fix #111

antchfx · Dec 10, 2023 · cde8fe6 · cde8fe6
1 parent f30da80
commit cde8fe6
Show file tree

Hide file tree

Showing 2 changed files with 50 additions and 7 deletions.
diff --git a/parse.go b/parse.go
@@ -8,6 +8,7 @@ import (
 	"net/http"
 	"regexp"
 	"strings"
+	"sync"
 
 	"github.com/antchfx/xpath"
 	"golang.org/x/net/html/charset"
@@ -59,6 +60,8 @@ type parser struct {
 	streamNode          *Node         // Need to remember the last target node So we can clean it up upon next Read() call.
 	streamNodePrev      *Node         // Need to remember target node's prev so upon target node removal, we can restore correct prev.
 	reader              *cachedReader // Need to maintain a reference to the reader, so we can determine whether a node contains CDATA.
+	once                sync.Once
+	space2prefix        map[string]string
 }
 
 func createParser(r io.Reader) *parser {
@@ -77,9 +80,11 @@ func createParser(r io.Reader) *parser {
 }
 
 func (p *parser) parse() (*Node, error) {
-	var streamElementNodeCounter int
-	space2prefix := map[string]string{"http://www.w3.org/XML/1998/namespace": "xml"}
+	p.once.Do(func() {
+		p.space2prefix = map[string]string{"http://www.w3.org/XML/1998/namespace": "xml"}
+	})
 
+	var streamElementNodeCounter int
 	for {
 		p.reader.StartCaching()
 		tok, err := p.decoder.Token()
@@ -108,24 +113,24 @@ func (p *parser) parse() (*Node, error) {
 
 			for _, att := range tok.Attr {
 				if att.Name.Local == "xmlns" {
-					space2prefix[att.Value] = "" // reset empty if exist the default namespace
+					p.space2prefix[att.Value] = "" // reset empty if exist the default namespace
 					//	defaultNamespaceURL = att.Value
 				} else if att.Name.Space == "xmlns" {
 					// maybe there are have duplicate NamespaceURL?
-					space2prefix[att.Value] = att.Name.Local
+					p.space2prefix[att.Value] = att.Name.Local
 				}
 			}
 
 			if space := tok.Name.Space; space != "" {
-				if _, found := space2prefix[space]; !found && p.decoder.Strict {
+				if _, found := p.space2prefix[space]; !found && p.decoder.Strict {
 					return nil, fmt.Errorf("xmlquery: invalid XML document, namespace %s is missing", space)
 				}
 			}
 
 			attributes := make([]Attr, len(tok.Attr))
 			for i, att := range tok.Attr {
 				name := att.Name
-				if prefix, ok := space2prefix[name.Space]; ok {
+				if prefix, ok := p.space2prefix[name.Space]; ok {
 					name.Space = prefix
 				}
 				attributes[i] = Attr{
@@ -155,7 +160,7 @@ func (p *parser) parse() (*Node, error) {
 			}
 
 			if node.NamespaceURI != "" {
-				if v, ok := space2prefix[node.NamespaceURI]; ok {
+				if v, ok := p.space2prefix[node.NamespaceURI]; ok {
 					cached := string(p.reader.Cache())
 					if strings.HasPrefix(cached, fmt.Sprintf("%s:%s", v, node.Data)) || strings.HasPrefix(cached, fmt.Sprintf("<%s:%s", v, node.Data)) {
 						node.Prefix = v

diff --git a/parse_test.go b/parse_test.go
@@ -562,3 +562,41 @@ func TestXMLPreservation(t *testing.T) {
 	testOutputXML(t, "first call result",
 		`<?xml version="1.0" encoding="UTF-8"?><AAA><CCC><![CDATA[c1]]></CCC></AAA>`, doc)
 }
+
+func TestStreamParser_DefaultNamespace(t *testing.T) {
+	s := `
+	<Objects xmlns="http://example.com/schema/2007/someschema">
+		<Object id="ObjectA">ObjectA</Object>
+		<Object id="ObjectB">ObjectB</Object>
+		<Object id="ObjectC">ObjectD</Object>
+	</Objects>`
+
+	sp, err := CreateStreamParser(strings.NewReader(s), "//Objects/*[namespace-uri()=\"http://example.com/schema/2007/someschema\" and local-name()=\"Object\"]")
+	if err != nil {
+		t.Fatal(err.Error())
+	}
+
+	n, err := sp.Read()
+	if err != nil {
+		t.Fatal(err.Error())
+	}
+
+	var x = `<Object id="ObjectA">ObjectA</Object>`
+	testOutputXML(t, "first call result", x, n)
+
+	n, err = sp.Read()
+	if err != nil {
+		t.Fatal(err.Error())
+	}
+
+	x = `<Object id="ObjectB">ObjectB</Object>`
+	testOutputXML(t, "second call result", x, n)
+
+	n, err = sp.Read()
+	if err != nil {
+		t.Fatal(err.Error())
+	}
+
+	x = `<Object id="ObjectC">ObjectD</Object>`
+	testOutputXML(t, "third call result", x, n)
+}