Golang：从net.Conn中读取单个XML文档

Question

I have a client-server connection. They communicate through xml and multiple xml documents are transmitted during one session. xml.Unmarshal wants a slice of bytes, but I can't just ReadAll all bytes from socket (it will try to read more then a single xml and therefore block).

Is there a standard xml parser or a library, that can parse from a stream of bytes and read no more data than it actually needs?

Answer 1

You may use xml.Decoder from the standard library for this purpose. You may use xml.NewDecoder() to create a new xml.Decoder which expects an io.Reader to read the data from. net.Conn qualifies as it implements io.Reader. The Decoder.Decode() method will read and process 1 XML document.

Let's see an example. The source will contain 2 XML documents concatenated, and we call Decoder.Decode() twice to read and parse those 2 documents.

The XML source: 2 XML documents (2 <Person>):

const data = `<Person>
    <Name>Bob</Name>
    <Age>23</Age>
</Person>
<Person>
    <Name>Alice</Name>
    <Age>21</Age>
</Person>
`

Go struct to model the XML documents:

type Person struct {
    Name string
    Age  int
}

Code to read those 2 XML documents:

buf := bytes.NewBuffer([]byte(data))
d := xml.NewDecoder(buf)

for i := 0; i < 2; i++ {
    p := Person{}
    if err := d.Decode(&p); err != nil {
        fmt.Println(err)
    } else {
        fmt.Printf("%+v
", p)
    }
}

Output (try it on the Go Playground):

{Name:Bob Age:23}
{Name:Alice Age:21}

Note that Decoder.Decode() will return io.EOF if no more data is available. To read all XML documents from the input, you may do it like this:

for {
    p := Person{}
    if err := d.Decode(&p); err != nil {
        if err == io.EOF {
            fmt.Println("EOF, breaking")
            break
        }
        fmt.Println(err)
    } else {
        fmt.Printf("%+v
", p)
    }
}

Back to your example

If you want to read XML documents transmitted over a TCP connection, you may simply pass the net.Conn value (which implements io.Reader) to xml.NewDecoder():

var con net.Conn
// Initialize / obtain connection

d := xml.NewDecoder(con)
var doc YourDocType
if err := d.Decode(&doc); err != nil {
    // Handle error
    return
}
// No error, use doc:
fmt.Printf("%+v", doc)