不仅损害了用户的利益,最终也将......一起来看看 《交互设计之路》 这本书的介绍吧!

CSS 压缩/解压工具
CSS 压缩/解压工具

在线压缩/解压 CSS 代码

XML 在线格式化
XML 在线格式化

在线 XML 格式化压缩工具

html转js在线工具
html转js在线工具

html转js在线工具