Upstream PuerkitoBio/goquery state of the touched files before 0f91a3810f

This commit is contained in:
upstream committed 2026-09-04 07:52:26 +02:00
commit 445cd7eabe
2 files changed
+518

No files matched your search

+266
View File
@@ -0,0 +1,266 @@
package goquery
import (
"strings"
"golang.org/x/net/html"
)
var classTrimReplacer = strings.NewReplacer("\t", " ", "\r", " ", "\n", " ")
// Attr gets the specified attribute's value for the first element in the
// Selection. To get the value for each element individually, use a looping
// construct such as Each or Map method.
func (s *Selection) Attr(attrName string) (string, bool) {
if len(s.Nodes) != 0 {
if attr := getAttributePtr(attrName, s.Nodes[0]); attr != nil {
return attr.Val, true
}
}
return "", false
}
// AttrOr works like Attr but returns default value if attribute is not present.
func (s *Selection) AttrOr(attrName, defaultValue string) string {
if val, exists := s.Attr(attrName); exists {
return val
}
return defaultValue
}
// RemoveAttr removes the named attribute from each element in the set of matched elements.
func (s *Selection) RemoveAttr(attrName string) *Selection {
for _, n := range s.Nodes {
removeAttr(n, attrName)
}
return s
}
// SetAttr sets the given attribute on each element in the set of matched elements.
func (s *Selection) SetAttr(attrName, val string) *Selection {
for _, n := range s.Nodes {
attr := getAttributePtr(attrName, n)
if attr == nil {
n.Attr = append(n.Attr, html.Attribute{Key: attrName, Val: val})
} else {
attr.Val = val
}
}
return s
}
// Text gets the combined text contents of each element in the set of matched
// elements, including their descendants.
func (s *Selection) Text() string {
var builder strings.Builder
for _, n := range s.Nodes {
s.textHelper(n, &builder)
}
return builder.String()
}
func (s *Selection) textHelper(n *html.Node, builder *strings.Builder) {
if n.Type == html.TextNode {
// Keep newlines and spaces, like jQuery
builder.WriteString(n.Data)
}
for c := n.FirstChild; c != nil; c = c.NextSibling {
s.textHelper(c, builder)
}
}
// Size is an alias for Length.
func (s *Selection) Size() int {
return s.Length()
}
// Length returns the number of elements in the Selection object.
func (s *Selection) Length() int {
return len(s.Nodes)
}
// Html gets the HTML contents of the first element in the set of matched
// elements. It includes text and comment nodes.
func (s *Selection) Html() (ret string, e error) {
// Since there is no .innerHtml, the HTML content must be re-created from
// the nodes using html.Render.
var builder strings.Builder
if len(s.Nodes) > 0 {
for c := s.Nodes[0].FirstChild; c != nil; c = c.NextSibling {
e = html.Render(&builder, c)
if e != nil {
return
}
}
ret = builder.String()
}
return
}
// AddClass adds the given class(es) to each element in the set of matched elements.
// Multiple class names can be specified, separated by a space or via multiple arguments.
func (s *Selection) AddClass(class ...string) *Selection {
classStr := strings.TrimSpace(strings.Join(class, " "))
if classStr == "" {
return s
}
tcls := getClassesSlice(classStr)
for _, n := range s.Nodes {
curClasses, attr := getClassesAndAttr(n)
for _, newClass := range tcls {
if !strings.Contains(curClasses, " "+newClass+" ") {
curClasses += newClass + " "
}
}
setClasses(n, attr, curClasses)
}
return s
}
// HasClass determines whether any of the matched elements are assigned the
// given class.
func (s *Selection) HasClass(class string) bool {
rawClass := class
class = " " + class + " "
for _, n := range s.Nodes {
if n.Type != html.ElementNode {
continue
}
if attr := getAttributePtr("class", n); attr != nil {
if !strings.Contains(attr.Val, rawClass) {
continue
}
val := classTrimReplacer.Replace(attr.Val)
if strings.Contains(" "+val+" ", class) {
return true
}
}
}
return false
}
// RemoveClass removes the given class(es) from each element in the set of matched elements.
// Multiple class names can be specified, separated by a space or via multiple arguments.
// If no class name is provided, all classes are removed.
func (s *Selection) RemoveClass(class ...string) *Selection {
var rclasses []string
classStr := strings.TrimSpace(strings.Join(class, " "))
remove := classStr == ""
if !remove {
rclasses = getClassesSlice(classStr)
}
for _, n := range s.Nodes {
if remove {
removeAttr(n, "class")
} else {
classes, attr := getClassesAndAttr(n)
for _, rcl := range rclasses {
classes = strings.ReplaceAll(classes, " "+rcl+" ", " ")
}
setClasses(n, attr, classes)
}
}
return s
}
// ToggleClass adds or removes the given class(es) for each element in the set of matched elements.
// Multiple class names can be specified, separated by a space or via multiple arguments.
func (s *Selection) ToggleClass(class ...string) *Selection {
classStr := strings.TrimSpace(strings.Join(class, " "))
if classStr == "" {
return s
}
tcls := getClassesSlice(classStr)
for _, n := range s.Nodes {
classes, attr := getClassesAndAttr(n)
for _, tcl := range tcls {
spaceAroundTcl := " " + tcl + " "
if strings.Contains(classes, spaceAroundTcl) {
classes = strings.ReplaceAll(classes, spaceAroundTcl, " ")
} else {
classes += tcl + " "
}
}
setClasses(n, attr, classes)
}
return s
}
func getAttributePtr(attrName string, n *html.Node) *html.Attribute {
if n == nil {
return nil
}
for i, a := range n.Attr {
if a.Key == attrName {
return &n.Attr[i]
}
}
return nil
}
// Get and normalize the "class" attribute from the node.
func getClassesAndAttr(n *html.Node) (classes string, attr *html.Attribute) {
// Applies only to element nodes
if n.Type == html.ElementNode {
attr = getAttributePtr("class", n)
if attr == nil {
n.Attr = append(n.Attr, html.Attribute{
Key: "class",
Val: "",
})
attr = &n.Attr[len(n.Attr)-1]
}
}
if attr == nil {
classes = " "
} else {
classes = classTrimReplacer.Replace(" " + attr.Val + " ")
}
return
}
func getClassesSlice(classes string) []string {
return strings.Fields(classes)
}
func removeAttr(n *html.Node, attrName string) {
for i, a := range n.Attr {
if a.Key == attrName {
n.Attr[i], n.Attr[len(n.Attr)-1], n.Attr =
n.Attr[len(n.Attr)-1], html.Attribute{}, n.Attr[:len(n.Attr)-1]
return
}
}
}
func setClasses(n *html.Node, attr *html.Attribute, classes string) {
classes = strings.TrimSpace(classes)
if classes == "" {
removeAttr(n, "class")
return
}
attr.Val = classes
}
+252
View File
@@ -0,0 +1,252 @@
package goquery
import (
"regexp"
"strings"
"testing"
)
func TestAttrExists(t *testing.T) {
if val, ok := Doc().Find("a").Attr("href"); !ok {
t.Error("Expected a value for the href attribute.")
} else {
t.Logf("Href of first anchor: %v.", val)
}
}
func TestAttrOr(t *testing.T) {
if val := Doc().Find("a").AttrOr("fake-attribute", "alternative"); val != "alternative" {
t.Error("Expected an alternative value for 'fake-attribute' attribute.")
} else {
t.Logf("Value returned for not existing attribute: %v.", val)
}
if val := Doc().Find("zz").AttrOr("fake-attribute", "alternative"); val != "alternative" {
t.Error("Expected an alternative value for 'fake-attribute' on an empty selection.")
} else {
t.Logf("Value returned for empty selection: %v.", val)
}
}
func TestAttrNotExist(t *testing.T) {
if val, ok := Doc().Find("div.row-fluid").Attr("href"); ok {
t.Errorf("Expected no value for the href attribute, got %v.", val)
}
}
func TestRemoveAttr(t *testing.T) {
sel := Doc2Clone().Find("div")
sel.RemoveAttr("id")
_, ok := sel.Attr("id")
if ok {
t.Error("Expected there to be no id attributes set")
}
}
func TestSetAttr(t *testing.T) {
sel := Doc2Clone().Find("#main")
sel.SetAttr("id", "not-main")
val, ok := sel.Attr("id")
if !ok {
t.Error("Expected an id attribute on main")
}
if val != "not-main" {
t.Errorf("Expected an attribute id to be not-main, got %s", val)
}
}
func TestSetAttr2(t *testing.T) {
sel := Doc2Clone().Find("#main")
sel.SetAttr("foo", "bar")
val, ok := sel.Attr("foo")
if !ok {
t.Error("Expected an 'foo' attribute on main")
}
if val != "bar" {
t.Errorf("Expected an attribute 'foo' to be 'bar', got '%s'", val)
}
}
func TestText(t *testing.T) {
txt := Doc().Find("h1").Text()
if strings.Trim(txt, " \n\r\t") != "Provok.in" {
t.Errorf("Expected text to be Provok.in, found %s.", txt)
}
}
func TestText2(t *testing.T) {
txt := Doc().Find(".hero-unit .container-fluid .row-fluid:nth-child(1)").Text()
if ok, e := regexp.MatchString(`^\s+Provok\.in\s+Prove your point.\s+$`, txt); !ok || e != nil {
t.Errorf("Expected text to be Provok.in Prove your point., found %s.", txt)
if e != nil {
t.Logf("Error: %s.", e.Error())
}
}
}
func TestText3(t *testing.T) {
txt := Doc().Find(".pvk-gutter").First().Text()
// There's an   character in there...
if ok, e := regexp.MatchString(`^[\s\x{00A0}]+$`, txt); !ok || e != nil {
t.Errorf("Expected spaces, found <%v>.", txt)
if e != nil {
t.Logf("Error: %s.", e.Error())
}
}
}
func TestHtml(t *testing.T) {
txt, e := Doc().Find("h1").Html()
if e != nil {
t.Errorf("Error: %s.", e)
}
if ok, e := regexp.MatchString(`^\s*<a href="/">Provok<span class="green">\.</span><span class="red">i</span>n</a>\s*$`, txt); !ok || e != nil {
t.Errorf("Unexpected HTML content, found %s.", txt)
if e != nil {
t.Logf("Error: %s.", e.Error())
}
}
}
func TestNbsp(t *testing.T) {
src := `<p>Some&nbsp;text</p>`
d, err := NewDocumentFromReader(strings.NewReader(src))
if err != nil {
t.Fatal(err)
}
txt := d.Find("p").Text()
ix := strings.Index(txt, "\u00a0")
if ix != 4 {
t.Errorf("Text: expected a non-breaking space at index 4, got %d", ix)
}
h, err := d.Find("p").Html()
if err != nil {
t.Fatal(err)
}
ix = strings.Index(h, "\u00a0")
if ix != 4 {
t.Errorf("Html: expected a non-breaking space at index 4, got %d", ix)
}
}
func TestAddClass(t *testing.T) {
sel := Doc2Clone().Find("#main")
sel.AddClass("main main main")
// Make sure that class was only added once
if a, ok := sel.Attr("class"); !ok || a != "main" {
t.Error("Expected #main to have class main")
}
}
func TestAddClassSimilar(t *testing.T) {
sel := Doc2Clone().Find("#nf5")
sel.AddClass("odd")
assertClass(t, sel, "odd")
assertClass(t, sel, "odder")
printSel(t, sel.Parent())
}
func TestAddEmptyClass(t *testing.T) {
sel := Doc2Clone().Find("#main")
sel.AddClass("")
// Make sure that class was only added once
if a, ok := sel.Attr("class"); ok {
t.Errorf("Expected #main to not to have a class, have: %s", a)
}
}
func TestAddClasses(t *testing.T) {
sel := Doc2Clone().Find("#main")
sel.AddClass("a b")
// Make sure that class was only added once
if !sel.HasClass("a") || !sel.HasClass("b") {
t.Errorf("#main does not have classes")
}
}
func TestHasClass(t *testing.T) {
sel := Doc().Find("div")
if !sel.HasClass("span12") {
t.Error("Expected at least one div to have class span12.")
}
}
func TestHasClassNone(t *testing.T) {
sel := Doc().Find("h2")
if sel.HasClass("toto") {
t.Error("Expected h1 to have no class.")
}
}
func TestHasClassNotFirst(t *testing.T) {
sel := Doc().Find(".alert")
if !sel.HasClass("alert-error") {
t.Error("Expected .alert to also have class .alert-error.")
}
}
func TestRemoveClass(t *testing.T) {
sel := Doc2Clone().Find("#nf1")
sel.RemoveClass("one row")
if !sel.HasClass("even") || sel.HasClass("one") || sel.HasClass("row") {
classes, _ := sel.Attr("class")
t.Error("Expected #nf1 to have class even, has ", classes)
}
}
func TestRemoveClassSimilar(t *testing.T) {
sel := Doc2Clone().Find("#nf5, #nf6")
assertLength(t, sel.Nodes, 2)
sel.RemoveClass("odd")
assertClass(t, sel.Eq(0), "odder")
printSel(t, sel)
}
func TestRemoveAllClasses(t *testing.T) {
sel := Doc2Clone().Find("#nf1")
sel.RemoveClass()
if a, ok := sel.Attr("class"); ok {
t.Error("All classes were not removed, has ", a)
}
sel = Doc2Clone().Find("#main")
sel.RemoveClass()
if a, ok := sel.Attr("class"); ok {
t.Error("All classes were not removed, has ", a)
}
}
func TestToggleClass(t *testing.T) {
sel := Doc2Clone().Find("#nf1")
sel.ToggleClass("one")
if sel.HasClass("one") {
t.Error("Expected #nf1 to not have class one")
}
sel.ToggleClass("one")
if !sel.HasClass("one") {
t.Error("Expected #nf1 to have class one")
}
sel.ToggleClass("one even row")
if a, ok := sel.Attr("class"); ok {
t.Errorf("Expected #nf1 to have no classes, have %q", a)
}
}