diff options
Diffstat (limited to 'doc/go_tutorial.html')
-rw-r--r-- | doc/go_tutorial.html | 1455 |
1 files changed, 1455 insertions, 0 deletions
diff --git a/doc/go_tutorial.html b/doc/go_tutorial.html new file mode 100644 index 000000000..0b366bb2b --- /dev/null +++ b/doc/go_tutorial.html @@ -0,0 +1,1455 @@ +<!-- A Tutorial for the Go Programming Language --> +<h2>Introduction</h2> +<p> +This document is a tutorial introduction to the basics of the Go programming +language, intended for programmers familiar with C or C++. It is not a comprehensive +guide to the language; at the moment the document closest to that is the +<a href='/doc/go_spec.html'>language specification</a>. +After you've read this tutorial, you should look at +<a href='/doc/effective_go.html'>Effective Go</a>, +which digs deeper into how the language is used and +talks about the style and idioms of programming in Go. +Also, slides from a 3-day course about Go are available. +They provide some background and a lot of examples: +<a href='/doc/GoCourseDay1.pdf'>Day 1</a>, +<a href='/doc/GoCourseDay2.pdf'>Day 2</a>, +<a href='/doc/GoCourseDay3.pdf'>Day 3</a>. +<p> +The presentation here proceeds through a series of modest programs to illustrate +key features of the language. All the programs work (at time of writing) and are +checked into the repository in the directory <a href='/doc/progs'><code>/doc/progs/</code></a>. +<p> +<h2>Hello, World</h2> +<p> +Let's start in the usual way: +<p> +<pre><!--{{code "progs/helloworld.go" `/package/` "$"}} +-->package main + +import fmt "fmt" // Package implementing formatted I/O. + +func main() { + fmt.Printf("Hello, world; or Καλημέρα κόσμε; or こんにちは 世界\n") +} +</pre> +<p> +Every Go source file declares, using a <code>package</code> statement, which package it's part of. +It may also import other packages to use their facilities. +This program imports the package <code>fmt</code> to gain access to +our old, now capitalized and package-qualified, friend, <code>fmt.Printf</code>. +<p> +Functions are introduced with the <code>func</code> keyword. +The <code>main</code> package's <code>main</code> function is where the program starts running (after +any initialization). +<p> +String constants can contain Unicode characters, encoded in UTF-8. +(In fact, Go source files are defined to be encoded in UTF-8.) +<p> +The comment convention is the same as in C++: +<p> +<pre> +/* ... */ +// ... +</pre> +<p> +Later we'll have much more to say about printing. +<p> +<h2>Semicolons</h2> +<p> +You might have noticed that our program has no semicolons. In Go +code, the only place you typically see semicolons is separating the +clauses of <code>for</code> loops and the like; they are not necessary after +every statement. +<p> +In fact, what happens is that the formal language uses semicolons, +much as in C or Java, but they are inserted automatically +at the end of every line that looks like the end of a statement. You +don't need to type them yourself. +<p> +For details about how this is done you can see the language +specification, but in practice all you need to know is that you +never need to put a semicolon at the end of a line. (You can put +them in if you want to write multiple statements per line.) As an +extra help, you can also leave out a semicolon immediately before +a closing brace. +<p> +This approach makes for clean-looking, semicolon-free code. The +one surprise is that it's important to put the opening +brace of a construct such as an <code>if</code> statement on the same line as +the <code>if</code>; if you don't, there are situations that may not compile +or may give the wrong result. The language forces the brace style +to some extent. +<p> +<h2>Compiling</h2> +<p> +Go is a compiled language. At the moment there are two compilers. +<code>Gccgo</code> is a Go compiler that uses the GCC back end. There is also a +suite of compilers with different (and odd) names for each architecture: +<code>6g</code> for the 64-bit x86, <code>8g</code> for the 32-bit x86, and more. These +compilers run significantly faster but generate less efficient code +than <code>gccgo</code>. At the time of writing (late 2009), they also have +a more robust run-time system although <code>gccgo</code> is catching up. +<p> +Here's how to compile and run our program. With <code>6g</code>, say, +<p> +<pre> +$ 6g helloworld.go # compile; object goes into helloworld.6 +$ 6l helloworld.6 # link; output goes into 6.out +$ 6.out +Hello, world; or Καλημέρα κόσμε; or こんにちは 世界 +$ +</pre> +<p> +With <code>gccgo</code> it looks a little more traditional. +<p> +<pre> +$ gccgo helloworld.go +$ a.out +Hello, world; or Καλημέρα κόσμε; or こんにちは 世界 +$ +</pre> +<p> +<h2>Echo</h2> +<p> +Next up, here's a version of the Unix utility <code>echo(1)</code>: +<p> +<pre><!--{{code "progs/echo.go" `/package/` "$"}} +-->package main + +import ( + "os" + "flag" // command line option parser +) + +var omitNewline = flag.Bool("n", false, "don't print final newline") + +const ( + Space = " " + Newline = "\n" +) + +func main() { + flag.Parse() // Scans the arg list and sets up flags + var s string = "" + for i := 0; i < flag.NArg(); i++ { + if i > 0 { + s += Space + } + s += flag.Arg(i) + } + if !*omitNewline { + s += Newline + } + os.Stdout.WriteString(s) +} +</pre> +<p> +This program is small but it's doing a number of new things. In the last example, +we saw <code>func</code> introduce a function. The keywords <code>var</code>, <code>const</code>, and <code>type</code> +(not used yet) also introduce declarations, as does <code>import</code>. +Notice that we can group declarations of the same sort into +parenthesized lists, one item per line, as in the <code>import</code> and <code>const</code> clauses here. +But it's not necessary to do so; we could have said +<p> +<pre> +const Space = " " +const Newline = "\n" +</pre> +<p> +This program imports the <code>"os"</code> package to access its <code>Stdout</code> variable, of type +<code>*os.File</code>. The <code>import</code> statement is actually a declaration: in its general form, +as used in our ``hello world'' program, +it names the identifier (<code>fmt</code>) +that will be used to access members of the package imported from the file (<code>"fmt"</code>), +found in the current directory or in a standard location. +In this program, though, we've dropped the explicit name from the imports; by default, +packages are imported using the name defined by the imported package, +which by convention is of course the file name itself. Our ``hello world'' program +could have said just <code>import "fmt"</code>. +<p> +You can specify your +own import names if you want but it's only necessary if you need to resolve +a naming conflict. +<p> +Given <code>os.Stdout</code> we can use its <code>WriteString</code> method to print the string. +<p> +After importing the <code>flag</code> package, we use a <code>var</code> declaration +to create and initialize a global variable, called <code>omitNewline</code>, +to hold the value of echo's <code>-n</code> flag. +The variable has type <code>*bool</code>, pointer to <code>bool</code>. +<p> +In <code>main.main</code>, we parse the arguments (the call to <code>flag.Parse</code>) and then create a local +string variable with which to build the output. +<p> +The declaration statement has the form +<p> +<pre> +var s string = "" +</pre> +<p> +This is the <code>var</code> keyword, followed by the name of the variable, followed by +its type, followed by an equals sign and an initial value for the variable. +<p> +Go tries to be terse, and this declaration could be shortened. Since the +string constant is of type string, we don't have to tell the compiler that. +We could write +<p> +<pre> +var s = "" +</pre> +<p> +or we could go even shorter and write the idiom +<p> +<pre> +s := "" +</pre> +<p> +The <code>:=</code> operator is used a lot in Go to represent an initializing declaration. +There's one in the <code>for</code> clause on the next line: +<p> +<pre><!--{{code "progs/echo.go" `/for/`}} +--> for i := 0; i < flag.NArg(); i++ { +</pre> +<p> +The <code>flag</code> package has parsed the arguments and left the non-flag arguments +in a list that can be iterated over in the obvious way. +<p> +The Go <code>for</code> statement differs from that of C in a number of ways. First, +it's the only looping construct; there is no <code>while</code> or <code>do</code>. Second, +there are no parentheses on the clause, but the braces on the body +are mandatory. The same applies to the <code>if</code> and <code>switch</code> statements. +Later examples will show some other ways <code>for</code> can be written. +<p> +The body of the loop builds up the string <code>s</code> by appending (using <code>+=</code>) +the arguments and separating spaces. After the loop, if the <code>-n</code> flag is not +set, the program appends a newline. Finally, it writes the result. +<p> +Notice that <code>main.main</code> is a niladic function with no return type. +It's defined that way. Falling off the end of <code>main.main</code> means +''success''; if you want to signal an erroneous return, call +<p> +<pre> +os.Exit(1) +</pre> +<p> +The <code>os</code> package contains other essentials for getting +started; for instance, <code>os.Args</code> is a slice used by the +<code>flag</code> package to access the command-line arguments. +<p> +<h2>An Interlude about Types</h2> +<p> +Go has some familiar types such as <code>int</code> and <code>uint</code> (unsigned <code>int</code>), which represent +values of the ''appropriate'' size for the machine. It also defines +explicitly-sized types such as <code>int8</code>, <code>float64</code>, and so on, plus +unsigned integer types such as <code>uint</code>, <code>uint32</code>, etc. +These are distinct types; even if <code>int</code> and <code>int32</code> are both 32 bits in size, +they are not the same type. There is also a <code>byte</code> synonym for +<code>uint8</code>, which is the element type for strings. +<p> +Floating-point types are always sized: <code>float32</code> and <code>float64</code>, +plus <code>complex64</code> (two <code>float32s</code>) and <code>complex128</code> +(two <code>float64s</code>). Complex numbers are outside the +scope of this tutorial. +<p> +Speaking of <code>string</code>, that's a built-in type as well. Strings are +<i>immutable values</i>—they are not just arrays of <code>byte</code> values. +Once you've built a string <i>value</i>, you can't change it, although +of course you can change a string <i>variable</i> simply by +reassigning it. This snippet from <code>strings.go</code> is legal code: +<p> +<pre><!--{{code "progs/strings.go" `/hello/` `/ciao/`}} +--> s := "hello" + if s[1] != 'e' { + os.Exit(1) + } + s = "good bye" + var p *string = &s + *p = "ciao" +</pre> +<p> +However the following statements are illegal because they would modify +a <code>string</code> value: +<p> +<pre> +s[0] = 'x' +(*p)[1] = 'y' +</pre> +<p> +In C++ terms, Go strings are a bit like <code>const strings</code>, while pointers +to strings are analogous to <code>const string</code> references. +<p> +Yes, there are pointers. However, Go simplifies their use a little; +read on. +<p> +Arrays are declared like this: +<p> +<pre> +var arrayOfInt [10]int +</pre> +<p> +Arrays, like strings, are values, but they are mutable. This differs +from C, in which <code>arrayOfInt</code> would be usable as a pointer to <code>int</code>. +In Go, since arrays are values, it's meaningful (and useful) to talk +about pointers to arrays. +<p> +The size of the array is part of its type; however, one can declare +a <i>slice</i> variable to hold a reference to any array, of any size, +with the same element type. +A <i>slice +expression</i> has the form <code>a[low : high]</code>, representing +the internal array indexed from <code>low</code> through <code>high-1</code>; the resulting +slice is indexed from <code>0</code> through <code>high-low-1</code>. +In short, slices look a lot like arrays but with +no explicit size (<code>[]</code> vs. <code>[10]</code>) and they reference a segment of +an underlying, usually anonymous, regular array. Multiple slices +can share data if they represent pieces of the same array; +multiple arrays can never share data. +<p> +Slices are much more common in Go programs than +regular arrays; they're more flexible, have reference semantics, +and are efficient. What they lack is the precise control of storage +layout of a regular array; if you want to have a hundred elements +of an array stored within your structure, you should use a regular +array. To create one, use a compound value <i>constructor</i>—an +expression formed +from a type followed by a brace-bounded expression like this: +<p> +<pre> +[3]int{1,2,3} +</pre> +<p> +In this case the constructor builds an array of 3 <code>ints</code>. +<p> +When passing an array to a function, you almost always want +to declare the formal parameter to be a slice. When you call +the function, slice the array to create +(efficiently) a slice reference and pass that. +By default, the lower and upper bounds of a slice match the +ends of the existing object, so the concise notation <code>[:]</code> +will slice the whole array. +<p> +Using slices one can write this function (from <code>sum.go</code>): +<p> +<pre><!--{{code "progs/sum.go" `/sum/` `/^}/`}} +-->func sum(a []int) int { // returns an int + s := 0 + for i := 0; i < len(a); i++ { + s += a[i] + } + return s +} +</pre> +<p> +Note how the return type (<code>int</code>) is defined for <code>sum</code> by stating it +after the parameter list. +<p> +To call the function, we slice the array. This intricate call (we'll show +a simpler way in a moment) constructs +an array and slices it: +<p> +<pre> +s := sum([3]int{1,2,3}[:]) +</pre> +<p> +If you are creating a regular array but want the compiler to count the +elements for you, use <code>...</code> as the array size: +<p> +<pre> +s := sum([...]int{1,2,3}[:]) +</pre> +<p> +That's fussier than necessary, though. +In practice, unless you're meticulous about storage layout within a +data structure, a slice itself—using empty brackets with no size—is all you need: +<p> +<pre> +s := sum([]int{1,2,3}) +</pre> +<p> +There are also maps, which you can initialize like this: +<p> +<pre> +m := map[string]int{"one":1 , "two":2} +</pre> +<p> +The built-in function <code>len</code>, which returns number of elements, +makes its first appearance in <code>sum</code>. It works on strings, arrays, +slices, maps, and channels. +<p> +By the way, another thing that works on strings, arrays, slices, maps +and channels is the <code>range</code> clause on <code>for</code> loops. Instead of writing +<p> +<pre> +for i := 0; i < len(a); i++ { ... } +</pre> +<p> +to loop over the elements of a slice (or map or ...) , we could write +<p> +<pre> +for i, v := range a { ... } +</pre> +<p> +This assigns <code>i</code> to the index and <code>v</code> to the value of the successive +elements of the target of the range. See +<a href='/doc/effective_go.html'>Effective Go</a> +for more examples of its use. +<p> +<p> +<h2>An Interlude about Allocation</h2> +<p> +Most types in Go are values. If you have an <code>int</code> or a <code>struct</code> +or an array, assignment +copies the contents of the object. +To allocate a new variable, use the built-in function <code>new</code>, which +returns a pointer to the allocated storage. +<p> +<pre> +type T struct { a, b int } +var t *T = new(T) +</pre> +<p> +or the more idiomatic +<p> +<pre> +t := new(T) +</pre> +<p> +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 function <code>make</code>: +<p> +<pre> +m := make(map[string]int) +</pre> +<p> +This statement initializes a new map ready to store entries. +If you just declare the map, as in +<p> +<pre> +var m map[string]int +</pre> +<p> +it creates a <code>nil</code> reference that cannot hold anything. To use the map, +you must first initialize the reference using <code>make</code> or by assignment from an +existing map. +<p> +Note that <code>new(T)</code> returns type <code>*T</code> while <code>make(T)</code> returns type +<code>T</code>. If you (mistakenly) allocate a reference object with <code>new</code> rather than <code>make</code>, +you receive a pointer to a nil reference, equivalent to +declaring an uninitialized variable and taking its address. +<p> +<h2>An Interlude about Constants</h2> +<p> +Although integers come in lots of sizes in Go, integer constants do not. +There are no constants like <code>0LL</code> or <code>0x0UL</code>. 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. +<p> +<pre> +const hardEight = (1 << 100) >> 97 // legal +</pre> +<p> +There are nuances that deserve redirection to the legalese of the +language specification but here are some illustrative examples: +<p> +<pre> +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 float64, the default type for floating constants +i3div2 := 3/2 // integer division - result is 1 +f3div2 := 3./2. // floating-point division - result is 1.5 +</pre> +<p> +Conversions only work for simple cases such as converting <code>ints</code> of one +sign or size to another and between integers and floating-point numbers, +plus a couple of other instances outside the scope of a tutorial. +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. +<p> +<h2>An I/O Package</h2> +<p> +Next we'll look at a simple package for doing file I/O with an +open/close/read/write interface. Here's the start of <code>file.go</code>: +<p> +<pre><!--{{code "progs/file.go" `/package/` `/^}/`}} +-->package file + +import ( + "os" + "syscall" +) + +type File struct { + fd int // file descriptor number + name string // file name at Open time +} +</pre> +<p> +The first few lines declare the name of the +package—<code>file</code>—and then import two packages. The <code>os</code> +package 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. +<p> +The other item is the low-level, external <code>syscall</code> package, which provides +a primitive interface to the underlying operating system's calls. +<p> +Next is a type definition: the <code>type</code> keyword introduces a type declaration, +in this case a data structure called <code>File</code>. +To make things a little more interesting, our <code>File</code> includes the name of the file +that the file descriptor refers to. +<p> +Because <code>File</code> starts 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''. +<p> +In the case of <code>File</code>, all its fields are lower case and so invisible to users, but we +will soon give it some exported, upper-case methods. +<p> +First, though, here is a factory to create a <code>File</code>: +<p> +<pre><!--{{code "progs/file.go" `/newFile/` `/^}/`}} +-->func newFile(fd int, name string) *File { + if fd < 0 { + return nil + } + return &File{fd, name} +} +</pre> +<p> +This returns a pointer to a new <code>File</code> structure 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 +<p> +<pre> +n := new(File) +n.fd = fd +n.name = name +return n +</pre> +<p> +but for simple structures like <code>File</code> it's easier to return the address of a +composite literal, as is done here in the <code>return</code> statement from <code>newFile</code>. +<p> +We can use the factory to construct some familiar, exported variables of type <code>*File</code>: +<p> +<pre><!--{{code "progs/file.go" `/var/` `/^.$/`}} +-->var ( + Stdin = newFile(syscall.Stdin, "/dev/stdin") + Stdout = newFile(syscall.Stdout, "/dev/stdout") + Stderr = newFile(syscall.Stderr, "/dev/stderr") +) + +</pre> +<p> +The <code>newFile</code> function was not exported because it's internal. The proper, +exported factory to use is <code>OpenFile</code> (we'll explain that name in a moment): +<p> +<pre><!--{{code "progs/file.go" `/func.OpenFile/` `/^}/`}} +-->func OpenFile(name string, mode int, perm uint32) (file *File, err os.Error) { + r, e := syscall.Open(name, mode, perm) + if e != 0 { + err = os.Errno(e) + } + return newFile(r, name), err +} +</pre> +<p> +There are a number of new things in these few lines. First, <code>OpenFile</code> returns +multiple values, a <code>File</code> and 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 function +<code>syscall.Open</code> +also has a multi-value return, which we can grab with the multi-variable +declaration on the first line; it declares <code>r</code> and <code>e</code> to hold the two values, +both of type <code>int</code> (although you'd have to look at the <code>syscall</code> package +to see that). Finally, <code>OpenFile</code> returns two values: a pointer to the new <code>File</code> +and the error. If <code>syscall.Open</code> fails, the file descriptor <code>r</code> will +be negative and <code>newFile</code> will return <code>nil</code>. +<p> +About those errors: The <code>os</code> library 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. In <code>Open</code> we use a +conversion to translate Unix's integer <code>errno</code> value into the integer type +<code>os.Errno</code>, which implements <code>os.Error</code>. +<p> +Why <code>OpenFile</code> and not <code>Open</code>? To mimic Go's <code>os</code> package, which +our exercise is emulating. The <code>os</code> package takes the opportunity +to make the two commonest cases - open for read and create for +write - the simplest, just <code>Open</code> and <code>Create</code>. <code>OpenFile</code> is the +general case, analogous to the Unix system call <code>Open</code>. Here is +the implementation of our <code>Open</code> and <code>Create</code>; they're trivial +wrappers that eliminate common errors by capturing +the tricky standard arguments to open and, especially, to create a file: +<p> +<pre><!--{{code "progs/file.go" `/^const/` `/^}/`}} +-->const ( + O_RDONLY = syscall.O_RDONLY + O_RDWR = syscall.O_RDWR + O_CREATE = syscall.O_CREAT + O_TRUNC = syscall.O_TRUNC +) + +func Open(name string) (file *File, err os.Error) { + return OpenFile(name, O_RDONLY, 0) +} +</pre> +<p> +<pre><!--{{code "progs/file.go" `/func.Create/` `/^}/`}} +-->func Create(name string) (file *File, err os.Error) { + return OpenFile(name, O_RDWR|O_CREATE|O_TRUNC, 0666) +} +</pre> +<p> +Back to our main story. +Now that we can build <code>Files</code>, 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 <code>*File</code>, +each of which declares a receiver variable <code>file</code>. +<p> +<pre><!--{{code "progs/file.go" `/Close/` "$"}} +-->func (file *File) Close() os.Error { + if file == nil { + return os.EINVAL + } + e := syscall.Close(file.fd) + file.fd = -1 // so it can't be closed again + if e != 0 { + return os.Errno(e) + } + return nil +} + +func (file *File) Read(b []byte) (ret int, err os.Error) { + if file == nil { + return -1, os.EINVAL + } + r, e := syscall.Read(file.fd, b) + if e != 0 { + err = os.Errno(e) + } + return int(r), err +} + +func (file *File) Write(b []byte) (ret int, err os.Error) { + if file == nil { + return -1, os.EINVAL + } + r, e := syscall.Write(file.fd, b) + if e != 0 { + err = os.Errno(e) + } + return int(r), err +} + +func (file *File) String() string { + return file.name +} +</pre> +<p> +There is no implicit <code>this</code> and the receiver variable must be used to access +members of the structure. Methods are not declared within +the <code>struct</code> declaration itself. The <code>struct</code> declaration defines only data members. +In fact, methods can be created for almost any type you name, such as an integer or +array, not just for <code>structs</code>. We'll see an example with arrays later. +<p> +The <code>String</code> method is so called because of a printing convention we'll +describe later. +<p> +The methods use the public variable <code>os.EINVAL</code> to return the (<code>os.Error</code> +version of the) Unix error code <code>EINVAL</code>. The <code>os</code> library defines a standard +set of such error values. +<p> +We can now use our new package: +<p> +<pre><!--{{code "progs/helloworld3.go" `/package/` "$"}} +-->package main + +import ( + "./file" + "fmt" + "os" +) + +func main() { + hello := []byte("hello, world\n") + file.Stdout.Write(hello) + f, err := file.Open("/does/not/exist") + if f == nil { + fmt.Printf("can't open file; err=%s\n", err.String()) + os.Exit(1) + } +} +</pre> +<p> +The ''<code>./</code>'' in the import of ''<code>./file</code>'' tells the compiler +to use our own package rather than +something from the directory of installed packages. +(Also, ''<code>file.go</code>'' must be compiled before we can import the +package.) +<p> +Now we can compile and run the program. On Unix, this would be the result: +<p> +<pre> +$ 6g file.go # compile file package +$ 6g helloworld3.go # compile main package +$ 6l -o helloworld3 helloworld3.6 # link - no need to mention "file" +$ helloworld3 +hello, world +can't open file; err=No such file or directory +$ +</pre> +<p> +<h2>Rotting cats</h2> +<p> +Building on the <code>file</code> package, here's a simple version of the Unix utility <code>cat(1)</code>, +<code>progs/cat.go</code>: +<p> +<pre><!--{{code "progs/cat.go" `/package/` "$"}} +-->package main + +import ( + "./file" + "flag" + "fmt" + "os" +) + +func cat(f *file.File) { + const NBUF = 512 + var buf [NBUF]byte + for { + switch nr, er := f.Read(buf[:]); true { + case nr < 0: + fmt.Fprintf(os.Stderr, "cat: error reading from %s: %s\n", f.String(), er.String()) + os.Exit(1) + case nr == 0: // EOF + return + case nr > 0: + if nw, ew := file.Stdout.Write(buf[0:nr]); nw != nr { + fmt.Fprintf(os.Stderr, "cat: error writing from %s: %s\n", f.String(), ew.String()) + os.Exit(1) + } + } + } +} + +func main() { + flag.Parse() // Scans the arg list and sets up flags + if flag.NArg() == 0 { + cat(file.Stdin) + } + for i := 0; i < flag.NArg(); i++ { + f, err := file.Open(flag.Arg(i)) + if f == nil { + fmt.Fprintf(os.Stderr, "cat: can't open %s: error %s\n", flag.Arg(i), err) + os.Exit(1) + } + cat(f) + f.Close() + } +} +</pre> +<p> +By now this should be easy to follow, but the <code>switch</code> statement introduces some +new features. Like a <code>for</code> loop, an <code>if</code> or <code>switch</code> can include an +initialization statement. The <code>switch</code> statement in <code>cat</code> uses one to create variables +<code>nr</code> and <code>er</code> to hold the return values from the call to <code>f.Read</code>. (The <code>if</code> a few lines later +has the same idea.) The <code>switch</code> statement 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. +<p> +Since the <code>switch</code> value is just <code>true</code>, we could leave it off—as is also +the situation +in a <code>for</code> statement, a missing value means <code>true</code>. In fact, such a <code>switch</code> +is a form of <code>if-else</code> chain. While we're here, it should be mentioned that in +<code>switch</code> statements each <code>case</code> has an implicit <code>break</code>. +<p> +The argument to <code>file.Stdout.Write</code> is created by slicing the array <code>buf</code>. +Slices provide the standard Go way to handle I/O buffers. +<p> +Now let's make a variant of <code>cat</code> that optionally does <code>rot13</code> on its input. +It's easy to do by just processing the bytes, but instead we will exploit +Go's notion of an <i>interface</i>. +<p> +The <code>cat</code> subroutine uses only two methods of <code>f</code>: <code>Read</code> and <code>String</code>, +so let's start by defining an interface that has exactly those two methods. +Here is code from <code>progs/cat_rot13.go</code>: +<p> +<pre><!--{{code "progs/cat_rot13.go" `/type.reader/` `/^}/`}} +-->type reader interface { + Read(b []byte) (ret int, err os.Error) + String() string +} +</pre> +<p> +Any type that has the two methods of <code>reader</code>—regardless of whatever +other methods the type may also have—is said to <i>implement</i> the +interface. Since <code>file.File</code> implements these methods, it implements the +<code>reader</code> interface. We could tweak the <code>cat</code> subroutine to accept a <code>reader</code> +instead of a <code>*file.File</code> and it would work just fine, but let's embellish a little +first by writing a second type that implements <code>reader</code>, one that wraps an +existing <code>reader</code> and does <code>rot13</code> on 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 the <code>reader</code> interface. +<p> +<pre><!--{{code "progs/cat_rot13.go" `/type.rotate13/` `/end.of.rotate13/`}} +-->type rotate13 struct { + source reader +} + +func newRotate13(source reader) *rotate13 { + return &rotate13{source} +} + +func (r13 *rotate13) Read(b []byte) (ret int, err os.Error) { + r, e := r13.source.Read(b) + for i := 0; i < r; i++ { + b[i] = rot13(b[i]) + } + return r, e +} + +func (r13 *rotate13) String() string { + return r13.source.String() +} +// end of rotate13 implementation +</pre> +<p> +(The <code>rot13</code> function called in <code>Read</code> is trivial and not worth reproducing here.) +<p> +To use the new feature, we define a flag: +<p> +<pre><!--{{code "progs/cat_rot13.go" `/rot13Flag/`}} +-->var rot13Flag = flag.Bool("rot13", false, "rot13 the input") +</pre> +<p> +and use it from within a mostly unchanged <code>cat</code> function: +<p> +<pre><!--{{code "progs/cat_rot13.go" `/func.cat/` `/^}/`}} +-->func cat(r reader) { + const NBUF = 512 + var buf [NBUF]byte + + if *rot13Flag { + r = newRotate13(r) + } + for { + switch nr, er := r.Read(buf[:]); { + case nr < 0: + fmt.Fprintf(os.Stderr, "cat: error reading from %s: %s\n", r.String(), er.String()) + os.Exit(1) + case nr == 0: // EOF + return + case nr > 0: + nw, ew := file.Stdout.Write(buf[0:nr]) + if nw != nr { + fmt.Fprintf(os.Stderr, "cat: error writing from %s: %s\n", r.String(), ew.String()) + os.Exit(1) + } + } + } +} +</pre> +<p> +(We could also do the wrapping in <code>main</code> and leave <code>cat</code> mostly alone, except +for changing the type of the argument; consider that an exercise.) +The <code>if</code> at the top of <code>cat</code> sets it all up: If the <code>rot13</code> flag is true, wrap the <code>reader</code> +we received into a <code>rotate13</code> and proceed. Note that the interface variables +are values, not pointers: the argument is of type <code>reader</code>, not <code>*reader</code>, +even though under the covers it holds a pointer to a <code>struct</code>. +<p> +Here it is in action: +<p> +<pre> +$ echo abcdefghijklmnopqrstuvwxyz | ./cat +abcdefghijklmnopqrstuvwxyz +$ echo abcdefghijklmnopqrstuvwxyz | ./cat --rot13 +nopqrstuvwxyzabcdefghijklm +$ +</pre> +<p> +Fans of dependency injection may take cheer from how easily interfaces +allow us to substitute the implementation of a file descriptor. +<p> +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 more <i>ad hoc</i>, +as we saw with <code>rot13</code>. The type <code>file.File</code> implements <code>reader</code>; it could also +implement a <code>writer</code>, or any other interface built from its methods that +fits the current situation. Consider the <i>empty interface</i> +<p> +<pre> +type Empty interface {} +</pre> +<p> +<i>Every</i> type implements the empty interface, which makes it +useful for things like containers. +<p> +<h2>Sorting</h2> +<p> +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. +<p> +As an example, consider this simple sort algorithm taken from <code>progs/sort.go</code>: +<p> +<pre><!--{{code "progs/sort.go" `/func.Sort/` `/^}/`}} +-->func Sort(data Interface) { + for i := 1; i < data.Len(); i++ { + for j := i; j > 0 && data.Less(j, j-1); j-- { + data.Swap(j, j-1) + } + } +} +</pre> +<p> +The code needs only three methods, which we wrap into sort's <code>Interface</code>: +<p> +<pre><!--{{code "progs/sort.go" `/interface/` `/^}/`}} +-->type Interface interface { + Len() int + Less(i, j int) bool + Swap(i, j int) +} +</pre> +<p> +We can apply <code>Sort</code> to any type that implements <code>Len</code>, <code>Less</code>, and <code>Swap</code>. +The <code>sort</code> package includes the necessary methods to allow sorting of +arrays of integers, strings, etc.; here's the code for arrays of <code>int</code> +<p> +<pre><!--{{code "progs/sort.go" `/type.*IntSlice/` `/Swap/`}} +-->type IntSlice []int + +func (p IntSlice) Len() int { return len(p) } +func (p IntSlice) Less(i, j int) bool { return p[i] < p[j] } +func (p IntSlice) Swap(i, j int) { p[i], p[j] = p[j], p[i] } +</pre> +<p> +Here we see methods defined for non-<code>struct</code> types. You can define methods +for any type you define and name in your package. +<p> +And now a routine to test it out, from <code>progs/sortmain.go</code>. This +uses a function in the <code>sort</code> package, omitted here for brevity, +to test that the result is sorted. +<p> +<pre><!--{{code "progs/sortmain.go" `/func.ints/` `/^}/`}} +-->func ints() { + data := []int{74, 59, 238, -784, 9845, 959, 905, 0, 0, 42, 7586, -5467984, 7586} + a := sort.IntSlice(data) + sort.Sort(a) + if !sort.IsSorted(a) { + panic("fail") + } +} +</pre> +<p> +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: +<p> +<pre><!--{{code "progs/sortmain.go" `/type.day/` `/Swap/`}} +-->type day struct { + num int + shortName string + longName string +} + +type dayArray struct { + data []*day +} + +func (p *dayArray) Len() int { return len(p.data) } +func (p *dayArray) Less(i, j int) bool { return p.data[i].num < p.data[j].num } +func (p *dayArray) Swap(i, j int) { p.data[i], p.data[j] = p.data[j], p.data[i] } +</pre> +<p> +<p> +<h2>Printing</h2> +<p> +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. +<p> +We've seen simple uses of the package <code>fmt</code>, which +implements <code>Printf</code>, <code>Fprintf</code>, and so on. +Within the <code>fmt</code> package, <code>Printf</code> is declared with this signature: +<p> +<pre> +Printf(format string, v ...interface{}) (n int, errno os.Error) +</pre> +<p> +The token <code>...</code> introduces a variable-length argument list that in C would +be handled using the <code>stdarg.h</code> macros. +In Go, variadic functions are passed a slice of the arguments of the +specified type. In <code>Printf</code>'s case, the declaration says <code>...interface{}</code> +so the actual type is a slice of empty interface values, <code>[]interface{}</code>. +<code>Printf</code> can examine the arguments by iterating over the slice +and, for each element, using a type switch or the reflection library +to interpret the value. +It's off topic here but such run-time type analysis +helps explain some of the nice properties of Go's <code>Printf</code>, +due to the ability of <code>Printf</code> to discover the type of its arguments +dynamically. +<p> +For example, in C each format must correspond to the type of its +argument. It's easier in many cases in Go. Instead of <code>%llud</code> you +can just say <code>%d</code>; <code>Printf</code> knows the size and signedness of the +integer and can do the right thing for you. The snippet +<p> +<pre><!--{{code "progs/print.go" 10 11}} +--> var u64 uint64 = 1<<64 - 1 + fmt.Printf("%d %d\n", u64, int64(u64)) +</pre> +<p> +prints +<p> +<pre> +18446744073709551615 -1 +</pre> +<p> +In fact, if you're lazy the format <code>%v</code> will print, in a simple +appropriate style, any value, even an array or structure. The output of +<p> +<pre><!--{{code "progs/print.go" 14 20}} +--> type T struct { + a int + b string + } + t := T{77, "Sunset Strip"} + a := []int{1, 2, 3, 4} + fmt.Printf("%v %v %v\n", u64, t, a) +</pre> +<p> +is +<p> +<pre> +18446744073709551615 {77 Sunset Strip} [1 2 3 4] +</pre> +<p> +You can drop the formatting altogether if you use <code>Print</code> or <code>Println</code> +instead of <code>Printf</code>. Those routines do fully automatic formatting. +The <code>Print</code> function just prints its elements out using the equivalent +of <code>%v</code> while <code>Println</code> inserts spaces between arguments +and adds a newline. The output of each of these two lines is identical +to that of the <code>Printf</code> call above. +<p> +<pre><!--{{code "progs/print.go" 21 22}} +--> fmt.Print(u64, " ", t, " ", a, "\n") + fmt.Println(u64, t, a) +</pre> +<p> +If you have your own type you'd like <code>Printf</code> or <code>Print</code> to format, +just give it a <code>String</code> 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. +<p> +<pre><!--{{code "progs/print_string.go" 9 "$"}} +-->type testType struct { + a int + b string +} + +func (t *testType) String() string { + return fmt.Sprint(t.a) + " " + t.b +} + +func main() { + t := &testType{77, "Sunset Strip"} + fmt.Println(t) +} +</pre> +<p> +Since <code>*testType</code> has a <code>String</code> method, the +default formatter for that type will use it and produce the output +<p> +<pre> +77 Sunset Strip +</pre> +<p> +Observe that the <code>String</code> method calls <code>Sprint</code> (the obvious Go +variant that returns a string) to do its formatting; special formatters +can use the <code>fmt</code> library recursively. +<p> +Another feature of <code>Printf</code> is that the format <code>%T</code> will print a string +representation of the type of a value, which can be handy when debugging +polymorphic code. +<p> +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. +<p> +You might ask, though, how <code>Printf</code> can tell whether a type implements +the <code>String</code> method. Actually what it does is ask if the value can +be converted to an interface variable that implements the method. +Schematically, given a value <code>v</code>, it does this: +<p> +<p> +<pre> +type Stringer interface { + String() string +} +</pre> +<p> +<pre> +s, ok := v.(Stringer) // Test whether v implements "String()" +if ok { + result = s.String() +} else { + result = defaultOutput(v) +} +</pre> +<p> +The code uses a ``type assertion'' (<code>v.(Stringer)</code>) to test if the value stored in +<code>v</code> satisfies the <code>Stringer</code> interface; if it does, <code>s</code> +will become an interface variable implementing the method and <code>ok</code> will +be <code>true</code>. 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, <code>ok</code> will be false. +<p> +In this snippet the name <code>Stringer</code> follows the convention that we add ''[e]r'' +to interfaces describing simple method sets like this. +<p> +One last wrinkle. To complete the suite, besides <code>Printf</code> etc. and <code>Sprintf</code> +etc., there are also <code>Fprintf</code> etc. Unlike in C, <code>Fprintf</code>'s first argument is +not a file. Instead, it is a variable of type <code>io.Writer</code>, which is an +interface type defined in the <code>io</code> library: +<p> +<pre> +type Writer interface { + Write(p []byte) (n int, err os.Error) +} +</pre> +<p> +(This interface is another conventional name, this time for <code>Write</code>; there are also +<code>io.Reader</code>, <code>io.ReadWriter</code>, and so on.) +Thus you can call <code>Fprintf</code> on any type that implements a standard <code>Write</code> +method, not just files but also network channels, buffers, whatever +you want. +<p> +<h2>Prime numbers</h2> +<p> +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. +<p> +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. +<p> +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. +<p> +<br> +<p> + <img src='sieve.gif'> +<p> +<br> +<p> +To create a stream of integers, we use a Go <i>channel</i>, 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, use +<code>make</code> to create a new channel. +<p> +Here is the first function in <code>progs/sieve.go</code>: +<p> +<pre><!--{{code "progs/sieve.go" `/Send/` `/^}/`}} +-->// Send the sequence 2, 3, 4, ... to channel 'ch'. +func generate(ch chan int) { + for i := 2; ; i++ { + ch <- i // Send 'i' to channel 'ch'. + } +} +</pre> +<p> +The <code>generate</code> function sends the sequence 2, 3, 4, 5, ... to its +argument channel, <code>ch</code>, using the binary communications operator <code><-</code>. +Channel operations block, so if there's no recipient for the value on <code>ch</code>, +the send operation will wait until one becomes available. +<p> +The <code>filter</code> function 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 <code><-</code> (receive) retrieves the next value on the channel. +<p> +<pre><!--{{code "progs/sieve.go" `/Copy.the/` `/^}/`}} +-->// Copy the values from channel 'in' to channel 'out', +// removing those divisible by 'prime'. +func filter(in, out chan int, prime int) { + for { + i := <-in // Receive value of new variable 'i' from 'in'. + if i%prime != 0 { + out <- i // Send 'i' to channel 'out'. + } + } +} +</pre> +<p> +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 Go <i>goroutines</i>. To start a goroutine, +invoke the function, prefixing the call with the keyword <code>go</code>; +this starts the function running in parallel with the current +computation but in the same address space: +<p> +<pre> +go sum(hugeArray) // calculate sum in the background +</pre> +<p> +If you want to know when the calculation is done, pass a channel +on which it can report back: +<p> +<pre> +ch := make(chan int) +go sum(hugeArray, ch) +// ... do something else for a while +result := <-ch // wait for, and retrieve, result +</pre> +<p> +Back to our prime sieve. Here's how the sieve pipeline is stitched +together: +<p> +<pre><!--{{code "progs/sieve.go" `/func.main/` `/^}/`}} +-->func main() { + ch := make(chan int) // Create a new channel. + go generate(ch) // Start generate() as a goroutine. + for i := 0; i < 100; i++ { // Print the first hundred primes. + prime := <-ch + fmt.Println(prime) + ch1 := make(chan int) + go filter(ch, ch1, prime) + ch = ch1 + } +} +</pre> +<p> +The first line of <code>main</code> creates the initial channel to pass to <code>generate</code>, which it +then starts up. As each prime pops out of the channel, a new <code>filter</code> +is added to the pipeline and <i>its</i> output becomes the new value +of <code>ch</code>. +<p> +The sieve program can be tweaked to use a pattern common +in this style of programming. Here is a variant version +of <code>generate</code>, from <code>progs/sieve1.go</code>: +<p> +<pre><!--{{code "progs/sieve1.go" `/func.generate/` `/^}/`}} +-->func generate() chan int { + ch := make(chan int) + go func() { + for i := 2; ; i++ { + ch <- i + } + }() + return ch +} +</pre> +<p> +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. +<p> +The function literal notation used in the <code>go</code> statement allows us to construct an +anonymous function and invoke it on the spot. Notice that the local +variable <code>ch</code> is available to the function literal and lives on even +after <code>generate</code> returns. +<p> +The same change can be made to <code>filter</code>: +<p> +<pre><!--{{code "progs/sieve1.go" `/func.filter/` `/^}/`}} +-->func filter(in chan int, prime int) chan int { + out := make(chan int) + go func() { + for { + if i := <-in; i%prime != 0 { + out <- i + } + } + }() + return out +} +</pre> +<p> +The <code>sieve</code> function's main loop becomes simpler and clearer as a +result, and while we're at it let's turn it into a factory too: +<p> +<pre><!--{{code "progs/sieve1.go" `/func.sieve/` `/^}/`}} +-->func sieve() chan int { + out := make(chan int) + go func() { + ch := generate() + for { + prime := <-ch + out <- prime + ch = filter(ch, prime) + } + }() + return out +} +</pre> +<p> +Now <code>main</code>'s interface to the prime sieve is a channel of primes: +<p> +<pre><!--{{code "progs/sieve1.go" `/func.main/` `/^}/`}} +-->func main() { + primes := sieve() + for i := 0; i < 100; i++ { // Print the first hundred primes. + fmt.Println(<-primes) + } +} +</pre> +<p> +<h2>Multiplexing</h2> +<p> +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 a <code>request</code> type, which embeds a channel +that will be used for the reply. +<p> +<pre><!--{{code "progs/server.go" `/type.request/` `/^}/`}} +-->type request struct { + a, b int + replyc chan int +} +</pre> +<p> +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: +<p> +<pre><!--{{code "progs/server.go" `/type.binOp/` `/^}/`}} +-->type binOp func(a, b int) int + +func run(op binOp, req *request) { + reply := op(req.a, req.b) + req.replyc <- reply +} +</pre> +<p> +The type declaration makes <code>binOp</code> represent a function taking two integers and +returning a third. +<p> +The <code>server</code> routine loops forever, receiving requests and, to avoid blocking due to +a long-running operation, starting a goroutine to do the actual work. +<p> +<pre><!--{{code "progs/server.go" `/func.server/` `/^}/`}} +-->func server(op binOp, service chan *request) { + for { + req := <-service + go run(op, req) // don't wait for it + } +} +</pre> +<p> +We construct a server in a familiar way, starting it and returning a channel +connected to it: +<p> +<pre><!--{{code "progs/server.go" `/func.startServer/` `/^}/`}} +-->func startServer(op binOp) chan *request { + req := make(chan *request) + go server(op, req) + return req +} +</pre> +<p> +Here's a simple test. It starts a server with an addition operator and sends out +<code>N</code> requests without waiting for the replies. Only after all the requests are sent +does it check the results. +<p> +<pre><!--{{code "progs/server.go" `/func.main/` `/^}/`}} +-->func main() { + adder := startServer(func(a, b int) int { return a + b }) + const N = 100 + var reqs [N]request + for i := 0; i < N; i++ { + req := &reqs[i] + req.a = i + req.b = i + N + req.replyc = make(chan int) + adder <- req + } + for i := N - 1; i >= 0; i-- { // doesn't matter what order + if <-reqs[i].replyc != N+2*i { + fmt.Println("fail at", i) + } + } + fmt.Println("done") +} +</pre> +<p> +One annoyance with this program is that it doesn't shut down the server cleanly; when <code>main</code> returns +there are a number of lingering goroutines blocked on communication. To solve this, +we can provide a second, <code>quit</code> channel to the server: +<p> +<pre><!--{{code "progs/server1.go" `/func.startServer/` `/^}/`}} +-->func startServer(op binOp) (service chan *request, quit chan bool) { + service = make(chan *request) + quit = make(chan bool) + go server(op, service, quit) + return service, quit +} +</pre> +<p> +It passes the quit channel to the <code>server</code> function, which uses it like this: +<p> +<pre><!--{{code "progs/server1.go" `/func.server/` `/^}/`}} +-->func server(op binOp, service chan *request, quit chan bool) { + for { + select { + case req := <-service: + go run(op, req) // don't wait for it + case <-quit: + return + } + } +} +</pre> +<p> +Inside <code>server</code>, the <code>select</code> statement 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, the <code>select</code> allows +the server to honor requests until it receives a quit message, at which point it +returns, terminating its execution. +<p> +<p> +All that's left is to strobe the <code>quit</code> channel +at the end of main: +<p> +<pre><!--{{code "progs/server1.go" `/adder,.quit/`}} +--> adder, quit := startServer(func(a, b int) int { return a + b }) +</pre> +... +<pre><!--{{code "progs/server1.go" `/quit....true/`}} +--> quit <- true +</pre> +<p> +There's a lot more to Go programming and concurrent programming in general but this +quick tour should give you some of the basics. |