This commit is contained in:
程广 2024-11-07 22:48:55 +08:00
commit 95b872aa61
13 changed files with 9982 additions and 0 deletions

56
atom_test.go Normal file
View File

@ -0,0 +1,56 @@
// Copyright 2011 The Go Authors. All rights reserved.
// Use of this source code is governed by a BSD-style
// license that can be found in the LICENSE file.
package goxml
import "time"
var atomValue = &Feed{
XMLName: Name{"http://www.w3.org/2005/Atom", "feed"},
Title: "Example Feed",
Link: []Link{{Href: "http://example.org/"}},
Updated: ParseTime("2003-12-13T18:30:02Z"),
Author: Person{Name: "John Doe"},
ID: "urn:uuid:60a76c80-d399-11d9-b93C-0003939e0af6",
Entry: []Entry{
{
Title: "Atom-Powered Robots Run Amok",
Link: []Link{{Href: "http://example.org/2003/12/13/atom03"}},
ID: "urn:uuid:1225c695-cfb8-4ebb-aaaa-80da344efa6a",
Updated: ParseTime("2003-12-13T18:30:02Z"),
Summary: NewText("Some text."),
},
},
}
var atomXML = `` +
`<feed xmlns="http://www.w3.org/2005/Atom" updated="2003-12-13T18:30:02Z">` +
`<title>Example Feed</title>` +
`<id>urn:uuid:60a76c80-d399-11d9-b93C-0003939e0af6</id>` +
`<link href="http://example.org/"></link>` +
`<author><name>John Doe</name><uri></uri><email></email></author>` +
`<entry>` +
`<title>Atom-Powered Robots Run Amok</title>` +
`<id>urn:uuid:1225c695-cfb8-4ebb-aaaa-80da344efa6a</id>` +
`<link href="http://example.org/2003/12/13/atom03"></link>` +
`<updated>2003-12-13T18:30:02Z</updated>` +
`<author><name></name><uri></uri><email></email></author>` +
`<summary>Some text.</summary>` +
`</entry>` +
`</feed>`
func ParseTime(str string) time.Time {
t, err := time.Parse(time.RFC3339, str)
if err != nil {
panic(err)
}
return t
}
func NewText(text string) Text {
return Text{
Body: text,
}
}

View File

@ -0,0 +1,85 @@
// Copyright 2018 The Go Authors. All rights reserved.
// Use of this source code is governed by a BSD-style
// license that can be found in the LICENSE file.
package goxml_test
import (
"fmt"
"log"
"strings"
"git.pyer.club/kingecg/goxml"
)
type Animal int
const (
Unknown Animal = iota
Gopher
Zebra
)
func (a *Animal) UnmarshalXML(d *goxml.Decoder, start goxml.StartElement) error {
var s string
if err := d.DecodeElement(&s, &start); err != nil {
return err
}
switch strings.ToLower(s) {
default:
*a = Unknown
case "gopher":
*a = Gopher
case "zebra":
*a = Zebra
}
return nil
}
func (a Animal) MarshalXML(e *goxml.Encoder, start goxml.StartElement) error {
var s string
switch a {
default:
s = "unknown"
case Gopher:
s = "gopher"
case Zebra:
s = "zebra"
}
return e.EncodeElement(s, start)
}
func Example_customMarshalXML() {
blob := `
<animals>
<animal>gopher</animal>
<animal>armadillo</animal>
<animal>zebra</animal>
<animal>unknown</animal>
<animal>gopher</animal>
<animal>bee</animal>
<animal>gopher</animal>
<animal>zebra</animal>
</animals>`
var zoo struct {
Animals []Animal `xml:"animal"`
}
if err := goxml.Unmarshal([]byte(blob), &zoo); err != nil {
log.Fatal(err)
}
census := make(map[Animal]int)
for _, animal := range zoo.Animals {
census[animal] += 1
}
fmt.Printf("Zoo Census:\n* Gophers: %d\n* Zebras: %d\n* Unknown: %d\n",
census[Gopher], census[Zebra], census[Unknown])
// Output:
// Zoo Census:
// * Gophers: 3
// * Zebras: 2
// * Unknown: 3
}

152
example_test.go Normal file
View File

@ -0,0 +1,152 @@
// Copyright 2012 The Go Authors. All rights reserved.
// Use of this source code is governed by a BSD-style
// license that can be found in the LICENSE file.
package goxml_test
import (
"fmt"
"os"
"git.pyer.club/kingecg/goxml"
)
func ExampleMarshalIndent() {
type Address struct {
City, State string
}
type Person struct {
XMLName goxml.Name `xml:"person"`
Id int `xml:"id,attr"`
FirstName string `xml:"name>first"`
LastName string `xml:"name>last"`
Age int `xml:"age"`
Height float32 `xml:"height,omitempty"`
Married bool
Address
Comment string `xml:",comment"`
}
v := &Person{Id: 13, FirstName: "John", LastName: "Doe", Age: 42}
v.Comment = " Need more details. "
v.Address = Address{"Hanga Roa", "Easter Island"}
output, err := goxml.MarshalIndent(v, " ", " ")
if err != nil {
fmt.Printf("error: %v\n", err)
}
os.Stdout.Write(output)
// Output:
// <person id="13">
// <name>
// <first>John</first>
// <last>Doe</last>
// </name>
// <age>42</age>
// <Married>false</Married>
// <City>Hanga Roa</City>
// <State>Easter Island</State>
// <!-- Need more details. -->
// </person>
}
func ExampleEncoder() {
type Address struct {
City, State string
}
type Person struct {
XMLName goxml.Name `xml:"person"`
Id int `xml:"id,attr"`
FirstName string `xml:"name>first"`
LastName string `xml:"name>last"`
Age int `xml:"age"`
Height float32 `xml:"height,omitempty"`
Married bool
Address
Comment string `xml:",comment"`
}
v := &Person{Id: 13, FirstName: "John", LastName: "Doe", Age: 42}
v.Comment = " Need more details. "
v.Address = Address{"Hanga Roa", "Easter Island"}
enc := goxml.NewEncoder(os.Stdout)
enc.Indent(" ", " ")
if err := enc.Encode(v); err != nil {
fmt.Printf("error: %v\n", err)
}
// Output:
// <person id="13">
// <name>
// <first>John</first>
// <last>Doe</last>
// </name>
// <age>42</age>
// <Married>false</Married>
// <City>Hanga Roa</City>
// <State>Easter Island</State>
// <!-- Need more details. -->
// </person>
}
// This example demonstrates unmarshaling an XML excerpt into a value with
// some preset fields. Note that the Phone field isn't modified and that
// the XML <Company> element is ignored. Also, the Groups field is assigned
// considering the element path provided in its tag.
func ExampleUnmarshal() {
type Email struct {
Where string `xml:"where,attr"`
Addr string
}
type Address struct {
City, State string
}
type Result struct {
XMLName goxml.Name `xml:"Person"`
Name string `xml:"FullName"`
Phone string
Email []Email
Groups []string `xml:"Group>Value"`
Address
}
v := Result{Name: "none", Phone: "none"}
data := `
<Person>
<FullName>Grace R. Emlin</FullName>
<Company>Example Inc.</Company>
<Email where="home">
<Addr>gre@example.com</Addr>
</Email>
<Email where='work'>
<Addr>gre@work.com</Addr>
</Email>
<Group>
<Value>Friends</Value>
<Value>Squash</Value>
</Group>
<City>Hanga Roa</City>
<State>Easter Island</State>
</Person>
`
err := goxml.Unmarshal([]byte(data), &v)
if err != nil {
fmt.Printf("error: %v", err)
return
}
fmt.Printf("XMLName: %#v\n", v.XMLName)
fmt.Printf("Name: %q\n", v.Name)
fmt.Printf("Phone: %q\n", v.Phone)
fmt.Printf("Email: %v\n", v.Email)
fmt.Printf("Groups: %v\n", v.Groups)
fmt.Printf("Address: %v\n", v.Address)
// Output:
// XMLName: goxml.Name{Space:"", Local:"Person"}
// Name: "Grace R. Emlin"
// Phone: "none"
// Email: [{home gre@example.com} {work gre@work.com}]
// Groups: [Friends Squash]
// Address: {Hanga Roa Easter Island}
}

View File

@ -0,0 +1,80 @@
// Copyright 2018 The Go Authors. All rights reserved.
// Use of this source code is governed by a BSD-style
// license that can be found in the LICENSE file.
package goxml_test
import (
"fmt"
"log"
"strings"
"git.pyer.club/kingecg/goxml"
)
type Size int
const (
Unrecognized Size = iota
Small
Large
)
func (s *Size) UnmarshalText(text []byte) error {
switch strings.ToLower(string(text)) {
default:
*s = Unrecognized
case "small":
*s = Small
case "large":
*s = Large
}
return nil
}
func (s Size) MarshalText() ([]byte, error) {
var name string
switch s {
default:
name = "unrecognized"
case Small:
name = "small"
case Large:
name = "large"
}
return []byte(name), nil
}
func Example_textMarshalXML() {
blob := `
<sizes>
<size>small</size>
<size>regular</size>
<size>large</size>
<size>unrecognized</size>
<size>small</size>
<size>normal</size>
<size>small</size>
<size>large</size>
</sizes>`
var inventory struct {
Sizes []Size `xml:"size"`
}
if err := goxml.Unmarshal([]byte(blob), &inventory); err != nil {
log.Fatal(err)
}
counts := make(map[Size]int)
for _, size := range inventory.Sizes {
counts[size] += 1
}
fmt.Printf("Inventory Counts:\n* Small: %d\n* Large: %d\n* Unrecognized: %d\n",
counts[Small], counts[Large], counts[Unrecognized])
// Output:
// Inventory Counts:
// * Small: 3
// * Large: 2
// * Unrecognized: 3
}

3
go.mod Normal file
View File

@ -0,0 +1,3 @@
module git.pyer.club/kingecg/goxml
go 1.23.1

1131
marshal.go Normal file

File diff suppressed because it is too large Load Diff

2591
marshal_test.go Normal file

File diff suppressed because it is too large Load Diff

44
prefix_test.go Normal file
View File

@ -0,0 +1,44 @@
package goxml_test
import (
"testing"
"git.pyer.club/kingecg/goxml"
)
type Person struct {
Name string `xml:"p:name"`
Age int `xml:"p:age"`
}
type Employee struct {
Person
Salary int `xml:"p:salary"`
}
func TestUnmarshal(t *testing.T) {
xml := `
<Employee>
<p:name>John Doe</p:name>
<p:age>42</p:age>
<p:salary>100000</p:salary>
</Employee>
`
e := Employee{}
err := goxml.Unmarshal([]byte(xml), &e)
if err != nil {
t.Fatal(err)
}
}
func TestMashal(t *testing.T) {
e := Employee{
Person{
Name: "John",
Age: 30,
},
50000,
}
b, _ := goxml.Marshal(e)
t.Log(string(b))
}

777
read.go Normal file
View File

@ -0,0 +1,777 @@
// Copyright 2009 The Go Authors. All rights reserved.
// Use of this source code is governed by a BSD-style
// license that can be found in the LICENSE file.
package goxml
import (
"bytes"
"encoding"
"errors"
"fmt"
"reflect"
"runtime"
"strconv"
"strings"
)
// BUG(rsc): Mapping between XML elements and data structures is inherently flawed:
// an XML element is an order-dependent collection of anonymous
// values, while a data structure is an order-independent collection
// of named values.
// See [encoding/json] for a textual representation more suitable
// to data structures.
// Unmarshal parses the XML-encoded data and stores the result in
// the value pointed to by v, which must be an arbitrary struct,
// slice, or string. Well-formed data that does not fit into v is
// discarded.
//
// Because Unmarshal uses the reflect package, it can only assign
// to exported (upper case) fields. Unmarshal uses a case-sensitive
// comparison to match XML element names to tag values and struct
// field names.
//
// Unmarshal maps an XML element to a struct using the following rules.
// In the rules, the tag of a field refers to the value associated with the
// key 'xml' in the struct field's tag (see the example above).
//
// - If the struct has a field of type []byte or string with tag
// ",innerxml", Unmarshal accumulates the raw XML nested inside the
// element in that field. The rest of the rules still apply.
//
// - If the struct has a field named XMLName of type Name,
// Unmarshal records the element name in that field.
//
// - If the XMLName field has an associated tag of the form
// "name" or "namespace-URL name", the XML element must have
// the given name (and, optionally, name space) or else Unmarshal
// returns an error.
//
// - If the XML element has an attribute whose name matches a
// struct field name with an associated tag containing ",attr" or
// the explicit name in a struct field tag of the form "name,attr",
// Unmarshal records the attribute value in that field.
//
// - If the XML element has an attribute not handled by the previous
// rule and the struct has a field with an associated tag containing
// ",any,attr", Unmarshal records the attribute value in the first
// such field.
//
// - If the XML element contains character data, that data is
// accumulated in the first struct field that has tag ",chardata".
// The struct field may have type []byte or string.
// If there is no such field, the character data is discarded.
//
// - If the XML element contains comments, they are accumulated in
// the first struct field that has tag ",comment". The struct
// field may have type []byte or string. If there is no such
// field, the comments are discarded.
//
// - If the XML element contains a sub-element whose name matches
// the prefix of a tag formatted as "a" or "a>b>c", unmarshal
// will descend into the XML structure looking for elements with the
// given names, and will map the innermost elements to that struct
// field. A tag starting with ">" is equivalent to one starting
// with the field name followed by ">".
//
// - If the XML element contains a sub-element whose name matches
// a struct field's XMLName tag and the struct field has no
// explicit name tag as per the previous rule, unmarshal maps
// the sub-element to that struct field.
//
// - If the XML element contains a sub-element whose name matches a
// field without any mode flags (",attr", ",chardata", etc), Unmarshal
// maps the sub-element to that struct field.
//
// - If the XML element contains a sub-element that hasn't matched any
// of the above rules and the struct has a field with tag ",any",
// unmarshal maps the sub-element to that struct field.
//
// - An anonymous struct field is handled as if the fields of its
// value were part of the outer struct.
//
// - A struct field with tag "-" is never unmarshaled into.
//
// If Unmarshal encounters a field type that implements the Unmarshaler
// interface, Unmarshal calls its UnmarshalXML method to produce the value from
// the XML element. Otherwise, if the value implements
// [encoding.TextUnmarshaler], Unmarshal calls that value's UnmarshalText method.
//
// Unmarshal maps an XML element to a string or []byte by saving the
// concatenation of that element's character data in the string or
// []byte. The saved []byte is never nil.
//
// Unmarshal maps an attribute value to a string or []byte by saving
// the value in the string or slice.
//
// Unmarshal maps an attribute value to an [Attr] by saving the attribute,
// including its name, in the Attr.
//
// Unmarshal maps an XML element or attribute value to a slice by
// extending the length of the slice and mapping the element or attribute
// to the newly created value.
//
// Unmarshal maps an XML element or attribute value to a bool by
// setting it to the boolean value represented by the string. Whitespace
// is trimmed and ignored.
//
// Unmarshal maps an XML element or attribute value to an integer or
// floating-point field by setting the field to the result of
// interpreting the string value in decimal. There is no check for
// overflow. Whitespace is trimmed and ignored.
//
// Unmarshal maps an XML element to a Name by recording the element
// name.
//
// Unmarshal maps an XML element to a pointer by setting the pointer
// to a freshly allocated value and then mapping the element to that value.
//
// A missing element or empty attribute value will be unmarshaled as a zero value.
// If the field is a slice, a zero value will be appended to the field. Otherwise, the
// field will be set to its zero value.
func Unmarshal(data []byte, v any) error {
return NewDecoder(bytes.NewReader(data)).Decode(v)
}
// Decode works like [Unmarshal], except it reads the decoder
// stream to find the start element.
func (d *Decoder) Decode(v any) error {
return d.DecodeElement(v, nil)
}
// DecodeElement works like [Unmarshal] except that it takes
// a pointer to the start XML element to decode into v.
// It is useful when a client reads some raw XML tokens itself
// but also wants to defer to [Unmarshal] for some elements.
func (d *Decoder) DecodeElement(v any, start *StartElement) error {
val := reflect.ValueOf(v)
if val.Kind() != reflect.Pointer {
return errors.New("non-pointer passed to Unmarshal")
}
if val.IsNil() {
return errors.New("nil pointer passed to Unmarshal")
}
return d.unmarshal(val.Elem(), start, 0)
}
// An UnmarshalError represents an error in the unmarshaling process.
type UnmarshalError string
func (e UnmarshalError) Error() string { return string(e) }
// Unmarshaler is the interface implemented by objects that can unmarshal
// an XML element description of themselves.
//
// UnmarshalXML decodes a single XML element
// beginning with the given start element.
// If it returns an error, the outer call to Unmarshal stops and
// returns that error.
// UnmarshalXML must consume exactly one XML element.
// One common implementation strategy is to unmarshal into
// a separate value with a layout matching the expected XML
// using d.DecodeElement, and then to copy the data from
// that value into the receiver.
// Another common strategy is to use d.Token to process the
// XML object one token at a time.
// UnmarshalXML may not use d.RawToken.
type Unmarshaler interface {
UnmarshalXML(d *Decoder, start StartElement) error
}
// UnmarshalerAttr is the interface implemented by objects that can unmarshal
// an XML attribute description of themselves.
//
// UnmarshalXMLAttr decodes a single XML attribute.
// If it returns an error, the outer call to [Unmarshal] stops and
// returns that error.
// UnmarshalXMLAttr is used only for struct fields with the
// "attr" option in the field tag.
type UnmarshalerAttr interface {
UnmarshalXMLAttr(attr Attr) error
}
// receiverType returns the receiver type to use in an expression like "%s.MethodName".
func receiverType(val any) string {
t := reflect.TypeOf(val)
if t.Name() != "" {
return t.String()
}
return "(" + t.String() + ")"
}
// unmarshalInterface unmarshals a single XML element into val.
// start is the opening tag of the element.
func (d *Decoder) unmarshalInterface(val Unmarshaler, start *StartElement) error {
// Record that decoder must stop at end tag corresponding to start.
d.pushEOF()
d.unmarshalDepth++
err := val.UnmarshalXML(d, *start)
d.unmarshalDepth--
if err != nil {
d.popEOF()
return err
}
if !d.popEOF() {
return fmt.Errorf("xml: %s.UnmarshalXML did not consume entire <%s> element", receiverType(val), start.Name.Local)
}
return nil
}
// unmarshalTextInterface unmarshals a single XML element into val.
// The chardata contained in the element (but not its children)
// is passed to the text unmarshaler.
func (d *Decoder) unmarshalTextInterface(val encoding.TextUnmarshaler) error {
var buf []byte
depth := 1
for depth > 0 {
t, err := d.Token()
if err != nil {
return err
}
switch t := t.(type) {
case CharData:
if depth == 1 {
buf = append(buf, t...)
}
case StartElement:
depth++
case EndElement:
depth--
}
}
return val.UnmarshalText(buf)
}
// unmarshalAttr unmarshals a single XML attribute into val.
func (d *Decoder) unmarshalAttr(val reflect.Value, attr Attr) error {
if val.Kind() == reflect.Pointer {
if val.IsNil() {
val.Set(reflect.New(val.Type().Elem()))
}
val = val.Elem()
}
if val.CanInterface() && val.Type().Implements(unmarshalerAttrType) {
// This is an unmarshaler with a non-pointer receiver,
// so it's likely to be incorrect, but we do what we're told.
return val.Interface().(UnmarshalerAttr).UnmarshalXMLAttr(attr)
}
if val.CanAddr() {
pv := val.Addr()
if pv.CanInterface() && pv.Type().Implements(unmarshalerAttrType) {
return pv.Interface().(UnmarshalerAttr).UnmarshalXMLAttr(attr)
}
}
// Not an UnmarshalerAttr; try encoding.TextUnmarshaler.
if val.CanInterface() && val.Type().Implements(textUnmarshalerType) {
// This is an unmarshaler with a non-pointer receiver,
// so it's likely to be incorrect, but we do what we're told.
return val.Interface().(encoding.TextUnmarshaler).UnmarshalText([]byte(attr.Value))
}
if val.CanAddr() {
pv := val.Addr()
if pv.CanInterface() && pv.Type().Implements(textUnmarshalerType) {
return pv.Interface().(encoding.TextUnmarshaler).UnmarshalText([]byte(attr.Value))
}
}
if val.Type().Kind() == reflect.Slice && val.Type().Elem().Kind() != reflect.Uint8 {
// Slice of element values.
// Grow slice.
n := val.Len()
val.Grow(1)
val.SetLen(n + 1)
// Recur to read element into slice.
if err := d.unmarshalAttr(val.Index(n), attr); err != nil {
val.SetLen(n)
return err
}
return nil
}
if val.Type() == attrType {
val.Set(reflect.ValueOf(attr))
return nil
}
return copyValue(val, []byte(attr.Value))
}
var (
attrType = reflect.TypeFor[Attr]()
unmarshalerType = reflect.TypeFor[Unmarshaler]()
unmarshalerAttrType = reflect.TypeFor[UnmarshalerAttr]()
textUnmarshalerType = reflect.TypeFor[encoding.TextUnmarshaler]()
)
const (
maxUnmarshalDepth = 10000
maxUnmarshalDepthWasm = 5000 // go.dev/issue/56498
)
var errUnmarshalDepth = errors.New("exceeded max depth")
// Unmarshal a single XML element into val.
func (d *Decoder) unmarshal(val reflect.Value, start *StartElement, depth int) error {
if depth >= maxUnmarshalDepth || runtime.GOARCH == "wasm" && depth >= maxUnmarshalDepthWasm {
return errUnmarshalDepth
}
// Find start element if we need it.
if start == nil {
for {
tok, err := d.Token()
if err != nil {
return err
}
if t, ok := tok.(StartElement); ok {
start = &t
break
}
}
}
// Load value from interface, but only if the result will be
// usefully addressable.
if val.Kind() == reflect.Interface && !val.IsNil() {
e := val.Elem()
if e.Kind() == reflect.Pointer && !e.IsNil() {
val = e
}
}
if val.Kind() == reflect.Pointer {
if val.IsNil() {
val.Set(reflect.New(val.Type().Elem()))
}
val = val.Elem()
}
if val.CanInterface() && val.Type().Implements(unmarshalerType) {
// This is an unmarshaler with a non-pointer receiver,
// so it's likely to be incorrect, but we do what we're told.
return d.unmarshalInterface(val.Interface().(Unmarshaler), start)
}
if val.CanAddr() {
pv := val.Addr()
if pv.CanInterface() && pv.Type().Implements(unmarshalerType) {
return d.unmarshalInterface(pv.Interface().(Unmarshaler), start)
}
}
if val.CanInterface() && val.Type().Implements(textUnmarshalerType) {
return d.unmarshalTextInterface(val.Interface().(encoding.TextUnmarshaler))
}
if val.CanAddr() {
pv := val.Addr()
if pv.CanInterface() && pv.Type().Implements(textUnmarshalerType) {
return d.unmarshalTextInterface(pv.Interface().(encoding.TextUnmarshaler))
}
}
var (
data []byte
saveData reflect.Value
comment []byte
saveComment reflect.Value
saveXML reflect.Value
saveXMLIndex int
saveXMLData []byte
saveAny reflect.Value
sv reflect.Value
tinfo *typeInfo
err error
)
switch v := val; v.Kind() {
default:
return errors.New("unknown type " + v.Type().String())
case reflect.Interface:
// TODO: For now, simply ignore the field. In the near
// future we may choose to unmarshal the start
// element on it, if not nil.
return d.Skip()
case reflect.Slice:
typ := v.Type()
if typ.Elem().Kind() == reflect.Uint8 {
// []byte
saveData = v
break
}
// Slice of element values.
// Grow slice.
n := v.Len()
v.Grow(1)
v.SetLen(n + 1)
// Recur to read element into slice.
if err := d.unmarshal(v.Index(n), start, depth+1); err != nil {
v.SetLen(n)
return err
}
return nil
case reflect.Bool, reflect.Float32, reflect.Float64, reflect.Int, reflect.Int8, reflect.Int16, reflect.Int32, reflect.Int64, reflect.Uint, reflect.Uint8, reflect.Uint16, reflect.Uint32, reflect.Uint64, reflect.Uintptr, reflect.String:
saveData = v
case reflect.Struct:
typ := v.Type()
if typ == nameType {
v.Set(reflect.ValueOf(start.Name))
break
}
sv = v
tinfo, err = getTypeInfo(typ, true)
if err != nil {
return err
}
// Validate and assign element name.
if tinfo.xmlname != nil {
finfo := tinfo.xmlname
if finfo.name != "" && finfo.name != start.Name.Local {
return UnmarshalError("expected element type <" + finfo.name + "> but have <" + start.Name.Local + ">")
}
if finfo.xmlns != "" && finfo.xmlns != start.Name.Space {
e := "expected element <" + finfo.name + "> in name space " + finfo.xmlns + " but have "
if start.Name.Space == "" {
e += "no name space"
} else {
e += start.Name.Space
}
return UnmarshalError(e)
}
fv := finfo.value(sv, initNilPointers)
if _, ok := fv.Interface().(Name); ok {
fv.Set(reflect.ValueOf(start.Name))
}
}
// Assign attributes.
for _, a := range start.Attr {
handled := false
any := -1
for i := range tinfo.fields {
finfo := &tinfo.fields[i]
switch finfo.flags & fMode {
case fAttr:
strv := finfo.value(sv, initNilPointers)
if a.Name.Local == finfo.name && (finfo.xmlns == "" || finfo.xmlns == a.Name.Space) {
if err := d.unmarshalAttr(strv, a); err != nil {
return err
}
handled = true
}
case fAny | fAttr:
if any == -1 {
any = i
}
}
}
if !handled && any >= 0 {
finfo := &tinfo.fields[any]
strv := finfo.value(sv, initNilPointers)
if err := d.unmarshalAttr(strv, a); err != nil {
return err
}
}
}
// Determine whether we need to save character data or comments.
for i := range tinfo.fields {
finfo := &tinfo.fields[i]
switch finfo.flags & fMode {
case fCDATA, fCharData:
if !saveData.IsValid() {
saveData = finfo.value(sv, initNilPointers)
}
case fComment:
if !saveComment.IsValid() {
saveComment = finfo.value(sv, initNilPointers)
}
case fAny, fAny | fElement:
if !saveAny.IsValid() {
saveAny = finfo.value(sv, initNilPointers)
}
case fInnerXML:
if !saveXML.IsValid() {
saveXML = finfo.value(sv, initNilPointers)
if d.saved == nil {
saveXMLIndex = 0
d.saved = new(bytes.Buffer)
} else {
saveXMLIndex = d.savedOffset()
}
}
}
}
}
// Find end element.
// Process sub-elements along the way.
Loop:
for {
var savedOffset int
if saveXML.IsValid() {
savedOffset = d.savedOffset()
}
tok, err := d.Token()
if err != nil {
return err
}
switch t := tok.(type) {
case StartElement:
consumed := false
if sv.IsValid() {
// unmarshalPath can call unmarshal, so we need to pass the depth through so that
// we can continue to enforce the maximum recursion limit.
consumed, err = d.unmarshalPath(tinfo, sv, nil, &t, depth)
if err != nil {
return err
}
if !consumed && saveAny.IsValid() {
consumed = true
if err := d.unmarshal(saveAny, &t, depth+1); err != nil {
return err
}
}
}
if !consumed {
if err := d.Skip(); err != nil {
return err
}
}
case EndElement:
if saveXML.IsValid() {
saveXMLData = d.saved.Bytes()[saveXMLIndex:savedOffset]
if saveXMLIndex == 0 {
d.saved = nil
}
}
break Loop
case CharData:
if saveData.IsValid() {
data = append(data, t...)
}
case Comment:
if saveComment.IsValid() {
comment = append(comment, t...)
}
}
}
if saveData.IsValid() && saveData.CanInterface() && saveData.Type().Implements(textUnmarshalerType) {
if err := saveData.Interface().(encoding.TextUnmarshaler).UnmarshalText(data); err != nil {
return err
}
saveData = reflect.Value{}
}
if saveData.IsValid() && saveData.CanAddr() {
pv := saveData.Addr()
if pv.CanInterface() && pv.Type().Implements(textUnmarshalerType) {
if err := pv.Interface().(encoding.TextUnmarshaler).UnmarshalText(data); err != nil {
return err
}
saveData = reflect.Value{}
}
}
if err := copyValue(saveData, data); err != nil {
return err
}
switch t := saveComment; t.Kind() {
case reflect.String:
t.SetString(string(comment))
case reflect.Slice:
t.Set(reflect.ValueOf(comment))
}
switch t := saveXML; t.Kind() {
case reflect.String:
t.SetString(string(saveXMLData))
case reflect.Slice:
if t.Type().Elem().Kind() == reflect.Uint8 {
t.Set(reflect.ValueOf(saveXMLData))
}
}
return nil
}
func copyValue(dst reflect.Value, src []byte) (err error) {
dst0 := dst
if dst.Kind() == reflect.Pointer {
if dst.IsNil() {
dst.Set(reflect.New(dst.Type().Elem()))
}
dst = dst.Elem()
}
// Save accumulated data.
switch dst.Kind() {
case reflect.Invalid:
// Probably a comment.
default:
return errors.New("cannot unmarshal into " + dst0.Type().String())
case reflect.Int, reflect.Int8, reflect.Int16, reflect.Int32, reflect.Int64:
if len(src) == 0 {
dst.SetInt(0)
return nil
}
itmp, err := strconv.ParseInt(strings.TrimSpace(string(src)), 10, dst.Type().Bits())
if err != nil {
return err
}
dst.SetInt(itmp)
case reflect.Uint, reflect.Uint8, reflect.Uint16, reflect.Uint32, reflect.Uint64, reflect.Uintptr:
if len(src) == 0 {
dst.SetUint(0)
return nil
}
utmp, err := strconv.ParseUint(strings.TrimSpace(string(src)), 10, dst.Type().Bits())
if err != nil {
return err
}
dst.SetUint(utmp)
case reflect.Float32, reflect.Float64:
if len(src) == 0 {
dst.SetFloat(0)
return nil
}
ftmp, err := strconv.ParseFloat(strings.TrimSpace(string(src)), dst.Type().Bits())
if err != nil {
return err
}
dst.SetFloat(ftmp)
case reflect.Bool:
if len(src) == 0 {
dst.SetBool(false)
return nil
}
value, err := strconv.ParseBool(strings.TrimSpace(string(src)))
if err != nil {
return err
}
dst.SetBool(value)
case reflect.String:
dst.SetString(string(src))
case reflect.Slice:
if len(src) == 0 {
// non-nil to flag presence
src = []byte{}
}
dst.SetBytes(src)
}
return nil
}
// unmarshalPath walks down an XML structure looking for wanted
// paths, and calls unmarshal on them.
// The consumed result tells whether XML elements have been consumed
// from the Decoder until start's matching end element, or if it's
// still untouched because start is uninteresting for sv's fields.
func (d *Decoder) unmarshalPath(tinfo *typeInfo, sv reflect.Value, parents []string, start *StartElement, depth int) (consumed bool, err error) {
recurse := false
Loop:
for i := range tinfo.fields {
finfo := &tinfo.fields[i]
if finfo.flags&fElement == 0 || len(finfo.parents) < len(parents) || finfo.xmlns != "" && finfo.xmlns != start.Name.Space {
continue
}
for j := range parents {
if parents[j] != finfo.parents[j] {
continue Loop
}
}
if len(finfo.parents) == len(parents) && finfo.name == start.Name.Local {
// It's a perfect match, unmarshal the field.
return true, d.unmarshal(finfo.value(sv, initNilPointers), start, depth+1)
}
if len(finfo.parents) > len(parents) && finfo.parents[len(parents)] == start.Name.Local {
// It's a prefix for the field. Break and recurse
// since it's not ok for one field path to be itself
// the prefix for another field path.
recurse = true
// We can reuse the same slice as long as we
// don't try to append to it.
parents = finfo.parents[:len(parents)+1]
break
}
}
if !recurse {
// We have no business with this element.
return false, nil
}
// The element is not a perfect match for any field, but one
// or more fields have the path to this element as a parent
// prefix. Recurse and attempt to match these.
for {
var tok Token
tok, err = d.Token()
if err != nil {
return true, err
}
switch t := tok.(type) {
case StartElement:
// the recursion depth of unmarshalPath is limited to the path length specified
// by the struct field tag, so we don't increment the depth here.
consumed2, err := d.unmarshalPath(tinfo, sv, parents, &t, depth)
if err != nil {
return true, err
}
if !consumed2 {
if err := d.Skip(); err != nil {
return true, err
}
}
case EndElement:
return true, nil
}
}
}
// Skip reads tokens until it has consumed the end element
// matching the most recent start element already consumed,
// skipping nested structures.
// It returns nil if it finds an end element matching the start
// element; otherwise it returns an error describing the problem.
func (d *Decoder) Skip() error {
var depth int64
for {
tok, err := d.Token()
if err != nil {
return err
}
switch tok.(type) {
case StartElement:
depth++
case EndElement:
if depth == 0 {
return nil
}
depth--
}
}
}

1128
read_test.go Normal file

File diff suppressed because it is too large Load Diff

372
typeinfo.go Normal file
View File

@ -0,0 +1,372 @@
// Copyright 2011 The Go Authors. All rights reserved.
// Use of this source code is governed by a BSD-style
// license that can be found in the LICENSE file.
package goxml
import (
"fmt"
"reflect"
"strings"
"sync"
)
// typeInfo holds details for the xml representation of a type.
type typeInfo struct {
xmlname *fieldInfo
fields []fieldInfo
}
// fieldInfo holds details for the xml representation of a single field.
type fieldInfo struct {
idx []int
name string
xmlns string
flags fieldFlags
parents []string
}
type fieldFlags int
const (
fElement fieldFlags = 1 << iota
fAttr
fCDATA
fCharData
fInnerXML
fComment
fAny
fOmitEmpty
fMode = fElement | fAttr | fCDATA | fCharData | fInnerXML | fComment | fAny
xmlName = "XMLName"
)
var tinfoMap sync.Map // map[reflect.Type]*typeInfo
var nameType = reflect.TypeFor[Name]()
// getTypeInfo returns the typeInfo structure with details necessary
// for marshaling and unmarshaling typ.
func getTypeInfo(typ reflect.Type, inDecode bool) (*typeInfo, error) {
if ti, ok := tinfoMap.Load(typ); ok {
return ti.(*typeInfo), nil
}
tinfo := &typeInfo{}
if typ.Kind() == reflect.Struct && typ != nameType {
n := typ.NumField()
for i := 0; i < n; i++ {
f := typ.Field(i)
if (!f.IsExported() && !f.Anonymous) || f.Tag.Get("xml") == "-" {
continue // Private field
}
// For embedded structs, embed its fields.
if f.Anonymous {
t := f.Type
if t.Kind() == reflect.Pointer {
t = t.Elem()
}
if t.Kind() == reflect.Struct {
inner, err := getTypeInfo(t, inDecode)
if err != nil {
return nil, err
}
if tinfo.xmlname == nil {
tinfo.xmlname = inner.xmlname
}
for _, finfo := range inner.fields {
finfo.idx = append([]int{i}, finfo.idx...)
if err := addFieldInfo(typ, tinfo, &finfo); err != nil {
return nil, err
}
}
continue
}
}
finfo, err := structFieldInfo(typ, &f, inDecode)
if err != nil {
return nil, err
}
if f.Name == xmlName {
tinfo.xmlname = finfo
continue
}
// Add the field if it doesn't conflict with other fields.
if err := addFieldInfo(typ, tinfo, finfo); err != nil {
return nil, err
}
}
}
ti, _ := tinfoMap.LoadOrStore(typ, tinfo)
return ti.(*typeInfo), nil
}
// structFieldInfo builds and returns a fieldInfo for f.
func structFieldInfo(typ reflect.Type, f *reflect.StructField, inDecode bool) (*fieldInfo, error) {
finfo := &fieldInfo{idx: f.Index}
// Split the tag from the xml namespace if necessary.
tag := f.Tag.Get("xml")
if ns, t, ok := strings.Cut(tag, " "); ok {
finfo.xmlns, tag = ns, t
}
if inDecode {
if ns, t, ok := strings.Cut(tag, ":"); ok {
finfo.xmlns, tag = ns, t
}
}
// Parse flags.
tokens := strings.Split(tag, ",")
if len(tokens) == 1 {
finfo.flags = fElement
} else {
tag = tokens[0]
for _, flag := range tokens[1:] {
switch flag {
case "attr":
finfo.flags |= fAttr
case "cdata":
finfo.flags |= fCDATA
case "chardata":
finfo.flags |= fCharData
case "innerxml":
finfo.flags |= fInnerXML
case "comment":
finfo.flags |= fComment
case "any":
finfo.flags |= fAny
case "omitempty":
finfo.flags |= fOmitEmpty
}
}
// Validate the flags used.
valid := true
switch mode := finfo.flags & fMode; mode {
case 0:
finfo.flags |= fElement
case fAttr, fCDATA, fCharData, fInnerXML, fComment, fAny, fAny | fAttr:
if f.Name == xmlName || tag != "" && mode != fAttr {
valid = false
}
default:
// This will also catch multiple modes in a single field.
valid = false
}
if finfo.flags&fMode == fAny {
finfo.flags |= fElement
}
if finfo.flags&fOmitEmpty != 0 && finfo.flags&(fElement|fAttr) == 0 {
valid = false
}
if !valid {
return nil, fmt.Errorf("xml: invalid tag in field %s of type %s: %q",
f.Name, typ, f.Tag.Get("xml"))
}
}
// Use of xmlns without a name is not allowed.
if finfo.xmlns != "" && tag == "" {
return nil, fmt.Errorf("xml: namespace without name in field %s of type %s: %q",
f.Name, typ, f.Tag.Get("xml"))
}
if f.Name == xmlName {
// The XMLName field records the XML element name. Don't
// process it as usual because its name should default to
// empty rather than to the field name.
finfo.name = tag
return finfo, nil
}
if tag == "" {
// If the name part of the tag is completely empty, get
// default from XMLName of underlying struct if feasible,
// or field name otherwise.
if xmlname := lookupXMLName(f.Type, inDecode); xmlname != nil {
finfo.xmlns, finfo.name = xmlname.xmlns, xmlname.name
} else {
finfo.name = f.Name
}
return finfo, nil
}
// Prepare field name and parents.
parents := strings.Split(tag, ">")
if parents[0] == "" {
parents[0] = f.Name
}
if parents[len(parents)-1] == "" {
return nil, fmt.Errorf("xml: trailing '>' in field %s of type %s", f.Name, typ)
}
finfo.name = parents[len(parents)-1]
if len(parents) > 1 {
if (finfo.flags & fElement) == 0 {
return nil, fmt.Errorf("xml: %s chain not valid with %s flag", tag, strings.Join(tokens[1:], ","))
}
finfo.parents = parents[:len(parents)-1]
}
// If the field type has an XMLName field, the names must match
// so that the behavior of both marshaling and unmarshaling
// is straightforward and unambiguous.
if finfo.flags&fElement != 0 {
ftyp := f.Type
xmlname := lookupXMLName(ftyp, inDecode)
if xmlname != nil && xmlname.name != finfo.name {
return nil, fmt.Errorf("xml: name %q in tag of %s.%s conflicts with name %q in %s.XMLName",
finfo.name, typ, f.Name, xmlname.name, ftyp)
}
}
return finfo, nil
}
// lookupXMLName returns the fieldInfo for typ's XMLName field
// in case it exists and has a valid xml field tag, otherwise
// it returns nil.
func lookupXMLName(typ reflect.Type, inDecode bool) (xmlname *fieldInfo) {
for typ.Kind() == reflect.Pointer {
typ = typ.Elem()
}
if typ.Kind() != reflect.Struct {
return nil
}
for i, n := 0, typ.NumField(); i < n; i++ {
f := typ.Field(i)
if f.Name != xmlName {
continue
}
finfo, err := structFieldInfo(typ, &f, inDecode)
if err == nil && finfo.name != "" {
return finfo
}
// Also consider errors as a non-existent field tag
// and let getTypeInfo itself report the error.
break
}
return nil
}
// addFieldInfo adds finfo to tinfo.fields if there are no
// conflicts, or if conflicts arise from previous fields that were
// obtained from deeper embedded structures than finfo. In the latter
// case, the conflicting entries are dropped.
// A conflict occurs when the path (parent + name) to a field is
// itself a prefix of another path, or when two paths match exactly.
// It is okay for field paths to share a common, shorter prefix.
func addFieldInfo(typ reflect.Type, tinfo *typeInfo, newf *fieldInfo) error {
var conflicts []int
Loop:
// First, figure all conflicts. Most working code will have none.
for i := range tinfo.fields {
oldf := &tinfo.fields[i]
if oldf.flags&fMode != newf.flags&fMode {
continue
}
if oldf.xmlns != "" && newf.xmlns != "" && oldf.xmlns != newf.xmlns {
continue
}
minl := min(len(newf.parents), len(oldf.parents))
for p := 0; p < minl; p++ {
if oldf.parents[p] != newf.parents[p] {
continue Loop
}
}
if len(oldf.parents) > len(newf.parents) {
if oldf.parents[len(newf.parents)] == newf.name {
conflicts = append(conflicts, i)
}
} else if len(oldf.parents) < len(newf.parents) {
if newf.parents[len(oldf.parents)] == oldf.name {
conflicts = append(conflicts, i)
}
} else {
if newf.name == oldf.name && newf.xmlns == oldf.xmlns {
conflicts = append(conflicts, i)
}
}
}
// Without conflicts, add the new field and return.
if conflicts == nil {
tinfo.fields = append(tinfo.fields, *newf)
return nil
}
// If any conflict is shallower, ignore the new field.
// This matches the Go field resolution on embedding.
for _, i := range conflicts {
if len(tinfo.fields[i].idx) < len(newf.idx) {
return nil
}
}
// Otherwise, if any of them is at the same depth level, it's an error.
for _, i := range conflicts {
oldf := &tinfo.fields[i]
if len(oldf.idx) == len(newf.idx) {
f1 := typ.FieldByIndex(oldf.idx)
f2 := typ.FieldByIndex(newf.idx)
return &TagPathError{typ, f1.Name, f1.Tag.Get("xml"), f2.Name, f2.Tag.Get("xml")}
}
}
// Otherwise, the new field is shallower, and thus takes precedence,
// so drop the conflicting fields from tinfo and append the new one.
for c := len(conflicts) - 1; c >= 0; c-- {
i := conflicts[c]
copy(tinfo.fields[i:], tinfo.fields[i+1:])
tinfo.fields = tinfo.fields[:len(tinfo.fields)-1]
}
tinfo.fields = append(tinfo.fields, *newf)
return nil
}
// A TagPathError represents an error in the unmarshaling process
// caused by the use of field tags with conflicting paths.
type TagPathError struct {
Struct reflect.Type
Field1, Tag1 string
Field2, Tag2 string
}
func (e *TagPathError) Error() string {
return fmt.Sprintf("%s field %q with tag %q conflicts with field %q with tag %q", e.Struct, e.Field1, e.Tag1, e.Field2, e.Tag2)
}
const (
initNilPointers = true
dontInitNilPointers = false
)
// value returns v's field value corresponding to finfo.
// It's equivalent to v.FieldByIndex(finfo.idx), but when passed
// initNilPointers, it initializes and dereferences pointers as necessary.
// When passed dontInitNilPointers and a nil pointer is reached, the function
// returns a zero reflect.Value.
func (finfo *fieldInfo) value(v reflect.Value, shouldInitNilPointers bool) reflect.Value {
for i, x := range finfo.idx {
if i > 0 {
t := v.Type()
if t.Kind() == reflect.Pointer && t.Elem().Kind() == reflect.Struct {
if v.IsNil() {
if !shouldInitNilPointers {
return reflect.Value{}
}
v.Set(reflect.New(v.Type().Elem()))
}
v = v.Elem()
}
}
v = v.Field(x)
}
return v
}

2075
xml.go Normal file

File diff suppressed because it is too large Load Diff

1488
xml_test.go Normal file

File diff suppressed because it is too large Load Diff