diff --git a/go.mod b/go.mod
index f302ca6ab4f48c1b0d9c115e8e969e9a7ae9fb69..f7ef22234c3d4a440a7a0eef74d15066cafad403 100644
--- a/go.mod
+++ b/go.mod
@@ -3,7 +3,7 @@ module git.autistici.org/ale/crawl
go 1.15
require (
- github.com/PuerkitoBio/goquery v1.7.1
+ github.com/PuerkitoBio/goquery v1.8.0
github.com/PuerkitoBio/purell v0.1.0
github.com/PuerkitoBio/urlesc v0.0.0-20170810143723-de5bf2ad4578 // indirect
github.com/google/go-cmp v0.5.6
diff --git a/go.sum b/go.sum
index b977e7259b9c323c68db509c2cad1e58a8a4b8b4..de34829985a3e3c4acaeeb6c2a6ba3be8e2eeec3 100644
--- a/go.sum
+++ b/go.sum
@@ -2,6 +2,8 @@ github.com/PuerkitoBio/goquery v1.5.0 h1:uGvmFXOA73IKluu/F84Xd1tt/z07GYm8X49XKHP
github.com/PuerkitoBio/goquery v1.5.0/go.mod h1:qD2PgZ9lccMbQlc7eEOjaeRlFQON7xY8kdmcsrnKqMg=
github.com/PuerkitoBio/goquery v1.7.1 h1:oE+T06D+1T7LNrn91B4aERsRIeCLJ/oPSa6xB9FPnz4=
github.com/PuerkitoBio/goquery v1.7.1/go.mod h1:XY0pP4kfraEmmV1O7Uf6XyjoslwsneBbgeDjLYuN8xY=
+github.com/PuerkitoBio/goquery v1.8.0 h1:PJTF7AmFCFKk1N6V6jmKfrNH9tV5pNE6lZMkG0gta/U=
+github.com/PuerkitoBio/goquery v1.8.0/go.mod h1:ypIiRMtY7COPGk+I/YbZLbxsxn9g5ejnI2HSMtkjZvI=
github.com/PuerkitoBio/purell v0.0.0-20180310210909-975f53781597 h1:1H3FyRw7YsqIty9WHPOVEGJaFJ1sfGVZ3PPDUw3ob2w=
github.com/PuerkitoBio/purell v0.0.0-20180310210909-975f53781597/go.mod h1:c11w/QuzBsJSee3cPx9rAFu61PvFxuPbtSwDGJws/X0=
github.com/PuerkitoBio/purell v0.1.0 h1:N8Bcc53nei5frgNYgAKo93qMUVdU5LUGHCBv8efdVcM=
@@ -12,6 +14,8 @@ github.com/andybalholm/cascadia v1.0.0 h1:hOCXnnZ5A+3eVDX8pvgl4kofXv2ELss0bKcqRy
github.com/andybalholm/cascadia v1.0.0/go.mod h1:GsXiBklL0woXo1j/WYWtSYYC4ouU9PqHO0sqidkEA4Y=
github.com/andybalholm/cascadia v1.2.0 h1:vuRCkM5Ozh/BfmsaTm26kbjm0mIOM3yS5Ek/F5h18aE=
github.com/andybalholm/cascadia v1.2.0/go.mod h1:YCyR8vOZT9aZ1CHEd8ap0gMVm2aFgxBp0T0eFw1RUQY=
+github.com/andybalholm/cascadia v1.3.1 h1:nhxRkql1kdYCc8Snf7D5/D3spOX+dBgjA6u8x004T2c=
+github.com/andybalholm/cascadia v1.3.1/go.mod h1:R4bJ1UQfqADjvDa4P6HZHLh/3OxWWEqc0Sk8XGwHqvA=
github.com/fsnotify/fsnotify v1.4.7 h1:IXs+QLmnXW2CcXuY+8Mzv/fWEsPGWxqefPtCP5CnV9I=
github.com/fsnotify/fsnotify v1.4.7/go.mod h1:jwhsz4b93w/PPRr/qN1Yymfu8t87LnFCMoQvtojpjFo=
github.com/golang/protobuf v1.2.0 h1:P3YflyNX/ehuJFLhxviNdFxQPkGK5cDcApsge1SqnvM=
@@ -48,6 +52,8 @@ golang.org/x/net v0.0.0-20190926025831-c00fd9afed17 h1:qPnAdmjNA41t3QBTx2mFGf/SD
golang.org/x/net v0.0.0-20190926025831-c00fd9afed17/go.mod h1:z5CRVTTTmAJ677TzLLGU+0bjPO0LkuOLi4/5GtJWs/s=
golang.org/x/net v0.0.0-20210614182718-04defd469f4e h1:XpT3nA5TvE525Ne3hInMh6+GETgn27Zfm9dxsThnX2Q=
golang.org/x/net v0.0.0-20210614182718-04defd469f4e/go.mod h1:9nx3DQGgdP8bBQD5qxJ1jj9UTztislL4KSBs9R2vV5Y=
+golang.org/x/net v0.0.0-20210916014120-12bc252f5db8 h1:/6y1LfuqNuQdHAm0jjtPtgRcxIxjVZgm5OTu8/QhZvk=
+golang.org/x/net v0.0.0-20210916014120-12bc252f5db8/go.mod h1:9nx3DQGgdP8bBQD5qxJ1jj9UTztislL4KSBs9R2vV5Y=
golang.org/x/sync v0.0.0-20180314180146-1d60e4601c6f h1:wMNYb4v58l5UBM7MYRLPG6ZhfOqbKu7X5eyFl8ZhKvA=
golang.org/x/sync v0.0.0-20180314180146-1d60e4601c6f/go.mod h1:RxMgew5VJxzue5/jJTE5uejpjVlOe/izrB70Jof72aM=
golang.org/x/sys v0.0.0-20180909124046-d0be0721c37e/go.mod h1:STP8DvDyc/dI5b8T5hshtkjS+E42TnysNCUPdjciGhY=
diff --git a/vendor/github.com/PuerkitoBio/goquery/.travis.yml b/vendor/github.com/PuerkitoBio/goquery/.travis.yml
deleted file mode 100644
index 8430c86a2a505815865340e25f933dd4d709318f..0000000000000000000000000000000000000000
--- a/vendor/github.com/PuerkitoBio/goquery/.travis.yml
+++ /dev/null
@@ -1,31 +0,0 @@
-arch:
- - amd64
- - ppc64le
-language: go
-
-go:
- - 1.7.x
- - 1.8.x
- - 1.9.x
- - 1.10.x
- - 1.11.x
- - 1.12.x
- - 1.13.x
- - 1.14.x
- - 1.15.x
- - tip
-
-jobs:
- exclude:
- - arch: ppc64le
- go: 1.7.x
- - arch: ppc64le
- go: 1.8.x
- - arch: ppc64le
- go: 1.9.x
- - arch: ppc64le
- go: 1.10.x
- - arch: ppc64le
- go: 1.11.x
- - arch: ppc64le
- go: 1.12.x
diff --git a/vendor/github.com/PuerkitoBio/goquery/README.md b/vendor/github.com/PuerkitoBio/goquery/README.md
index 6bb185c2cad3e4d0434ccd057c12ead9f6e8d2d4..775223401dd38af858ef7184ed4b0994cce0dab4 100644
--- a/vendor/github.com/PuerkitoBio/goquery/README.md
+++ b/vendor/github.com/PuerkitoBio/goquery/README.md
@@ -1,7 +1,6 @@
# goquery - a little like that j-thing, only in Go
-[](https://builds.sr.ht/~mna/goquery/commits/fedora.yml?)
-[](http://travis-ci.org/PuerkitoBio/goquery)
+[](https://github.com/PuerkitoBio/goquery/actions)
[](https://pkg.go.dev/github.com/PuerkitoBio/goquery)
[](https://sourcegraph.com/github.com/PuerkitoBio/goquery?badge)
@@ -41,6 +40,7 @@ Please note that because of the net/html dependency, goquery requires Go1.1+ and
**Note that goquery's API is now stable, and will not break.**
+* **2021-10-25 (v1.8.0)** : Add `Render` function to render a `Selection` to an `io.Writer` (thanks [@anthonygedeon](https://github.com/anthonygedeon)).
* **2021-07-11 (v1.7.1)** : Update go.mod dependencies and add dependabot config (thanks [@jauderho](https://github.com/jauderho)).
* **2021-06-14 (v1.7.0)** : Add `Single` and `SingleMatcher` functions to optimize first-match selection (thanks [@gdollardollar](https://github.com/gdollardollar)).
* **2021-01-11 (v1.6.1)** : Fix panic when calling `{Prepend,Append,Set}Html` on a `Selection` that contains non-Element nodes.
diff --git a/vendor/github.com/PuerkitoBio/goquery/go.mod b/vendor/github.com/PuerkitoBio/goquery/go.mod
index 3af3b154db61ca88cce58faa90678cc16f8062c2..4b5a30963a39fdcf6e9afd73fbf298801340fa81 100644
--- a/vendor/github.com/PuerkitoBio/goquery/go.mod
+++ b/vendor/github.com/PuerkitoBio/goquery/go.mod
@@ -1,8 +1,8 @@
module github.com/PuerkitoBio/goquery
require (
- github.com/andybalholm/cascadia v1.2.0
- golang.org/x/net v0.0.0-20210614182718-04defd469f4e
+ github.com/andybalholm/cascadia v1.3.1
+ golang.org/x/net v0.0.0-20210916014120-12bc252f5db8
)
go 1.13
diff --git a/vendor/github.com/PuerkitoBio/goquery/go.sum b/vendor/github.com/PuerkitoBio/goquery/go.sum
index 13b128d4e55f0a7d99f041bc7ae03dd3b46ed6d1..167f12d2db63d97a0dbb16fead68d9478eea6378 100644
--- a/vendor/github.com/PuerkitoBio/goquery/go.sum
+++ b/vendor/github.com/PuerkitoBio/goquery/go.sum
@@ -1,8 +1,7 @@
-github.com/andybalholm/cascadia v1.2.0 h1:vuRCkM5Ozh/BfmsaTm26kbjm0mIOM3yS5Ek/F5h18aE=
-github.com/andybalholm/cascadia v1.2.0/go.mod h1:YCyR8vOZT9aZ1CHEd8ap0gMVm2aFgxBp0T0eFw1RUQY=
-golang.org/x/net v0.0.0-20180218175443-cbe0f9307d01/go.mod h1:mL1N/T3taQHkDXs73rZJwtUhF3w3ftmwwsq0BUmARs4=
-golang.org/x/net v0.0.0-20210614182718-04defd469f4e h1:XpT3nA5TvE525Ne3hInMh6+GETgn27Zfm9dxsThnX2Q=
-golang.org/x/net v0.0.0-20210614182718-04defd469f4e/go.mod h1:9nx3DQGgdP8bBQD5qxJ1jj9UTztislL4KSBs9R2vV5Y=
+github.com/andybalholm/cascadia v1.3.1 h1:nhxRkql1kdYCc8Snf7D5/D3spOX+dBgjA6u8x004T2c=
+github.com/andybalholm/cascadia v1.3.1/go.mod h1:R4bJ1UQfqADjvDa4P6HZHLh/3OxWWEqc0Sk8XGwHqvA=
+golang.org/x/net v0.0.0-20210916014120-12bc252f5db8 h1:/6y1LfuqNuQdHAm0jjtPtgRcxIxjVZgm5OTu8/QhZvk=
+golang.org/x/net v0.0.0-20210916014120-12bc252f5db8/go.mod h1:9nx3DQGgdP8bBQD5qxJ1jj9UTztislL4KSBs9R2vV5Y=
golang.org/x/sys v0.0.0-20201119102817-f84b799fce68/go.mod h1:h1NjWce9XRLGQEsW7wpKNCjG9DtNlClVuFLEZdDNbEs=
golang.org/x/sys v0.0.0-20210423082822-04245dca01da/go.mod h1:h1NjWce9XRLGQEsW7wpKNCjG9DtNlClVuFLEZdDNbEs=
golang.org/x/term v0.0.0-20201126162022-7de9c90e9dd1/go.mod h1:bj7SfCRtBDWHUb9snDiAeCFNEtKQo2Wmx5Cou7ajbmo=
diff --git a/vendor/github.com/PuerkitoBio/goquery/utilities.go b/vendor/github.com/PuerkitoBio/goquery/utilities.go
index 3e11b1db1b42401308a8feda729ec3ef5ce133dd..6d243cdd63ad55f3fbb728d864fe9e6a5eaf8a02 100644
--- a/vendor/github.com/PuerkitoBio/goquery/utilities.go
+++ b/vendor/github.com/PuerkitoBio/goquery/utilities.go
@@ -2,6 +2,7 @@ package goquery
import (
"bytes"
+ "io"
"golang.org/x/net/html"
)
@@ -50,13 +51,24 @@ func nodeName(node *html.Node) string {
case html.ElementNode, html.DoctypeNode:
return node.Data
default:
- if node.Type >= 0 && int(node.Type) < len(nodeNames) {
+ if int(node.Type) < len(nodeNames) {
return nodeNames[node.Type]
}
return ""
}
}
+// Render renders the html of the first element from selector and writes it to
+// the writer. It behaves the same as OuterHtml but writes to w instead of
+// returning the string.
+func Render(w io.Writer, s *Selection) error {
+ if s.Length() == 0 {
+ return nil
+ }
+ n := s.Get(0)
+ return html.Render(w, n)
+}
+
// OuterHtml returns the outer HTML rendering of the first item in
// the selection - that is, the HTML including the first element's
// tag and attributes.
@@ -66,12 +78,7 @@ func nodeName(node *html.Node) string {
// a property provided by the DOM).
func OuterHtml(s *Selection) (string, error) {
var buf bytes.Buffer
-
- if s.Length() == 0 {
- return "", nil
- }
- n := s.Get(0)
- if err := html.Render(&buf, n); err != nil {
+ if err := Render(&buf, s); err != nil {
return "", err
}
return buf.String(), nil
diff --git a/vendor/github.com/andybalholm/cascadia/go.mod b/vendor/github.com/andybalholm/cascadia/go.mod
index 51a330b50004eb7d7d749c5e9b8c233fad107621..ee72d3fe0c8d726656b1fc83fa206a8e98dd1e74 100644
--- a/vendor/github.com/andybalholm/cascadia/go.mod
+++ b/vendor/github.com/andybalholm/cascadia/go.mod
@@ -1,5 +1,5 @@
module github.com/andybalholm/cascadia
-require golang.org/x/net v0.0.0-20180218175443-cbe0f9307d01
+go 1.16
-go 1.13
+require golang.org/x/net v0.0.0-20210916014120-12bc252f5db8
diff --git a/vendor/github.com/andybalholm/cascadia/go.sum b/vendor/github.com/andybalholm/cascadia/go.sum
new file mode 100644
index 0000000000000000000000000000000000000000..0f4194c577d5cad37b4e58de9303e66984630c7a
--- /dev/null
+++ b/vendor/github.com/andybalholm/cascadia/go.sum
@@ -0,0 +1,7 @@
+golang.org/x/net v0.0.0-20210916014120-12bc252f5db8 h1:/6y1LfuqNuQdHAm0jjtPtgRcxIxjVZgm5OTu8/QhZvk=
+golang.org/x/net v0.0.0-20210916014120-12bc252f5db8/go.mod h1:9nx3DQGgdP8bBQD5qxJ1jj9UTztislL4KSBs9R2vV5Y=
+golang.org/x/sys v0.0.0-20201119102817-f84b799fce68/go.mod h1:h1NjWce9XRLGQEsW7wpKNCjG9DtNlClVuFLEZdDNbEs=
+golang.org/x/sys v0.0.0-20210423082822-04245dca01da/go.mod h1:h1NjWce9XRLGQEsW7wpKNCjG9DtNlClVuFLEZdDNbEs=
+golang.org/x/term v0.0.0-20201126162022-7de9c90e9dd1/go.mod h1:bj7SfCRtBDWHUb9snDiAeCFNEtKQo2Wmx5Cou7ajbmo=
+golang.org/x/text v0.3.6/go.mod h1:5Zoc/QRtKVWzQhOtBMvqHzDpF6irO9z98xDceosuGiQ=
+golang.org/x/tools v0.0.0-20180917221912-90fa682c2a6e/go.mod h1:n7NCudcB/nEzxVGmLbDWY5pfWTLqBcC2KZ6jyYvM4mQ=
diff --git a/vendor/github.com/andybalholm/cascadia/parser.go b/vendor/github.com/andybalholm/cascadia/parser.go
index c40a39fed16e1e0a8c365fc3dc59955ae46dc7ad..f654c0c7a81dd6e08c41dec98a2b4a0a418906a3 100644
--- a/vendor/github.com/andybalholm/cascadia/parser.go
+++ b/vendor/github.com/andybalholm/cascadia/parser.go
@@ -36,7 +36,7 @@ func (p *parser) parseEscape() (result string, err error) {
for i = start; i < start+6 && i < len(p.s) && hexDigit(p.s[i]); i++ {
// empty
}
- v, _ := strconv.ParseUint(p.s[start:i], 16, 21)
+ v, _ := strconv.ParseUint(p.s[start:i], 16, 64)
if len(p.s) > i {
switch p.s[i] {
case '\r':
@@ -409,6 +409,19 @@ func (p *parser) parseAttributeSelector() (attrSelector, error) {
if p.i >= len(p.s) {
return attrSelector{}, errors.New("unexpected EOF in attribute selector")
}
+
+ // check if the attribute contains an ignore case flag
+ ignoreCase := false
+ if p.s[p.i] == 'i' || p.s[p.i] == 'I' {
+ ignoreCase = true
+ p.i++
+ }
+
+ p.skipWhitespace()
+ if p.i >= len(p.s) {
+ return attrSelector{}, errors.New("unexpected EOF in attribute selector")
+ }
+
if p.s[p.i] != ']' {
return attrSelector{}, fmt.Errorf("expected ']', found '%c' instead", p.s[p.i])
}
@@ -416,15 +429,17 @@ func (p *parser) parseAttributeSelector() (attrSelector, error) {
switch op {
case "=", "!=", "~=", "|=", "^=", "$=", "*=", "#=":
- return attrSelector{key: key, val: val, operation: op, regexp: rx}, nil
+ return attrSelector{key: key, val: val, operation: op, regexp: rx, insensitive: ignoreCase}, nil
default:
return attrSelector{}, fmt.Errorf("attribute operator %q is not supported", op)
}
}
-var errExpectedParenthesis = errors.New("expected '(' but didn't find it")
-var errExpectedClosingParenthesis = errors.New("expected ')' but didn't find it")
-var errUnmatchedParenthesis = errors.New("unmatched '('")
+var (
+ errExpectedParenthesis = errors.New("expected '(' but didn't find it")
+ errExpectedClosingParenthesis = errors.New("expected ')' but didn't find it")
+ errUnmatchedParenthesis = errors.New("unmatched '('")
+)
// parsePseudoclassSelector parses a pseudoclass selector like :not(p) or a pseudo-element
// For backwards compatibility, both ':' and '::' prefix are allowed for pseudo-elements.
@@ -552,6 +567,37 @@ func (p *parser) parsePseudoclassSelector() (out Sel, pseudoElement string, err
out = emptyElementPseudoClassSelector{}
case "root":
out = rootPseudoClassSelector{}
+ case "link":
+ out = linkPseudoClassSelector{}
+ case "lang":
+ if !p.consumeParenthesis() {
+ return out, "", errExpectedParenthesis
+ }
+ if p.i == len(p.s) {
+ return out, "", errUnmatchedParenthesis
+ }
+ val, err := p.parseIdentifier()
+ if err != nil {
+ return out, "", err
+ }
+ val = strings.ToLower(val)
+ p.skipWhitespace()
+ if p.i >= len(p.s) {
+ return out, "", errors.New("unexpected EOF in pseudo selector")
+ }
+ if !p.consumeClosingParenthesis() {
+ return out, "", errExpectedClosingParenthesis
+ }
+ out = langPseudoClassSelector{lang: val}
+ case "enabled":
+ out = enabledPseudoClassSelector{}
+ case "disabled":
+ out = disabledPseudoClassSelector{}
+ case "checked":
+ out = checkedPseudoClassSelector{}
+ case "visited", "hover", "active", "focus", "target":
+ // Not applicable in a static context: never match.
+ out = neverMatchSelector{value: ":" + name}
case "after", "backdrop", "before", "cue", "first-letter", "first-line", "grammar-error", "marker", "placeholder", "selection", "spelling-error":
return nil, name, nil
default:
@@ -714,6 +760,9 @@ func (p *parser) parseSimpleSelectorSequence() (Sel, error) {
case '*':
// It's the universal selector. Just skip over it, since it doesn't affect the meaning.
p.i++
+ if p.i+2 < len(p.s) && p.s[p.i:p.i+2] == "|*" { // other version of universal selector
+ p.i += 2
+ }
case '#', '.', '[', ':':
// There's no type selector. Wait to process the other till the main loop.
default:
diff --git a/vendor/github.com/andybalholm/cascadia/pseudo_classes.go b/vendor/github.com/andybalholm/cascadia/pseudo_classes.go
new file mode 100644
index 0000000000000000000000000000000000000000..3986b22cd4e27097da7d7a662f809a8949530346
--- /dev/null
+++ b/vendor/github.com/andybalholm/cascadia/pseudo_classes.go
@@ -0,0 +1,474 @@
+package cascadia
+
+import (
+ "bytes"
+ "fmt"
+ "regexp"
+ "strings"
+
+ "golang.org/x/net/html"
+ "golang.org/x/net/html/atom"
+)
+
+// This file implements the pseudo classes selectors,
+// which share the implementation of PseudoElement() and Specificity()
+
+type abstractPseudoClass struct{}
+
+func (s abstractPseudoClass) Specificity() Specificity {
+ return Specificity{0, 1, 0}
+}
+
+func (c abstractPseudoClass) PseudoElement() string {
+ return ""
+}
+
+type relativePseudoClassSelector struct {
+ name string // one of "not", "has", "haschild"
+ match SelectorGroup
+}
+
+func (s relativePseudoClassSelector) Match(n *html.Node) bool {
+ if n.Type != html.ElementNode {
+ return false
+ }
+ switch s.name {
+ case "not":
+ // matches elements that do not match a.
+ return !s.match.Match(n)
+ case "has":
+ // matches elements with any descendant that matches a.
+ return hasDescendantMatch(n, s.match)
+ case "haschild":
+ // matches elements with a child that matches a.
+ return hasChildMatch(n, s.match)
+ default:
+ panic(fmt.Sprintf("unsupported relative pseudo class selector : %s", s.name))
+ }
+}
+
+// hasChildMatch returns whether n has any child that matches a.
+func hasChildMatch(n *html.Node, a Matcher) bool {
+ for c := n.FirstChild; c != nil; c = c.NextSibling {
+ if a.Match(c) {
+ return true
+ }
+ }
+ return false
+}
+
+// hasDescendantMatch performs a depth-first search of n's descendants,
+// testing whether any of them match a. It returns true as soon as a match is
+// found, or false if no match is found.
+func hasDescendantMatch(n *html.Node, a Matcher) bool {
+ for c := n.FirstChild; c != nil; c = c.NextSibling {
+ if a.Match(c) || (c.Type == html.ElementNode && hasDescendantMatch(c, a)) {
+ return true
+ }
+ }
+ return false
+}
+
+// Specificity returns the specificity of the most specific selectors
+// in the pseudo-class arguments.
+// See https://www.w3.org/TR/selectors/#specificity-rules
+func (s relativePseudoClassSelector) Specificity() Specificity {
+ var max Specificity
+ for _, sel := range s.match {
+ newSpe := sel.Specificity()
+ if max.Less(newSpe) {
+ max = newSpe
+ }
+ }
+ return max
+}
+
+func (c relativePseudoClassSelector) PseudoElement() string {
+ return ""
+}
+
+type containsPseudoClassSelector struct {
+ abstractPseudoClass
+ value string
+ own bool
+}
+
+func (s containsPseudoClassSelector) Match(n *html.Node) bool {
+ var text string
+ if s.own {
+ // matches nodes that directly contain the given text
+ text = strings.ToLower(nodeOwnText(n))
+ } else {
+ // matches nodes that contain the given text.
+ text = strings.ToLower(nodeText(n))
+ }
+ return strings.Contains(text, s.value)
+}
+
+type regexpPseudoClassSelector struct {
+ abstractPseudoClass
+ regexp *regexp.Regexp
+ own bool
+}
+
+func (s regexpPseudoClassSelector) Match(n *html.Node) bool {
+ var text string
+ if s.own {
+ // matches nodes whose text directly matches the specified regular expression
+ text = nodeOwnText(n)
+ } else {
+ // matches nodes whose text matches the specified regular expression
+ text = nodeText(n)
+ }
+ return s.regexp.MatchString(text)
+}
+
+// writeNodeText writes the text contained in n and its descendants to b.
+func writeNodeText(n *html.Node, b *bytes.Buffer) {
+ switch n.Type {
+ case html.TextNode:
+ b.WriteString(n.Data)
+ case html.ElementNode:
+ for c := n.FirstChild; c != nil; c = c.NextSibling {
+ writeNodeText(c, b)
+ }
+ }
+}
+
+// nodeText returns the text contained in n and its descendants.
+func nodeText(n *html.Node) string {
+ var b bytes.Buffer
+ writeNodeText(n, &b)
+ return b.String()
+}
+
+// nodeOwnText returns the contents of the text nodes that are direct
+// children of n.
+func nodeOwnText(n *html.Node) string {
+ var b bytes.Buffer
+ for c := n.FirstChild; c != nil; c = c.NextSibling {
+ if c.Type == html.TextNode {
+ b.WriteString(c.Data)
+ }
+ }
+ return b.String()
+}
+
+type nthPseudoClassSelector struct {
+ abstractPseudoClass
+ a, b int
+ last, ofType bool
+}
+
+func (s nthPseudoClassSelector) Match(n *html.Node) bool {
+ if s.a == 0 {
+ if s.last {
+ return simpleNthLastChildMatch(s.b, s.ofType, n)
+ } else {
+ return simpleNthChildMatch(s.b, s.ofType, n)
+ }
+ }
+ return nthChildMatch(s.a, s.b, s.last, s.ofType, n)
+}
+
+// nthChildMatch implements :nth-child(an+b).
+// If last is true, implements :nth-last-child instead.
+// If ofType is true, implements :nth-of-type instead.
+func nthChildMatch(a, b int, last, ofType bool, n *html.Node) bool {
+ if n.Type != html.ElementNode {
+ return false
+ }
+
+ parent := n.Parent
+ if parent == nil {
+ return false
+ }
+
+ if parent.Type == html.DocumentNode {
+ return false
+ }
+
+ i := -1
+ count := 0
+ for c := parent.FirstChild; c != nil; c = c.NextSibling {
+ if (c.Type != html.ElementNode) || (ofType && c.Data != n.Data) {
+ continue
+ }
+ count++
+ if c == n {
+ i = count
+ if !last {
+ break
+ }
+ }
+ }
+
+ if i == -1 {
+ // This shouldn't happen, since n should always be one of its parent's children.
+ return false
+ }
+
+ if last {
+ i = count - i + 1
+ }
+
+ i -= b
+ if a == 0 {
+ return i == 0
+ }
+
+ return i%a == 0 && i/a >= 0
+}
+
+// simpleNthChildMatch implements :nth-child(b).
+// If ofType is true, implements :nth-of-type instead.
+func simpleNthChildMatch(b int, ofType bool, n *html.Node) bool {
+ if n.Type != html.ElementNode {
+ return false
+ }
+
+ parent := n.Parent
+ if parent == nil {
+ return false
+ }
+
+ if parent.Type == html.DocumentNode {
+ return false
+ }
+
+ count := 0
+ for c := parent.FirstChild; c != nil; c = c.NextSibling {
+ if c.Type != html.ElementNode || (ofType && c.Data != n.Data) {
+ continue
+ }
+ count++
+ if c == n {
+ return count == b
+ }
+ if count >= b {
+ return false
+ }
+ }
+ return false
+}
+
+// simpleNthLastChildMatch implements :nth-last-child(b).
+// If ofType is true, implements :nth-last-of-type instead.
+func simpleNthLastChildMatch(b int, ofType bool, n *html.Node) bool {
+ if n.Type != html.ElementNode {
+ return false
+ }
+
+ parent := n.Parent
+ if parent == nil {
+ return false
+ }
+
+ if parent.Type == html.DocumentNode {
+ return false
+ }
+
+ count := 0
+ for c := parent.LastChild; c != nil; c = c.PrevSibling {
+ if c.Type != html.ElementNode || (ofType && c.Data != n.Data) {
+ continue
+ }
+ count++
+ if c == n {
+ return count == b
+ }
+ if count >= b {
+ return false
+ }
+ }
+ return false
+}
+
+type onlyChildPseudoClassSelector struct {
+ abstractPseudoClass
+ ofType bool
+}
+
+// Match implements :only-child.
+// If `ofType` is true, it implements :only-of-type instead.
+func (s onlyChildPseudoClassSelector) Match(n *html.Node) bool {
+ if n.Type != html.ElementNode {
+ return false
+ }
+
+ parent := n.Parent
+ if parent == nil {
+ return false
+ }
+
+ if parent.Type == html.DocumentNode {
+ return false
+ }
+
+ count := 0
+ for c := parent.FirstChild; c != nil; c = c.NextSibling {
+ if (c.Type != html.ElementNode) || (s.ofType && c.Data != n.Data) {
+ continue
+ }
+ count++
+ if count > 1 {
+ return false
+ }
+ }
+
+ return count == 1
+}
+
+type inputPseudoClassSelector struct {
+ abstractPseudoClass
+}
+
+// Matches input, select, textarea and button elements.
+func (s inputPseudoClassSelector) Match(n *html.Node) bool {
+ return n.Type == html.ElementNode && (n.Data == "input" || n.Data == "select" || n.Data == "textarea" || n.Data == "button")
+}
+
+type emptyElementPseudoClassSelector struct {
+ abstractPseudoClass
+}
+
+// Matches empty elements.
+func (s emptyElementPseudoClassSelector) Match(n *html.Node) bool {
+ if n.Type != html.ElementNode {
+ return false
+ }
+
+ for c := n.FirstChild; c != nil; c = c.NextSibling {
+ switch c.Type {
+ case html.ElementNode:
+ return false
+ case html.TextNode:
+ if strings.TrimSpace(nodeText(c)) == "" {
+ continue
+ } else {
+ return false
+ }
+ }
+ }
+
+ return true
+}
+
+type rootPseudoClassSelector struct {
+ abstractPseudoClass
+}
+
+// Match implements :root
+func (s rootPseudoClassSelector) Match(n *html.Node) bool {
+ if n.Type != html.ElementNode {
+ return false
+ }
+ if n.Parent == nil {
+ return false
+ }
+ return n.Parent.Type == html.DocumentNode
+}
+
+func hasAttr(n *html.Node, attr string) bool {
+ return matchAttribute(n, attr, func(string) bool { return true })
+}
+
+type linkPseudoClassSelector struct {
+ abstractPseudoClass
+}
+
+// Match implements :link
+func (s linkPseudoClassSelector) Match(n *html.Node) bool {
+ return (n.DataAtom == atom.A || n.DataAtom == atom.Area || n.DataAtom == atom.Link) && hasAttr(n, "href")
+}
+
+type langPseudoClassSelector struct {
+ abstractPseudoClass
+ lang string
+}
+
+func (s langPseudoClassSelector) Match(n *html.Node) bool {
+ own := matchAttribute(n, "lang", func(val string) bool {
+ return val == s.lang || strings.HasPrefix(val, s.lang+"-")
+ })
+ if n.Parent == nil {
+ return own
+ }
+ return own || s.Match(n.Parent)
+}
+
+type enabledPseudoClassSelector struct {
+ abstractPseudoClass
+}
+
+func (s enabledPseudoClassSelector) Match(n *html.Node) bool {
+ if n.Type != html.ElementNode {
+ return false
+ }
+ switch n.DataAtom {
+ case atom.A, atom.Area, atom.Link:
+ return hasAttr(n, "href")
+ case atom.Optgroup, atom.Menuitem, atom.Fieldset:
+ return !hasAttr(n, "disabled")
+ case atom.Button, atom.Input, atom.Select, atom.Textarea, atom.Option:
+ return !hasAttr(n, "disabled") && !inDisabledFieldset(n)
+ }
+ return false
+}
+
+type disabledPseudoClassSelector struct {
+ abstractPseudoClass
+}
+
+func (s disabledPseudoClassSelector) Match(n *html.Node) bool {
+ if n.Type != html.ElementNode {
+ return false
+ }
+ switch n.DataAtom {
+ case atom.Optgroup, atom.Menuitem, atom.Fieldset:
+ return hasAttr(n, "disabled")
+ case atom.Button, atom.Input, atom.Select, atom.Textarea, atom.Option:
+ return hasAttr(n, "disabled") || inDisabledFieldset(n)
+ }
+ return false
+}
+
+func hasLegendInPreviousSiblings(n *html.Node) bool {
+ for s := n.PrevSibling; s != nil; s = s.PrevSibling {
+ if s.DataAtom == atom.Legend {
+ return true
+ }
+ }
+ return false
+}
+
+func inDisabledFieldset(n *html.Node) bool {
+ if n.Parent == nil {
+ return false
+ }
+ if n.Parent.DataAtom == atom.Fieldset && hasAttr(n.Parent, "disabled") &&
+ (n.DataAtom != atom.Legend || hasLegendInPreviousSiblings(n)) {
+ return true
+ }
+ return inDisabledFieldset(n.Parent)
+}
+
+type checkedPseudoClassSelector struct {
+ abstractPseudoClass
+}
+
+func (s checkedPseudoClassSelector) Match(n *html.Node) bool {
+ if n.Type != html.ElementNode {
+ return false
+ }
+ switch n.DataAtom {
+ case atom.Input, atom.Menuitem:
+ return hasAttr(n, "checked") && matchAttribute(n, "type", func(val string) bool {
+ t := toLowerASCII(val)
+ return t == "checkbox" || t == "radio"
+ })
+ case atom.Option:
+ return hasAttr(n, "selected")
+ }
+ return false
+}
diff --git a/vendor/github.com/andybalholm/cascadia/selector.go b/vendor/github.com/andybalholm/cascadia/selector.go
index e2a6dc4be16f99b89c54fbb94d1c609db571307b..87549be23958dfc5948fb4ea20ac998c02ea8d8a 100644
--- a/vendor/github.com/andybalholm/cascadia/selector.go
+++ b/vendor/github.com/andybalholm/cascadia/selector.go
@@ -1,7 +1,6 @@
package cascadia
import (
- "bytes"
"fmt"
"regexp"
"strings"
@@ -232,7 +231,7 @@ type classSelector struct {
// Matches elements by class attribute.
func (t classSelector) Match(n *html.Node) bool {
return matchAttribute(n, "class", func(s string) bool {
- return matchInclude(t.class, s)
+ return matchInclude(t.class, s, false)
})
}
@@ -266,6 +265,7 @@ func (c idSelector) PseudoElement() string {
type attrSelector struct {
key, val, operation string
regexp *regexp.Regexp
+ insensitive bool
}
// Matches elements by attribute value.
@@ -274,20 +274,20 @@ func (t attrSelector) Match(n *html.Node) bool {
case "":
return matchAttribute(n, t.key, func(string) bool { return true })
case "=":
- return matchAttribute(n, t.key, func(s string) bool { return s == t.val })
+ return matchAttribute(n, t.key, func(s string) bool { return matchInsensitiveValue(s, t.val, t.insensitive) })
case "!=":
- return attributeNotEqualMatch(t.key, t.val, n)
+ return attributeNotEqualMatch(t.key, t.val, n, t.insensitive)
case "~=":
// matches elements where the attribute named key is a whitespace-separated list that includes val.
- return matchAttribute(n, t.key, func(s string) bool { return matchInclude(t.val, s) })
+ return matchAttribute(n, t.key, func(s string) bool { return matchInclude(t.val, s, t.insensitive) })
case "|=":
- return attributeDashMatch(t.key, t.val, n)
+ return attributeDashMatch(t.key, t.val, n, t.insensitive)
case "^=":
- return attributePrefixMatch(t.key, t.val, n)
+ return attributePrefixMatch(t.key, t.val, n, t.insensitive)
case "$=":
- return attributeSuffixMatch(t.key, t.val, n)
+ return attributeSuffixMatch(t.key, t.val, n, t.insensitive)
case "*=":
- return attributeSubstringMatch(t.key, t.val, n)
+ return attributeSubstringMatch(t.key, t.val, n, t.insensitive)
case "#=":
return attributeRegexMatch(t.key, t.regexp, n)
default:
@@ -295,6 +295,17 @@ func (t attrSelector) Match(n *html.Node) bool {
}
}
+// matches elements where we ignore (or not) the case of the attribute value
+// the user attribute is the value set by the user to match elements
+// the real attribute is the attribute value found in the code parsed
+func matchInsensitiveValue(userAttr string, realAttr string, ignoreCase bool) bool {
+ if ignoreCase {
+ return strings.EqualFold(userAttr, realAttr)
+ }
+ return userAttr == realAttr
+
+}
+
// matches elements where the attribute named key satisifes the function f.
func matchAttribute(n *html.Node, key string, f func(string) bool) bool {
if n.Type != html.ElementNode {
@@ -310,12 +321,12 @@ func matchAttribute(n *html.Node, key string, f func(string) bool) bool {
// attributeNotEqualMatch matches elements where
// the attribute named key does not have the value val.
-func attributeNotEqualMatch(key, val string, n *html.Node) bool {
+func attributeNotEqualMatch(key, val string, n *html.Node, ignoreCase bool) bool {
if n.Type != html.ElementNode {
return false
}
for _, a := range n.Attr {
- if a.Key == key && a.Val == val {
+ if a.Key == key && matchInsensitiveValue(a.Val, val, ignoreCase) {
return false
}
}
@@ -323,13 +334,13 @@ func attributeNotEqualMatch(key, val string, n *html.Node) bool {
}
// returns true if s is a whitespace-separated list that includes val.
-func matchInclude(val, s string) bool {
+func matchInclude(val string, s string, ignoreCase bool) bool {
for s != "" {
i := strings.IndexAny(s, " \t\r\n\f")
if i == -1 {
- return s == val
+ return matchInsensitiveValue(s, val, ignoreCase)
}
- if s[:i] == val {
+ if matchInsensitiveValue(s[:i], val, ignoreCase) {
return true
}
s = s[i+1:]
@@ -338,16 +349,16 @@ func matchInclude(val, s string) bool {
}
// matches elements where the attribute named key equals val or starts with val plus a hyphen.
-func attributeDashMatch(key, val string, n *html.Node) bool {
+func attributeDashMatch(key, val string, n *html.Node, ignoreCase bool) bool {
return matchAttribute(n, key,
func(s string) bool {
- if s == val {
+ if matchInsensitiveValue(s, val, ignoreCase) {
return true
}
if len(s) <= len(val) {
return false
}
- if s[:len(val)] == val && s[len(val)] == '-' {
+ if matchInsensitiveValue(s[:len(val)], val, ignoreCase) && s[len(val)] == '-' {
return true
}
return false
@@ -356,36 +367,45 @@ func attributeDashMatch(key, val string, n *html.Node) bool {
// attributePrefixMatch returns a Selector that matches elements where
// the attribute named key starts with val.
-func attributePrefixMatch(key, val string, n *html.Node) bool {
+func attributePrefixMatch(key, val string, n *html.Node, ignoreCase bool) bool {
return matchAttribute(n, key,
func(s string) bool {
if strings.TrimSpace(s) == "" {
return false
}
+ if ignoreCase {
+ return strings.HasPrefix(strings.ToLower(s), strings.ToLower(val))
+ }
return strings.HasPrefix(s, val)
})
}
// attributeSuffixMatch matches elements where
// the attribute named key ends with val.
-func attributeSuffixMatch(key, val string, n *html.Node) bool {
+func attributeSuffixMatch(key, val string, n *html.Node, ignoreCase bool) bool {
return matchAttribute(n, key,
func(s string) bool {
if strings.TrimSpace(s) == "" {
return false
}
+ if ignoreCase {
+ return strings.HasSuffix(strings.ToLower(s), strings.ToLower(val))
+ }
return strings.HasSuffix(s, val)
})
}
// attributeSubstringMatch matches nodes where
// the attribute named key contains val.
-func attributeSubstringMatch(key, val string, n *html.Node) bool {
+func attributeSubstringMatch(key, val string, n *html.Node, ignoreCase bool) bool {
return matchAttribute(n, key,
func(s string) bool {
if strings.TrimSpace(s) == "" {
return false
}
+ if ignoreCase {
+ return strings.Contains(strings.ToLower(s), strings.ToLower(val))
+ }
return strings.Contains(s, val)
})
}
@@ -407,394 +427,22 @@ func (c attrSelector) PseudoElement() string {
return ""
}
-// ---------------- Pseudo class selectors ----------------
-// we use severals concrete types of pseudo-class selectors
-
-type relativePseudoClassSelector struct {
- name string // one of "not", "has", "haschild"
- match SelectorGroup
-}
-
-func (s relativePseudoClassSelector) Match(n *html.Node) bool {
- if n.Type != html.ElementNode {
- return false
- }
- switch s.name {
- case "not":
- // matches elements that do not match a.
- return !s.match.Match(n)
- case "has":
- // matches elements with any descendant that matches a.
- return hasDescendantMatch(n, s.match)
- case "haschild":
- // matches elements with a child that matches a.
- return hasChildMatch(n, s.match)
- default:
- panic(fmt.Sprintf("unsupported relative pseudo class selector : %s", s.name))
- }
-}
-
-// hasChildMatch returns whether n has any child that matches a.
-func hasChildMatch(n *html.Node, a Matcher) bool {
- for c := n.FirstChild; c != nil; c = c.NextSibling {
- if a.Match(c) {
- return true
- }
- }
- return false
-}
-
-// hasDescendantMatch performs a depth-first search of n's descendants,
-// testing whether any of them match a. It returns true as soon as a match is
-// found, or false if no match is found.
-func hasDescendantMatch(n *html.Node, a Matcher) bool {
- for c := n.FirstChild; c != nil; c = c.NextSibling {
- if a.Match(c) || (c.Type == html.ElementNode && hasDescendantMatch(c, a)) {
- return true
- }
- }
- return false
-}
-
-// Specificity returns the specificity of the most specific selectors
-// in the pseudo-class arguments.
-// See https://www.w3.org/TR/selectors/#specificity-rules
-func (s relativePseudoClassSelector) Specificity() Specificity {
- var max Specificity
- for _, sel := range s.match {
- newSpe := sel.Specificity()
- if max.Less(newSpe) {
- max = newSpe
- }
- }
- return max
-}
-
-func (c relativePseudoClassSelector) PseudoElement() string {
- return ""
-}
+// see pseudo_classes.go for pseudo classes selectors
-type containsPseudoClassSelector struct {
- own bool
+// on a static context, some selectors can't match anything
+type neverMatchSelector struct {
value string
}
-func (s containsPseudoClassSelector) Match(n *html.Node) bool {
- var text string
- if s.own {
- // matches nodes that directly contain the given text
- text = strings.ToLower(nodeOwnText(n))
- } else {
- // matches nodes that contain the given text.
- text = strings.ToLower(nodeText(n))
- }
- return strings.Contains(text, s.value)
-}
-
-func (s containsPseudoClassSelector) Specificity() Specificity {
- return Specificity{0, 1, 0}
-}
-
-func (c containsPseudoClassSelector) PseudoElement() string {
- return ""
-}
-
-type regexpPseudoClassSelector struct {
- own bool
- regexp *regexp.Regexp
-}
-
-func (s regexpPseudoClassSelector) Match(n *html.Node) bool {
- var text string
- if s.own {
- // matches nodes whose text directly matches the specified regular expression
- text = nodeOwnText(n)
- } else {
- // matches nodes whose text matches the specified regular expression
- text = nodeText(n)
- }
- return s.regexp.MatchString(text)
-}
-
-// writeNodeText writes the text contained in n and its descendants to b.
-func writeNodeText(n *html.Node, b *bytes.Buffer) {
- switch n.Type {
- case html.TextNode:
- b.WriteString(n.Data)
- case html.ElementNode:
- for c := n.FirstChild; c != nil; c = c.NextSibling {
- writeNodeText(c, b)
- }
- }
-}
-
-// nodeText returns the text contained in n and its descendants.
-func nodeText(n *html.Node) string {
- var b bytes.Buffer
- writeNodeText(n, &b)
- return b.String()
-}
-
-// nodeOwnText returns the contents of the text nodes that are direct
-// children of n.
-func nodeOwnText(n *html.Node) string {
- var b bytes.Buffer
- for c := n.FirstChild; c != nil; c = c.NextSibling {
- if c.Type == html.TextNode {
- b.WriteString(c.Data)
- }
- }
- return b.String()
-}
-
-func (s regexpPseudoClassSelector) Specificity() Specificity {
- return Specificity{0, 1, 0}
-}
-
-func (c regexpPseudoClassSelector) PseudoElement() string {
- return ""
-}
-
-type nthPseudoClassSelector struct {
- a, b int
- last, ofType bool
-}
-
-func (s nthPseudoClassSelector) Match(n *html.Node) bool {
- if s.a == 0 {
- if s.last {
- return simpleNthLastChildMatch(s.b, s.ofType, n)
- } else {
- return simpleNthChildMatch(s.b, s.ofType, n)
- }
- }
- return nthChildMatch(s.a, s.b, s.last, s.ofType, n)
-}
-
-// nthChildMatch implements :nth-child(an+b).
-// If last is true, implements :nth-last-child instead.
-// If ofType is true, implements :nth-of-type instead.
-func nthChildMatch(a, b int, last, ofType bool, n *html.Node) bool {
- if n.Type != html.ElementNode {
- return false
- }
-
- parent := n.Parent
- if parent == nil {
- return false
- }
-
- if parent.Type == html.DocumentNode {
- return false
- }
-
- i := -1
- count := 0
- for c := parent.FirstChild; c != nil; c = c.NextSibling {
- if (c.Type != html.ElementNode) || (ofType && c.Data != n.Data) {
- continue
- }
- count++
- if c == n {
- i = count
- if !last {
- break
- }
- }
- }
-
- if i == -1 {
- // This shouldn't happen, since n should always be one of its parent's children.
- return false
- }
-
- if last {
- i = count - i + 1
- }
-
- i -= b
- if a == 0 {
- return i == 0
- }
-
- return i%a == 0 && i/a >= 0
-}
-
-// simpleNthChildMatch implements :nth-child(b).
-// If ofType is true, implements :nth-of-type instead.
-func simpleNthChildMatch(b int, ofType bool, n *html.Node) bool {
- if n.Type != html.ElementNode {
- return false
- }
-
- parent := n.Parent
- if parent == nil {
- return false
- }
-
- if parent.Type == html.DocumentNode {
- return false
- }
-
- count := 0
- for c := parent.FirstChild; c != nil; c = c.NextSibling {
- if c.Type != html.ElementNode || (ofType && c.Data != n.Data) {
- continue
- }
- count++
- if c == n {
- return count == b
- }
- if count >= b {
- return false
- }
- }
- return false
-}
-
-// simpleNthLastChildMatch implements :nth-last-child(b).
-// If ofType is true, implements :nth-last-of-type instead.
-func simpleNthLastChildMatch(b int, ofType bool, n *html.Node) bool {
- if n.Type != html.ElementNode {
- return false
- }
-
- parent := n.Parent
- if parent == nil {
- return false
- }
-
- if parent.Type == html.DocumentNode {
- return false
- }
-
- count := 0
- for c := parent.LastChild; c != nil; c = c.PrevSibling {
- if c.Type != html.ElementNode || (ofType && c.Data != n.Data) {
- continue
- }
- count++
- if c == n {
- return count == b
- }
- if count >= b {
- return false
- }
- }
+func (s neverMatchSelector) Match(n *html.Node) bool {
return false
}
-// Specificity for nth-child pseudo-class.
-// Does not support a list of selectors
-func (s nthPseudoClassSelector) Specificity() Specificity {
- return Specificity{0, 1, 0}
-}
-
-func (c nthPseudoClassSelector) PseudoElement() string {
- return ""
-}
-
-type onlyChildPseudoClassSelector struct {
- ofType bool
-}
-
-// Match implements :only-child.
-// If `ofType` is true, it implements :only-of-type instead.
-func (s onlyChildPseudoClassSelector) Match(n *html.Node) bool {
- if n.Type != html.ElementNode {
- return false
- }
-
- parent := n.Parent
- if parent == nil {
- return false
- }
-
- if parent.Type == html.DocumentNode {
- return false
- }
-
- count := 0
- for c := parent.FirstChild; c != nil; c = c.NextSibling {
- if (c.Type != html.ElementNode) || (s.ofType && c.Data != n.Data) {
- continue
- }
- count++
- if count > 1 {
- return false
- }
- }
-
- return count == 1
-}
-
-func (s onlyChildPseudoClassSelector) Specificity() Specificity {
- return Specificity{0, 1, 0}
-}
-
-func (c onlyChildPseudoClassSelector) PseudoElement() string {
- return ""
-}
-
-type inputPseudoClassSelector struct{}
-
-// Matches input, select, textarea and button elements.
-func (s inputPseudoClassSelector) Match(n *html.Node) bool {
- return n.Type == html.ElementNode && (n.Data == "input" || n.Data == "select" || n.Data == "textarea" || n.Data == "button")
-}
-
-func (s inputPseudoClassSelector) Specificity() Specificity {
- return Specificity{0, 1, 0}
-}
-
-func (c inputPseudoClassSelector) PseudoElement() string {
- return ""
-}
-
-type emptyElementPseudoClassSelector struct{}
-
-// Matches empty elements.
-func (s emptyElementPseudoClassSelector) Match(n *html.Node) bool {
- if n.Type != html.ElementNode {
- return false
- }
-
- for c := n.FirstChild; c != nil; c = c.NextSibling {
- switch c.Type {
- case html.ElementNode, html.TextNode:
- return false
- }
- }
-
- return true
-}
-
-func (s emptyElementPseudoClassSelector) Specificity() Specificity {
- return Specificity{0, 1, 0}
-}
-
-func (c emptyElementPseudoClassSelector) PseudoElement() string {
- return ""
-}
-
-type rootPseudoClassSelector struct{}
-
-// Match implements :root
-func (s rootPseudoClassSelector) Match(n *html.Node) bool {
- if n.Type != html.ElementNode {
- return false
- }
- if n.Parent == nil {
- return false
- }
- return n.Parent.Type == html.DocumentNode
-}
-
-func (s rootPseudoClassSelector) Specificity() Specificity {
- return Specificity{0, 1, 0}
+func (s neverMatchSelector) Specificity() Specificity {
+ return Specificity{0, 0, 0}
}
-func (c rootPseudoClassSelector) PseudoElement() string {
+func (c neverMatchSelector) PseudoElement() string {
return ""
}
diff --git a/vendor/github.com/andybalholm/cascadia/serialize.go b/vendor/github.com/andybalholm/cascadia/serialize.go
index f15b07952775715c7dbbf8887449dfb7eaf563ac..61acf04e1c69470888485381ea95c1195d908f18 100644
--- a/vendor/github.com/andybalholm/cascadia/serialize.go
+++ b/vendor/github.com/andybalholm/cascadia/serialize.go
@@ -2,21 +2,35 @@ package cascadia
import (
"fmt"
+ "strconv"
"strings"
)
// implements the reverse operation Sel -> string
+var specialCharReplacer *strings.Replacer
+
+func init() {
+ var pairs []string
+ for _, s := range ",!\"#$%&'()*+ -./:;<=>?@[\\]^`{|}~" {
+ pairs = append(pairs, string(s), "\\"+string(s))
+ }
+ specialCharReplacer = strings.NewReplacer(pairs...)
+}
+
+// espace special CSS char
+func escape(s string) string { return specialCharReplacer.Replace(s) }
+
func (c tagSelector) String() string {
return c.tag
}
func (c idSelector) String() string {
- return "#" + c.id
+ return "#" + escape(c.id)
}
func (c classSelector) String() string {
- return "." + c.class
+ return "." + escape(c.class)
}
func (c attrSelector) String() string {
@@ -26,12 +40,20 @@ func (c attrSelector) String() string {
} else if c.operation != "" {
val = fmt.Sprintf(`"%s"`, val)
}
- return fmt.Sprintf(`[%s%s%s]`, c.key, c.operation, val)
+
+ ignoreCase := ""
+
+ if c.insensitive {
+ ignoreCase = " i"
+ }
+
+ return fmt.Sprintf(`[%s%s%s%s]`, c.key, c.operation, val, ignoreCase)
}
func (c relativePseudoClassSelector) String() string {
return fmt.Sprintf(":%s(%s)", c.name, c.match.String())
}
+
func (c containsPseudoClassSelector) String() string {
s := "contains"
if c.own {
@@ -39,6 +61,7 @@ func (c containsPseudoClassSelector) String() string {
}
return fmt.Sprintf(`:%s("%s")`, s, c.value)
}
+
func (c regexpPseudoClassSelector) String() string {
s := "matches"
if c.own {
@@ -46,6 +69,7 @@ func (c regexpPseudoClassSelector) String() string {
}
return fmt.Sprintf(":%s(%s)", s, c.regexp.String())
}
+
func (c nthPseudoClassSelector) String() string {
if c.a == 0 && c.b == 1 { // special cases
s := ":first-"
@@ -70,24 +94,56 @@ func (c nthPseudoClassSelector) String() string {
case [2]bool{false, false}:
name = "nth-child"
}
- return fmt.Sprintf(":%s(%dn+%d)", name, c.a, c.b)
+ s := fmt.Sprintf("+%d", c.b)
+ if c.b < 0 { // avoid +-8 invalid syntax
+ s = strconv.Itoa(c.b)
+ }
+ return fmt.Sprintf(":%s(%dn%s)", name, c.a, s)
}
+
func (c onlyChildPseudoClassSelector) String() string {
if c.ofType {
return ":only-of-type"
}
return ":only-child"
}
+
func (c inputPseudoClassSelector) String() string {
return ":input"
}
+
func (c emptyElementPseudoClassSelector) String() string {
return ":empty"
}
+
func (c rootPseudoClassSelector) String() string {
return ":root"
}
+func (c linkPseudoClassSelector) String() string {
+ return ":link"
+}
+
+func (c langPseudoClassSelector) String() string {
+ return fmt.Sprintf(":lang(%s)", c.lang)
+}
+
+func (c neverMatchSelector) String() string {
+ return c.value
+}
+
+func (c enabledPseudoClassSelector) String() string {
+ return ":enabled"
+}
+
+func (c disabledPseudoClassSelector) String() string {
+ return ":disabled"
+}
+
+func (c checkedPseudoClassSelector) String() string {
+ return ":checked"
+}
+
func (c compoundSelector) String() string {
if len(c.selectors) == 0 && c.pseudoElement == "" {
return "*"
diff --git a/vendor/modules.txt b/vendor/modules.txt
index 1d2f65d959c5e6ab9366434b068613fa89ec512d..2d68b824ba4a9f118df47f69713950efdbdd982a 100644
--- a/vendor/modules.txt
+++ b/vendor/modules.txt
@@ -1,4 +1,4 @@
-# github.com/PuerkitoBio/goquery v1.7.1
+# github.com/PuerkitoBio/goquery v1.8.0
## explicit
github.com/PuerkitoBio/goquery
# github.com/PuerkitoBio/purell v0.1.0
@@ -6,7 +6,7 @@ github.com/PuerkitoBio/goquery
github.com/PuerkitoBio/purell
# github.com/PuerkitoBio/urlesc v0.0.0-20170810143723-de5bf2ad4578
## explicit
-# github.com/andybalholm/cascadia v1.2.0
+# github.com/andybalholm/cascadia v1.3.1
github.com/andybalholm/cascadia
# github.com/golang/snappy v0.0.1
github.com/golang/snappy
@@ -39,6 +39,6 @@ github.com/syndtr/goleveldb/leveldb/table
github.com/syndtr/goleveldb/leveldb/util
# golang.org/x/crypto v0.0.0-20190308221718-c2843e01d9a2
## explicit
-# golang.org/x/net v0.0.0-20210614182718-04defd469f4e
+# golang.org/x/net v0.0.0-20210916014120-12bc252f5db8
golang.org/x/net/html
golang.org/x/net/html/atom