# 在Android中使用OSGi框架(Apache Felix)

## Dalvik VM

Android允许开发者使用Java开发应用，但出于某些原因，代码实际是运行在名为Dalvik的一个针对移动设备平台的虚拟机上，而不是标准的Java虚拟机。Dalvik并不使用标准的Java字节码格式，而是使用Android SDK中的一个工具dx将由Java编译出来的类文件转换为另外一种类文件格式(.dex格式)。这个转换是在编译期完成的。

## 准备Bundles

dx --dex --output=classes.dex JAR_file.jar


aapt add JAR_file.jar classes.dex


adb push JAR_file.jar path_emulator/JAR_file.jar


osgi-android: /
\- bin
\- bundle
\- conf
\- felix.sh


## ERP5中如何应用统一业务模型

UBM统一业务模型完全集成在ERP5的实现中。ERP5中的所有文档都基于该模型设计，ERP5中的所有工具也都采用该模型统一实现不同的业务活动，例如交易、生产、客户关系管理、财务和项目管理。 这使得我们可以基于ERP5快速开发新的应用，同时降低维护成本。因为所有的数据都采用统一的设计架构，因此代码量也大幅度的减少。

# 星际争霸2的大学专业？

[This class] does not teach about Starcraft, but rather aims to utilize the game and the complex situations that arise within it to present and develop the important skills professionals will undoubtedly need in the 21st Century workplace. （这个课程）不会教你关于星际争霸的知识，而是利用游戏中出现的各种复杂情况来展现和研究在21世纪的职场中所需要的各种重要的职业技巧。

This course includes required weekly game play, viewing and analysis of recorded matches, written assignments which emphasize analysis and synthesis of real/game-world concepts, and collaboration with other students. 这个课程的内容和要求包含：每周必须玩游戏，观看和分析比赛录像，写论文（着重分析游戏与现实世界中概念的结合），与其它同学的合作。

via crunchgear

# 10 Ways We Hurt Our Romantic Relationships

10 Ways We Hurt Our Romantic Relationshipsx 幸福爱情的十大忌讳【石榴刺猬全文译】

It's not easy to have a great relationship with your boy/girlfriend, partner, or spouse. But it's not impossible, either — it takes some work, of course, but it's work, work that's a joy when everything comes together.

A lot of times, though, the work isn't enough. We get in our own way with ideas and attitudes about relationships that are not only wrong, but often work to undermine our relationships no matter how hard we work at it.

I've watched a lot of breakups (some of them my own). I've seen dramatic flare-ups and drawn-out slow fades, and I've tried to pay attention to what seems to be going on. Here are a few of the things I've seen that cause people to destroy their own relationships.

1. You're playing to win 　事事争胜 One of the deadliest killers of relationships is the competitive urge. I don't mean competition in the sense that you can't stand to lose at tennis, I mean the attitude that the relationship itself is a kind of game that you're tying to win. People in competitive relationships are always looking for an advantage, the upper hand, some edge they can hold over their partner's head. If you feel that there are things you can't tell your partner because she or he will use it against you, you're in a competitive relationship — but not for long.

1. You don't trust 　缺乏信任 There are two aspects of trust that are important in relationships. One is trusting your partner enough to know that s/he won't cheat on you or otherwise hurt you — and to know that he or she trusts you that way, too. The other is trusting them enough to know they won't leave you or stop loving you no matter what you do or say. The second that level of trust is gone, whether because one of you takes advantage of that trust and does something horrible or because one of you thinks the other has, the relationship is over — even if it takes 10 more years for you to break up.

1. You don't talk 　欠缺沟通　 Too many people hold their tongues about things that bother or upset them in their relationship, either because they don't want to hurt their partner, or because they're trying to win. (See #1 above; example: “If you don't know why I'm mad, I'm certainly not going to tell you!”) While this might make things easier in the short term, in the long run it gradually erodes the foundation of the relationship away. Little issues grow into bigger and bigger problems — problems that don't get fixed because your partner is blissfully unaware, or worse, is totally aware of them but thinks they don't really bother you. Ultimately, keeping quiet reflects a lack of trust — and, as I said that's the death of a relationship.

1. You don't listen 　不会倾听 Listening — really listening — is hard. It's normal to want to defend ourselves when we hear something that seems like criticism, so instead of really hearing someone out, we interrupt to explain or excuse ourselves, or we turn inward to prepare our defense. But your partner deserves your active listening. S/he even deserves you to hear the between-the-lines content of daily chit-chat, to suss out his/her dreams and desires when even s/he doesn't even know exactly what they are. If you can't listen that way, at least to the person you love, there's a problem.

倾听——真正的倾听——是一件很难的事情。当我们听到对自己的批评或者责难的时候，通常我们都会选择防守来保护自己。因此当听到这些对自己的批评的时候，我们不会等对方说完，而是直接打断对方来为自己解释或者开脱。对方会“理所当然”的认为你“应该”仔细的听这些唠叨，然后明白对方想要的东西——尽管有的时候对方自己也不知道自己想要什么。如果你不能用这样的方式来倾听（至少对你爱的人这样），这就是个问题。

2. You spend like a single person 　自己过自己的 This was a hard lesson for me to learn — until it broke up a 7-year relationship. When you're single, you can buy whatever you want, whenever you want, with little regard for the future. It's not necessarily wise, but you're the only one who has to pay the consequences. When you are with someone in a long-term relationship, that is no longer a possibility. Your partner — and your children, if there are or will be any — will have to bear the brunt of your spending, so you'd better get in the habit of taking care of household necessities first and then, if there's anything left over, of discussing with your partner the best way to use it.

This is an increasing problem these days, because more and more people are opting to keep their finances separate, even when they're married. There's nothing wrong with that kind of arrangement in and of itself, but it demands more communication and involvement between the partners, not less. If you're spending money as if it was your money and nobody else has a right to tell you what to do with it, your relationship is doomed.

1. You're afraid of breaking up 　害怕分手 Nobody in a truly happy partnership is afraid of breaking up. If you are, that's a big warning sign that something's wrong. But often, what's wrong is the fear itself. Not only does it betray a lack of trust, but it shows a lack of self-confidence and self-esteem — you're afraid that there's no good reason for someone to want to be with you, and that sooner or later your partner will “wise up” and take off. So you pour more energy into keeping up the appearance of a happy relationships than you do into building yourself up as a person. Quite frankly, this isn't going to be very satisfying for you, and it also isn't going to be very satisfying for your partner.

1. You're dependent 　依赖性强 There's a thin line between companionship and support and dependency. If you depend on your partner — that is, if you absolutely cannot live without her or him — you've crossed that line. The pressure is now on your partner to fill whatever's missing in you — a pressure s/he will learn to resent. If you expect your partner to bring everything while you bring nothing to your relationship — and I'm talking finances as well as emotional support, here — you're in trouble. (Note: I'm not saying that you need to contribute equally to household finances — what I'm saying is that if you're not contributing to the household budget, and you're not contributing anywhere else, things are out of whack and that's never good.)

1. You expect Happiness 　期望得到快乐 A sign of a bad relationship is that one or both partners expect either to make the other happy or for their partner to make them happy. This is not only an unrealistic expectation to lay on yourself or on them — nobody can “make” you happy, except you — but it's an unrealistic expectation to lay on your relationship. Relationships aren't only about being happy, and there's lots of times when you won't and even shouldn't be. Being able to rely on someone even when you're upset, miserable, depressed, or grieving is a lot more important than being happy all the time. If you expect your partner to make you happy — or worse, you're frustrated because you aren't able to make your partner happy — your relationship isn't going to fare well when it hits a rough spot.

1. You never fight 　从不吵架 A good argument is essential, every now and then. In part, arguing helps bring out the little stuff before it becomes major, but also, fighting expresses anger which is a perfectly normal part of a human's emotional make-up. Your relationship has to be strong enough to hold all of who you are, not just the sunny stuff.

One reason couples don't fight is that they fear conflict — which reflects a lack of trust and a foundation of fear. That's bad. Another reason couples avoid arguments is that they've learned that anger is unreasonable and unproductive. They've learned that arguing represents a breakdown rather than a natural part of a relationship's development. While an argument isn't pleasant, it can help both partners to articulate issues they may not have even known they had — and help keep them from simmering until you cross a line you can't come back from.

1. You expect it to be easy/you expect it to be hard 　把爱情想的很轻松或者很难 There are two deeply problematic attitudes about relationships I hear often. One is that a relationship should be easy, that if you really love each other and are meant to be together, it will work itself out. The other is that anything worth having is going to be hard — and that therefore if it's hard, it must be worth having.

The outcome of both views is that you don't work at your relationship. You don't work because it's supposed to be easy and therefore not need any work, or you don't work because it's supposed to be hard and it wouldn't be hard if you worked at it. In both cases, you quickly get burnt out — either because the problems you're ignoring really don't go away just because you think they should. or because the problems you're cultivating are a constant drag on your energy. A relationship that's too much work might be suffering from one of the attitudes above, but a relationship that doesn't seem to need any work isn't any better.

Your choices 你的选择： There isn't any one answer to any of the problems above. There are choices though: you can either seek out an answer, something that addresses why you are hurting your relationship, or you can resign yourself to the failure of your relationship (and maybe the next one, and the next one, and…). Failure doesn't always mean you break up — many people aren't that lucky. But people can live quite unhappily in failed relationships for years and even decades because they're afraid they won't find anything better, or worse, they're afraid they deserve it. Don't you be one of them — if you suffer from any of these problems, figure out how to fix it, whether that means therapy, a solo mountain retreat, or just talking to your partner and committing yourselves to change. 上述问题都没有一个唯一确定的答案，但是你有不同选择：你可以尝试去给自己找一个答案，明白你为什么在伤害你俩的感情，你也可以选择退出这段失败的感情(也许还包括下一段，再下一段……)。失败的感情不一定意味着分手——很多人没有那么幸运。很多人可以一直在失败的感情中不愉快的生活很多年，甚至一辈子。因为他们害怕自己找不到更好的，更糟糕的是一些人认为这是自己应得的。如果你也是如此，不要跟他们一样，从问题中吸取教训，尝试去真正解决问题，例如寻求心理治疗，独自去远足一下，或者直接跟伴侣商量，确确实实的做一些改变……

# 避免Android开发中的ANR

## ANR是什么

ANRs (“Application Not Responding”)，意思是”应用没有响应“。

– 主线程 (“事件处理线程” / “UI线程”) 在5秒内没有响应输入事件

1、在主线程内进行网络操作

2、在主线程内进行一些缓慢的磁盘操作（例如执行没有优化过的SQL查询）

## 一些数据(Nexus One为例)

• ~0.04 ms – 通过管道进程从A->B再从B->A写一个字节；或者（从dalvik）读一个简单的/proc文件

• ~0.12 ms – 由A->B 再由B->A 进行一次Binder的RPC调用

• ~5-25 ms – 从未缓冲的flash • ~5-200+(!) ms – 向为缓冲的flash中写点东西（下面是具体数据）

•    16 ms – 60fps的视频中的一帧

•    41 ms – 24fps的视频中的一帧

• 100-200 ms – human perception of slow action

• 108/350/500/800 ms – 3G网络上ping（可变）

• ~1-6+ seconds – 通过HTTP在3G网络上获取6k的数据

private class DownloadFilesTask extends AsyncTask {
protected Long doInBackground(URL... urls) {  // on some background thread
int count = urls.length;
long totalSize = 0;
for (int i = 0; i < count; i++) {
publishProgress((int) ((i / (float) count) * 100));
}

}

protected void onProgressUpdate(Integer... progress) {  // on UI thread!
setProgressPercent(progress[0]);
}

protected void onPostExecute(Long result) {  // on UI thread!
}
}

private boolean handleWebSearchRequest(final ContentResolver cr) {
...
protected Void doInBackground(Void... unused) {
Browser.updateVisitedHistory(cr, newUrl, false);
return null;
}

}.execute()
...
return true;

}

1、必须从主线程调用，或者线程中有Handler或Looper。

• 用户退出了activity
• 系统内存不足
• 系统暂存了activity的状态留待后用
• 系统干掉了你的线程

## android.app.IntentService

Eclair(2.0, 2.1)文档中说：

“An abstract Service that serializes the handling of the Intents passed upon service start and handles them on a handler  thread.  To use this class extend it and implement onHandleIntent(Intent). The Service will automatically be  stopped when the last enqueued Intent is handled.” 有点令人困惑，因此...几乎没人用

Froyo (2.2) 的文档, 又澄清了一下....

android.app.IntentService

“IntentService is a base class for Services that handle asynchronous requests (expressed as Intents) on demand. Clients send requests  through startService(Intent) calls; the service is started as needed, handles each Intent in turn using a worker thread, and stops itself  when it runs out of work.

This 'work queue processor' pattern is commonly used to offload tasks from an application's main thread. The IntentService class exists to  simplify this pattern and take care of the mechanics. To use it, extend IntentService and implement onHandleIntent(Intent). IntentService  will receive the Intents, launch a worker thread, and stop the service as appropriate.

All requests are handled on a single worker thread -- they may take as long as necessary (and will not block the application's main loop), but  only one request will be processed at a time.”

## IntentService 的好处

• Acitivity的进程，当处理Intent的时候，会产生一个对应的Service
• Android的进程处理器现在会尽可能的不kill掉你
• 非常容易使用

public class DismissAllAlarmsService extends IntentService {
@Override public void onHandleIntent(Intent unusedIntent) {
ContentResolver resolver = getContentResolver();
...
resolver.update(uri, values, selection, null);
}
}
Intent intent = new Intent(context, DismissAllAlarmsService.class);
context.startService(intent);

## 其它技巧

2、显示一些动画，表示在处理中

3、使用进度条对话框

5、当不确定要耗时多久的时候，组合使用上述所有方法

## 总结

• 离开主线程！
• 磁盘和网络操作不是马上就能完的
• 了解sqlite在干嘛
• 进度展示很好

PS，在视频讲座中，作者还提到，Chrome团队为了避免Jank（响应超时而死掉），几乎所有的功能和任务都会在子线程里面去做。这一点也值得在Android中借鉴。

# 在Eclipse中查看Android SDK的源代码

### 链接到Eclipse

IPath android_src = new Path(AdtPlugin.getOsAbsoluteAndroidSources());

/** Returns the absolute android sources path in the sdk */
public static String getOsAbsoluteAndroidSources() {
return getOsSdkFolder() + getOsRelativeAndroidSources();
}

/** Returns the android sources path relative to the sdk folder */
public static String getOsRelativeAndroidSources() {
return AndroidConstants.FD_ANDROID_SOURCES;
}

public static final String FD_ANDROID_SOURCES = "sources";

### 解决方案 #1

SDK_PATH
|-- android.jar
+--docs/...
+--samples/...
+--sources
+--android
|      ...accounts, annotation, app, bluetooth, etc...
+--com/android/etc...
+--dalvik/...
+--java/...
+--javax/...

simon@simon-desktop:~/bin$cat make-src-tree #!/bin/bash # Author: Simon Liu , Beijing, China # Date: 2010-08-27 # License: GPL curr_dir=$PWD
dest_dir=~/workspace/src-tree
if [ ! -z "$1" ]; then dest_dir=$1
fi

rm -rf $dest_dir mkdir -p$dest_dir

for d in $(find -path .repo -prune \ -or -path .git -prune \ -or -path "*/src/com/*" -type d -print \ -or -path "*/src/org/*" -type d -print \ -or -path "*/java/com/*" -type d -print \ -or -path "*/java/org/*" -type d -print \ -or -path "*/java/android/*" -type d -print \ -or -path "*/java/javax/*" -type d -print) do sd=$(echo $d | sed 's#.*/src/$.*$/#\1/#g' | sed 's#.*/java/$.*$/#\1/#g') echo ";;$sd"

mkdir -p $dest_dir/$sd
for f in $(cd$d; find . -mindepth 0 -maxdepth 1 -type f | sed 's#^./##g'); do
echo "        $f" ln -s$curr_dir/$d/$f $dest_dir/$sd/$f done done  现在，当我再去查看Android SDK类的时候，可以看到源码了： ### 解决方案 #2 如果你实在不愿意把所有的Android源码拷贝到SDK目录里面去，也可以创建一个Eclispe User Library并把源码附加上去。 # Go 编程语言入门教程 原文：http://golang.org 翻译：刘金雨 http://log4think.com ## 介绍 本文档是关于Go编程语言基础的一个介绍性的入门教程，偏向于熟悉C或C++的读者。本文并非一份语言的完整指南，如果需要的话可以参考语言规范。读完本教程之后，你可以继续学习Effective Go，这份文档会更深入的挖掘如何使用Go语言。 此外还有一份《三日入门》的教程可供参考: 本文将会以一系列适当的程序来说明语言的一些关键特性。所有的示例程序都是可运行的（在撰写本文时），并且这些程序都会提交到版本库的/doc/progs/目录下。 程序片段都会标注上在源文件中的行号，为了清晰起见，空行前面的行号留空。 ## Hello World 先从一个最常见的开始: 05 package main 07 import fmt "fmt" // 本包实现了格式化输入输出 09 func main() { 10 fmt.Printf("Hello, world; or Καλημέρα κόσμε; or こんにちは 世界n"); 11 }  每份Go的源文件都会使用package语句声明它的包名。同时也可以通过导入其它包来使用其中定义的功能。这段代码导入了包fmt来调用我们的老朋友——现在它的开头字母是大写的，并且前面带有包名限定fmt.Printf 函数的声明使用关键字func，整个程序将会从为main包中的main函数开始（经过初始化之后）。 字符串常量可以包含Unicode字符，采用UTF-8编码（事实上，所有Go程序的源文件都是使用UTF-8编码的）。 注释的方式同C++一样：/* ... */// ... 稍后我们会继续提到print ## 编译 Go是一个编译型语言。目前有两个编译器，其中gccgo编译器采用了GCC作为后端，此外还有一系列根据其所适用的架构命名的编译器，例如6g适用于64位的x86结构，8g 适用于32位的x86结构，等等。这些编译器比gccgo运行的更快、生成的代码更加有效率。在撰写本文的时候（2009年底），他们还具有一个更加健壮的运行期系统，但是gccgo也正在迎头赶上。 下面来看看如何编译和运行程序。采用6g是这样的 $ 6g helloworld.go  编译; 中间代码位于 helloworld.6 中
$6l helloworld.6 链接; 输出至 6.out$ 6.out
Hello, world; or Καλημέρα κόσμε; or こんにちは 世界
$ gccgo的方式看起来更加传统一些。 $ gccgo helloworld.go
$a.out Hello, world; or Καλημέρα κόσμε; or こんにちは 世界$


## Echo

05    package main

07    import (
08        "os";
09        "flag";  // command line option parser
10    )

12    var omitNewline = flag.Bool("n", false, "don't print final newline")

14    const (
15        Space = " ";
16        Newline = "n";
17    )

19    func main() {
20        flag.Parse();   // Scans the arg list and sets up flags
21        var s string = "";
22        for i := 0; i < flag.NArg(); i++ {
23            if i &gt; 0 {
24                s += Space
25            }
26            s += flag.Arg(i);
27        }
28        if !*omitNewline {
29            s += Newline
30        }
31        os.Stdout.WriteString(s);
32    }


const Space = " "
const Newline = "n"


main.main中进行了参数解析，并创建了一个本地字符串类型的变量用于构造输出的内容。声明语句如下

var s string = "";


Go试图尽量保持简洁，这个声明也可以用更短的形式。因为初值是一个字符串类型的常量，没有必要再声明数据类型了，因此这个声明可以写成这样：

var s = "";


s := "";


22    for i := 0; i < flag.NArg(); i++ {


flag包会解析命令行参数，并将参数值保存在一个列表中。

Go语言中的for语句和C语言中的有几个不同之处。首先，for是唯一的循环语句，没有while语句或do语句。其次，for语句后面的三个子句不需要圆括号，但大括号是必须的。这一条对ifswitch语句同样适用。稍后还会有几个例子演示for语句的其它用法。

os.Exit(1)


os包还包含一些其他的常用功能，例如os.Args会被flag包用于访问命令行参数。

## 插播：数据类型 Types

Go支持一些常见的数据类型，例如intfloat，其值采用机器“适用”的大小来表示。也有定义了明确大小的数据类型，例如int8float64等，以及无符号整数类型，例如uintuint32等。这些都是完全不同的数据类型，即使intint32都是32位整数，但它们是不同的类型。对于表示字符串元素的类型byteuint8也是同样如此。

11    s := "hello";
12    if s[1] != 'e' { os.Exit(1) }
13    s = "good bye";
14    var p *string = &s;
15    p = "ciao";


s[0] = 'x';
(*p)[1] = 'y';


var arrayOfInt [10]int;


Slice 在 Go 程序中比数组更常见。它更灵活，并且具有引用的语义，效率也更高。其不足之处在于无法像数组一样精确控制存储方式，如果想在一个数据结构中保存一个具有 100个元素的序列，应该采用数组。

09    func sum(a []int) int {   // 返回一个整数
10        s := 0;
11        for i := 0; i < len(a); i++ {
12            s += a[i]
13        }
14        return s
15    }


19    s := sum(&[3]int{1,2,3});  // a slice of the array is passed to sum


s := sum(&[...]int{1,2,3});


s := sum([]int{1,2,3});


m := map[string]int{"one":1 , "two":2}


sum还第一次出现 了内置函数len()，用于返回元素数量。可以用于字符串、数组、slice、map、map和channel.

for i := 0; i < len(a); i++ { ... }


for i, v := range a { ... }


Go中的大多数数据类型都是值类型。对intstruct或数组的赋值会拷贝其内容。new()可以分配一个新的变量，并返回其分配的存储空间的地址。例如

type T struct { a, b int }
var t *T = new(T);


t := new(T);


Some types—maps, slices, and channels (see below)—have reference semantics. If you're holding a slice or a map and you modify its contents, other variables referencing the same underlying data will see the modification. For these three types you want to use the built-in functionmake():

m := make(map[string]int);


This statement initializes a new map ready to store entries. If you just declare the map, as in

var m map[string]int;


it creates anilreference that cannot hold anything. To use the map, you must first initialize the reference usingmake()or by assignment from an existing map.

Note thatnew(T)returns type*Twhilemake(T)returns typeT. If you (mistakenly) allocate a reference object withnew(), you receive a pointer to a nil reference, equivalent to declaring an uninitialized variable and taking its address.

Although integers come in lots of sizes in Go, integer constants do not. There are no constants like0LLor0x0UL. Instead, integer constants are evaluated as large-precision values that can overflow only when they are assigned to an integer variable with too little precision to represent the value.

const hardEight = (1 << 100) &gt;&gt; 97  // legal


There are nuances that deserve redirection to the legalese of the language specification but here are some illustrative examples:

var a uint64 = 0  // a has type uint64, value 0
a := uint64(0)    // equivalent; uses a "conversion"
i := 0x1234       // i gets default type: int
var j int = 1e6   // legal - 1000000 is representable in an int
x := 1.5          // a float
i3div2 := 3/2     // integer division - result is 1
f3div2 := 3./2.   // floating point division - result is 1.5


Conversions only work for simple cases such as convertingintsof one sign or size to another, and betweenintsandfloats, plus a few other simple cases. There are no automatic numeric conversions of any kind in Go, other than that of making constants have concrete size and type when assigned to a variable.

## An I/O Package

Next we'll look at a simple package for doing file I/O with the usual sort of open/close/read/write interface. Here's the start offile.go:

05    package file

07    import  (
08        "os";
09        "syscall";
10    )

12    type File  struct {
13        fd      int;    // file descriptor number
14        name    string; // file name at Open time
15    }


The first few lines declare the name of the package—file—and then import two packages. Theospackage hides the differences between various operating systems to give a consistent view of files and so on; here we're going to use its error handling utilities and reproduce the rudiments of its file I/O.

The other item is the low-level, externalsyscallpackage, which provides a primitive interface to the underlying operating system's calls.

Next is a type definition: thetypekeyword introduces a type declaration, in this case a data structure calledFile. To make things a little more interesting, ourFileincludes the name of the file that the file descriptor refers to.

BecauseFilestarts with a capital letter, the type is available outside the package, that is, by users of the package. In Go the rule about visibility of information is simple: if a name (of a top-level type, function, method, constant or variable, or of a structure field or method) is capitalized, users of the package may see it. Otherwise, the name and hence the thing being named is visible only inside the package in which it is declared. This is more than a convention; the rule is enforced by the compiler. In Go, the term for publicly visible names is ''exported''.

In the case ofFile, all its fields are lower case and so invisible to users, but we will soon give it some exported, upper-case methods.

First, though, here is a factory to create aFile:

17    func newFile(fd int, name string) *File {
18        if fd < 0 {
19            return nil
20        }
21        return &File{fd, name}
22    }


This returns a pointer to a newFilestructure with the file descriptor and name filled in. This code uses Go's notion of a ''composite literal'', analogous to the ones used to build maps and arrays, to construct a new heap-allocated object. We could write

n := new(File);
n.fd = fd;
n.name = name;
return n


but for simple structures likeFileit's easier to return the address of a nonce composite literal, as is done here on line 21.

We can use the factory to construct some familiar, exported variables of type*File:

24    var (
25        Stdin  = newFile(0, "/dev/stdin");
26        Stdout = newFile(1, "/dev/stdout");
27        Stderr = newFile(2, "/dev/stderr");
28    )


ThenewFilefunction was not exported because it's internal. The proper, exported factory to use isOpen:

30    func Open(name string, mode int, perm int) (file *File, err os.Error) {
31        r, e := syscall.Open(name, mode, perm);
32        if e != 0 {
33            err = os.Errno(e);
34        }
35        return newFile(r, name), err
36    }


There are a number of new things in these few lines. First,Openreturns multiple values, aFileand an error (more about errors in a moment). We declare the multi-value return as a parenthesized list of declarations; syntactically they look just like a second parameter list. The functionsyscall.Openalso has a multi-value return, which we can grab with the multi-variable declaration on line 31; it declaresrandeto hold the two values, both of typeint(although you'd have to look at thesyscallpackage to see that). Finally, line 35 returns two values: a pointer to the newFileand the error. Ifsyscall.Openfails, the file descriptorrwill be negative andNewFilewill returnnil.

About those errors: Theoslibrary includes a general notion of an error. It's a good idea to use its facility in your own interfaces, as we do here, for consistent error handling throughout Go code. InOpenwe use a conversion to translate Unix's integererrnovalue into the integer typeos.Errno, which implementsos.Error.

Now that we can buildFiles, we can write methods for them. To declare a method of a type, we define a function to have an explicit receiver of that type, placed in parentheses before the function name. Here are some methods for*File, each of which declares a receiver variablefile.

38    func (file *File) Close() os.Error {
39        if file == nil {
40            return os.EINVAL
41        }
42        e := syscall.Close(file.fd);
43        file.fd = -1;  // so it can't be closed again
44        if e != 0 {
45            return os.Errno(e);
46        }
47        return nil
48    }
50    func  (file *File) Read(b []byte) (ret int, err os.Error) {
51        if file  == nil {
52            return -1, os.EINVAL

53        }
54        r, e  := syscall.Read(file.fd, b);
55        if e != 0 {
56            err =  os.Errno(e);
57        }
58        return int(r), err
59    }

61    func  (file *File) Write(b []byte) (ret int, err os.Error) {
62        if  file == nil {
63            return -1, os.EINVAL

64        }
65        r, e := syscall.Write(file.fd, b);
66        if e != 0 {
67            err = os.Errno(e);
68        }
69        return int(r),  err
70    }

72    func (file *File) String() string {
73        return file.name
74    }


There is no implicitthisand the receiver variable must be used to access members of the structure. Methods are not declared within thestructdeclaration itself. Thestructdeclaration defines only data members. In fact, methods can be created for almost any type you name, such as an integer or array, not just forstructs. We'll see an example with arrays later.

TheStringmethod is so called because of a printing convention we'll describe later.

The methods use the public variableos.EINVALto return the (os.Errorversion of the) Unix error codeEINVAL. Theoslibrary defines a standard set of such error values.

We can now use our new package:

05    package main
07    import  (
08        "./file";
09        "fmt";
10        "os";
11    )

13    func  main() {
14        hello := []byte{'h', 'e', 'l', 'l', 'o', ',', ' ',  'w', 'o', 'r', 'l', 'd', 'n'};
15        file.Stdout.Write(hello);
16        file, err := file.Open("/does/not/exist",  0,  0);
17        if file == nil {
18            fmt.Printf("can't open file;  err=%sn",  err.String());
19            os.Exit(1);
20        }
21    }


The ''./'' in the import of ''./file'' tells the compiler to use our own package rather than something from the directory of installed packages.

Finally we can run the program:

% helloworld3
hello, world
can't open file; err=No such file or directory
%


## Rotting cats

Building on thefilepackage, here's a simple version of the Unix utilitycat(1),progs/cat.go:

05    package main

07    import (
08        "./file"
09        "flag"
10        "fmt"
11        "os"
12    )

14    func cat(f *file.File) {
15        const NBUF = 512
16        var buf [NBUF]byte
17        for {
18            switch nr, er := f.Read(&buf); true {
19            case nr < 0:
20                fmt.Fprintf(os.Stderr, "cat: error reading from %s: %s\n", f.String(), er.String())
21                os.Exit(1)
22            case nr == 0:  // EOF
23                return
24            case nr &gt; 0:
25                if nw, ew := file.Stdout.Write(buf[0:nr]); nw != nr {
26                    fmt.Fprintf(os.Stderr, "cat: error writing from %s: %s\n", f.String(), ew.String())
27                }
28            }
29        }
30    }

32    func main() {
33        flag.Parse()   // Scans the arg list and sets up flags
34        if flag.NArg() == 0 {
35            cat(file.Stdin)
36        }
37        for i := 0; i < flag.NArg(); i++ {
38            f, err := file.Open(flag.Arg(i), 0, 0)
39            if f == nil {
40                fmt.Fprintf(os.Stderr, "cat: can't open %s: error %s\n", flag.Arg(i), err)
41                os.Exit(1)
42            }
43            cat(f)
44            f.Close()
45        }
46    }


By now this should be easy to follow, but theswitchstatement introduces some new features. Like aforloop, aniforswitchcan include an initialization statement. Theswitchon line 18 uses one to create variablesnranderto hold the return values fromf.Read(). (Theifon line 25 has the same idea.) Theswitchstatement is general: it evaluates the cases from top to bottom looking for the first case that matches the value; the case expressions don't need to be constants or even integers, as long as they all have the same type.

Since theswitchvalue is justtrue, we could leave it off—as is also the situation in aforstatement, a missing value meanstrue. In fact, such aswitchis a form ofif-elsechain. While we're here, it should be mentioned that inswitchstatements eachcasehas an implicitbreak.

Line 25 callsWrite()by slicing the incoming buffer, which is itself a slice. Slices provide the standard Go way to handle I/O buffers.

Now let's make a variant ofcatthat optionally doesrot13on its input. It's easy to do by just processing the bytes, but instead we will exploit Go's notion of aninterface.

Thecat()subroutine uses only two methods off:Read()andString(), so let's start by defining an interface that has exactly those two methods. Here is code fromprogs/cat_rot13.go:

26    type reader interface {
27        Read(b []byte) (ret int, err os.Error);
28        String() string;
29    }


Any type that has the two methods ofreader—regardless of whatever other methods the type may also have—is said toimplementthe interface. Sincefile.Fileimplements these methods, it implements thereaderinterface. We could tweak thecatsubroutine to accept areaderinstead of a*file.Fileand it would work just fine, but let's embellish a little first by writing a second type that implementsreader, one that wraps an existingreaderand doesrot13on the data. To do this, we just define the type and implement the methods and with no other bookkeeping, we have a second implementation of thereaderinterface.

31    type rotate13 struct {
33    }

35    func newRotate13(source reader) *rotate13 {
36        return &rotate13{source}
37    }

39    func (r13 *rotate13) Read(b []byte) (ret int, err os.Error) {
41        for i := 0; i < r; i++ {
42            b[i] = rot13(b[i])
43        }
44        return r, e
45    }

47    func (r13 *rotate13) String() string {
48        return r13.source.String()
49    }
50    // end of rotate13 implementation


(Therot13function called on line 42 is trivial and not worth reproducing here.)

To use the new feature, we define a flag:

14    var rot13Flag = flag.Bool("rot13", false, "rot13 the input")


and use it from within a mostly unchangedcat()function:

52    func cat(r reader) {
53        const NBUF = 512
54        var buf [NBUF]byte

56        if *rot13Flag {
57            r = newRotate13(r)
58        }
59        for {
60            switch nr, er := r.Read(&buf); {
61            case nr < 0:
62                fmt.Fprintf(os.Stderr, "cat: error reading from %s: %s\n", r.String(), er.String())
63                os.Exit(1)
64            case nr == 0:  // EOF
65                return
66            case nr &gt; 0:
67                nw, ew := file.Stdout.Write(buf[0:nr])
68                if nw != nr {
69                    fmt.Fprintf(os.Stderr, "cat: error writing from %s: %s\n", r.String(), ew.String())
70                }
71            }
72        }
73    }


(We could also do the wrapping inmainand leavecat()mostly alone, except for changing the type of the argument; consider that an exercise.) Lines 56 through 58 set it all up: If therot13flag is true, wrap thereaderwe received into arotate13and proceed. Note that the interface variables are values, not pointers: the argument is of typereader, not*reader, even though under the covers it holds a pointer to astruct.

Here it is in action:

% echo abcdefghijklmnopqrstuvwxyz | ./cat
abcdefghijklmnopqrstuvwxyz
% echo abcdefghijklmnopqrstuvwxyz | ./cat --rot13
nopqrstuvwxyzabcdefghijklm
%


Fans of dependency injection may take cheer from how easily interfaces allow us to substitute the implementation of a file descriptor.

Interfaces are a distinctive feature of Go. An interface is implemented by a type if the type implements all the methods declared in the interface. This means that a type may implement an arbitrary number of different interfaces. There is no type hierarchy; things can be much moread hoc, as we saw withrot13. The typefile.Fileimplementsreader; it could also implement awriter, or any other interface built from its methods that fits the current situation. Consider theempty interface

type Empty interface {}


Everytype implements the empty interface, which makes it useful for things like containers.

## Sorting

Interfaces provide a simple form of polymorphism. They completely separate the definition of what an object does from how it does it, allowing distinct implementations to be represented at different times by the same interface variable.

As an example, consider this simple sort algorithm taken fromprogs/sort.go:

13    func Sort(data Interface) {
14        for i := 1; i < data.Len(); i++ {
15            for j := i; j &gt; 0 && data.Less(j, j-1); j-- {
16                data.Swap(j, j-1)
17            }
18        }
19    }


The code needs only three methods, which we wrap into sort'sInterface:

07    type Interface interface {
08        Len() int;
09        Less(i, j int) bool;
10        Swap(i, j int);
11    }


We can applySortto any type that implementsLen, Less, andSwap. Thesortpackage includes the necessary methods to allow sorting of arrays of integers, strings, etc.; here's the code for arrays ofint

33    type IntArray []int

35    func (p IntArray) Len() int            { return len(p) }
36    func (p IntArray) Less(i, j int) bool  { return p[i] < p[j] }
37    func (p IntArray) Swap(i, j int)       { p[i], p[j] = p[j], p[i] }


Here we see methods defined for non-structtypes. You can define methods for any type you define and name in your package.

And now a routine to test it out, fromprogs/sortmain.go. This uses a function in thesortpackage, omitted here for brevity, to test that the result is sorted.

12    func ints() {
13        data := []int{74, 59, 238, -784, 9845, 959, 905, 0, 0, 42, 7586, -5467984, 7586};
14        a := sort.IntArray(data);
15        sort.Sort(a);
16        if !sort.IsSorted(a) {
17            panic()
18        }
19    }


If we have a new type we want to be able to sort, all we need to do is to implement the three methods for that type, like this:

30    type day struct {
31        num        int
32        shortName  string
33        longName   string
34    }

36    type dayArray struct {
37        data []*day
38    }

40    func (p *dayArray) Len() int            { return len(p.data) }
41    func (p *dayArray) Less(i, j int) bool  { return p.data[i].num < p.data[j].num }
42    func (p *dayArray) Swap(i, j int)       { p.data[i], p.data[j] = p.data[j], p.data[i] }


## Printing

The examples of formatted printing so far have been modest. In this section we'll talk about how formatted I/O can be done well in Go.

We've seen simple uses of the packagefmt, which implementsPrintf, Fprintf, and so on. Within thefmtpackage,Printfis declared with this signature:

Printf(format string, v ...) (n int, errno os.Error)


That ...represents the variadic argument list that in C would be handled using thestdarg.hmacros but in Go is passed using an empty interface variable (interface {}) and then unpacked using the reflection library. It's off topic here but the use of reflection helps explain some of the nice properties of Go'sPrintf, due to the ability ofPrintfto discover the type of its arguments dynamically.

For example, in C each format must correspond to the type of its argument. It's easier in many cases in Go. Instead of%lludyou can just say%d;Printfknows the size and signedness of the integer and can do the right thing for you. The snippet

10    var u64 uint64 = 1<<64-1;
11    fmt.Printf("%d %dn", u64, int64(u64));


prints

 18446744073709551615 -1


In fact, if you're lazy the format%vwill print, in a simple appropriate style, any value, even an array or structure. The output of

14    type T struct { a int; b string };
15    t := T{77, "Sunset Strip"};
16    a := []int{1, 2, 3, 4};
17    fmt.Printf("%v %v %vn", u64, t, a);


is

18446744073709551615 {77 Sunset Strip} [1 2 3 4]


You can drop the formatting altogether if you usePrintorPrintlninstead ofPrintf. Those routines do fully automatic formatting. ThePrintfunction just prints its elements out using the equivalent of%vwhilePrintlninserts spaces between arguments and adds a newline. The output of each of these two lines is identical to that of thePrintfcall above.

18    fmt.Print(u64, " ", t, " ", a, "n");
19    fmt.Println(u64, t, a);


If you have your own type you'd likePrintforPrintto format, just give it aString()method that returns a string. The print routines will examine the value to inquire whether it implements the method and if so, use it rather than some other formatting. Here's a simple example.

09    type testType struct {
10        a int
11        b string
12    }

14    func (t *testType) String() string {
15        return fmt.Sprint(t.a) + " " + t.b
16    }

18    func main() {
19        t := &testType{77, "Sunset Strip"}
20        fmt.Println(t)
21    }


Since*testTypehas aString()method, the default formatter for that type will use it and produce the output

77 Sunset Strip


Observe that theString()method callsSprint(the obvious Go variant that returns a string) to do its formatting; special formatters can use thefmtlibrary recursively.

Another feature ofPrintfis that the format%Twill print a string representation of the type of a value, which can be handy when debugging polymorphic code.

It's possible to write full custom print formats with flags and precisions and such, but that's getting a little off the main thread so we'll leave it as an exploration exercise.

You might ask, though, howPrintfcan tell whether a type implements theString()method. Actually what it does is ask if the value can be converted to an interface variable that implements the method. Schematically, given a valuev, it does this:

type Stringer interface {
String() string
}

s, ok := v.(Stringer);  // Test whether v implements "String()"
if ok {
result = s.String()
} else {
result = defaultOutput(v)
}


The code uses a type assertion'' (v.(Stringer)) to test if the value stored invsatisfies theStringerinterface; if it does,swill become an interface variable implementing the method andokwill betrue. We then use the interface variable to call the method. (The ''comma, ok'' pattern is a Go idiom used to test the success of operations such as type conversion, map update, communications, and so on, although this is the only appearance in this tutorial.) If the value does not satisfy the interface,okwill be false.

In this snippet the nameStringerfollows the convention that we add ''[e]r'' to interfaces describing simple method sets like this.

One last wrinkle. To complete the suite, besidesPrintfetc. andSprintfetc., there are alsoFprintfetc. Unlike in C,Fprintf's first argument is not a file. Instead, it is a variable of typeio.Writer, which is an interface type defined in theiolibrary:

type Writer interface {
Write(p []byte) (n int, err os.Error);
}


(This interface is another conventional name, this time forWrite; there are alsoio.Reader,io.ReadWriter, and so on.) Thus you can callFprintfon any type that implements a standardWrite()method, not just files but also network channels, buffers, whatever you want.

## Prime numbers

Now we come to processes and communication—concurrent programming. It's a big subject so to be brief we assume some familiarity with the topic.

A classic program in the style is a prime sieve. (The sieve of Eratosthenes is computationally more efficient than the algorithm presented here, but we are more interested in concurrency than algorithmics at the moment.) It works by taking a stream of all the natural numbers and introducing a sequence of filters, one for each prime, to winnow the multiples of that prime. At each step we have a sequence of filters of the primes so far, and the next number to pop out is the next prime, which triggers the creation of the next filter in the chain.

Here's a flow diagram; each box represents a filter element whose creation is triggered by the first number that flowed from the elements before it.

To create a stream of integers, we use a Gochannel, which, borrowing from CSP's descendants, represents a communications channel that can connect two concurrent computations. In Go, channel variables are references to a run-time object that coordinates the communication; as with maps and slices, usemaketo create a new channel.

Here is the first function inprogs/sieve.go:

09    // Send the sequence 2, 3, 4, ... to channel 'ch'.
10    func generate(ch chan int) {
11        for i := 2; ; i++ {
12            ch <- i  // Send 'i' to channel 'ch'.
13        }
14    }


The generatefunction sends the sequence 2, 3, 4, 5, ... to its argument channel,ch, using the binary communications operator<-. Channel operations block, so if there's no recipient for the value onch, the send operation will wait until one becomes available.

Thefilterfunction has three arguments: an input channel, an output channel, and a prime number. It copies values from the input to the output, discarding anything divisible by the prime. The unary communications operator<-(receive) retrieves the next value on the channel.

16    // Copy the values from channel 'in' to channel 'out',
17    // removing those divisible by 'prime'.
18    func filter(in, out chan int, prime int) {
19        for {
20            i := <-in;  // Receive value of new variable 'i' from 'in'.
21            if i % prime != 0 {
22                out <- i  // Send 'i' to channel 'out'.
23            }
24        }
25    }


The generator and filters execute concurrently. Go has its own model of process/threads/light-weight processes/coroutines, so to avoid notational confusion we call concurrently executing computations in Gogoroutines. To start a goroutine, invoke the function, prefixing the call with the keywordgo; this starts the function running in parallel with the current computation but in the same address space:

go sum(hugeArray); // calculate sum in the background


If you want to know when the calculation is done, pass a channel on which it can report back:

ch := make(chan int);
go sum(hugeArray, ch);
// ... do something else for a while
result := <-ch;  // wait for, and retrieve, result


Back to our prime sieve. Here's how the sieve pipeline is stitched together:

28    func main() {
29        ch := make(chan int);  // Create a new channel.
30        go generate(ch);  // Start generate() as a goroutine.
31        for {
32            prime := <-ch;
33            fmt.Println(prime);
34            ch1 := make(chan int);
35            go filter(ch, ch1, prime);
36            ch = ch1
37        }
38    }


Line 29 creates the initial channel to pass togenerate, which it then starts up. As each prime pops out of the channel, a newfilteris added to the pipeline anditsoutput becomes the new value ofch.

The sieve program can be tweaked to use a pattern common in this style of programming. Here is a variant version ofgenerate, fromprogs/sieve1.go:

10    func generate() chan int {
11        ch := make(chan int);
12        go func(){
13            for i := 2; ; i++ {
14                ch <- i
15            }
16        }();
17        return ch;
18    }


This version does all the setup internally. It creates the output channel, launches a goroutine running a function literal, and returns the channel to the caller. It is a factory for concurrent execution, starting the goroutine and returning its connection.

The function literal notation (lines 12-16) allows us to construct an anonymous function and invoke it on the spot. Notice that the local variablechis available to the function literal and lives on even aftergeneratereturns.

The same change can be made tofilter:

21    func filter(in chan int, prime int) chan int {
22        out := make(chan int);
23        go func() {
24            for {
25                if i := <-in; i % prime != 0 {
26                    out <- i
27                }
28            }
29        }();
30        return out;
31    }


The sievefunction's main loop becomes simpler and clearer as a result, and while we're at it let's turn it into a factory too:

33    func sieve() chan int {
34        out := make(chan int);
35        go func() {
36            ch := generate();
37            for {
38                prime := <-ch;
39                out <- prime;
40                ch = filter(ch, prime);
41            }
42        }();
43        return out;
44    }


Nowmain's interface to the prime sieve is a channel of primes:

46    func main() {
47        primes := sieve();
48        for {
49            fmt.Println(<-primes);
50        }
51    }


## Multiplexing

With channels, it's possible to serve multiple independent client goroutines without writing an explicit multiplexer. The trick is to send the server a channel in the message, which it will then use to reply to the original sender. A realistic client-server program is a lot of code, so here is a very simple substitute to illustrate the idea. It starts by defining arequesttype, which embeds a channel that will be used for the reply.

09    type request struct {
10        a, b    int;
12    }


The server will be trivial: it will do simple binary operations on integers. Here's the code that invokes the operation and responds to the request:

14    type binOp func(a, b int) int

16    func run(op binOp, req *request) {
19    }


Line 18 defines the namebinOpto be a function taking two integers and returning a third.

Theserverroutine loops forever, receiving requests and, to avoid blocking due to a long-running operation, starting a goroutine to do the actual work.

21    func server(op binOp, service chan *request) {
22        for {
23            req := <-service;
24            go run(op, req);  // don't wait for it
25        }
26    }


We construct a server in a familiar way, starting it and returning a channel connected to it:

28    func startServer(op binOp) chan *request {
29        req := make(chan *request);
30        go server(op, req);
31        return req;
32    }


Here's a simple test. It starts a server with an addition operator and sends outNrequests without waiting for the replies. Only after all the requests are sent does it check the results.

34    func main() {
35        adder := startServer(func(a, b int) int { return a + b });
36        const N = 100;
37        var reqs [N]request;
38        for i := 0; i < N; i++ {
39            req := &reqs[i];
40            req.a = i;
41            req.b = i + N;
44        }
45        for i := N-1; i &gt;= 0; i-- {   // doesn't matter what order
46            if <-reqs[i].replyc != N + 2*i {
47                fmt.Println("fail at", i);
48            }
49        }
50        fmt.Println("done");
51    }


One annoyance with this program is that it doesn't shut down the server cleanly; whenmainreturns there are a number of lingering goroutines blocked on communication. To solve this, we can provide a second,quitchannel to the server:

32    func startServer(op binOp) (service chan *request, quit chan bool) {
33        service = make(chan *request);
34        quit = make(chan bool);
35        go server(op, service, quit);
36        return service, quit;
37    }


It passes the quit channel to theserverfunction, which uses it like this:

21    func server(op binOp, service chan *request, quit chan bool) {
22        for {
23            select {
24            case req := <-service:
25                go run(op, req);  // don't wait for it
26            case <-quit:
27                return;
28            }
29        }
30    }


Inside server, theselectstatement chooses which of the multiple communications listed by its cases can proceed. If all are blocked, it waits until one can proceed; if multiple can proceed, it chooses one at random. In this instance, theselectallows the server to honor requests until it receives a quit message, at which point it returns, terminating its execution.

All that's left is to strobe thequitchannel at the end of main:

42        adder, quit := startServer(func(a, b int) int { return a + b });
...
55        quit <- true;
`