html: parse <select> tags.

The additional test case in parse_test.go is:
<select><b><option><select><option></b></select>X

R=andybalholm
CC=golang-dev
https://golang.org/cl/5293051

diff --git a/src/pkg/html/parse.go b/src/pkg/html/parse.go
index ba7e705a7946f5f50956d87090c26f3d38380a24..5b25e2620d00af9fa222fee244eeb20d04e67a13 100644 (file)
--- a/src/pkg/html/parse.go
+++ b/src/pkg/html/parse.go
@@ -234,6 +234,41 @@ func (p *parser) setOriginalIM(im insertionMode) {
        p.originalIM = im
 }
 
+// Section 11.2.3.1, "reset the insertion mode".
+func (p *parser) resetInsertionMode() insertionMode {
+       for i := len(p.oe) - 1; i >= 0; i-- {
+               n := p.oe[i]
+               if i == 0 {
+                       // TODO: set n to the context element, for HTML fragment parsing.
+               }
+               switch n.Data {
+               case "select":
+                       return inSelectIM
+               case "td", "th":
+                       return inCellIM
+               case "tr":
+                       return inRowIM
+               case "tbody", "thead", "tfoot":
+                       return inTableBodyIM
+               case "caption":
+                       // TODO: return inCaptionIM
+               case "colgroup":
+                       // TODO: return inColumnGroupIM
+               case "table":
+                       return inTableIM
+               case "head":
+                       return inBodyIM
+               case "body":
+                       return inBodyIM
+               case "frameset":
+                       // TODO: return inFramesetIM
+               case "html":
+                       return beforeHeadIM
+               }
+       }
+       return inBodyIM
+}
+
 // Section 11.2.5.4.1.
 func initialIM(p *parser) (insertionMode, bool) {
        switch p.tok.Type {
@@ -478,6 +513,12 @@ func inBodyIM(p *parser) (insertionMode, bool) {
                        p.oe.pop()
                        p.acknowledgeSelfClosingTag()
                        p.framesetOK = false
+               case "select":
+                       p.reconstructActiveFormattingElements()
+                       p.addElement(p.tok.Data, p.tok.Attr)
+                       p.framesetOK = false
+                       // TODO: detect <select> inside a table.
+                       return inSelectIM, true
                default:
                        // TODO.
                        p.addElement(p.tok.Data, p.tok.Attr)
@@ -671,8 +712,7 @@ func inTableIM(p *parser) (insertionMode, bool) {
                switch p.tok.Data {
                case "table":
                        if p.popUntil(tableScopeStopTags, "table") {
-                               // TODO: "reset the insertion mode appropriately" as per 11.2.3.1.
-                               return inBodyIM, false
+                               return p.resetInsertionMode(), true
                        }
                        // Ignore the token.
                        return inTableIM, true
@@ -833,6 +873,68 @@ func inCellIM(p *parser) (insertionMode, bool) {
        return useTheRulesFor(p, inCellIM, inBodyIM)
 }
 
+// Section 11.2.5.4.16.
+func inSelectIM(p *parser) (insertionMode, bool) {
+       endSelect := false
+       switch p.tok.Type {
+       case ErrorToken:
+               // TODO.
+       case TextToken:
+               p.addText(p.tok.Data)
+       case StartTagToken:
+               switch p.tok.Data {
+               case "html":
+                       // TODO.
+               case "option":
+                       if p.top().Data == "option" {
+                               p.oe.pop()
+                       }
+                       p.addElement(p.tok.Data, p.tok.Attr)
+               case "optgroup":
+                       // TODO.
+               case "select":
+                       endSelect = true
+               case "input", "keygen", "textarea":
+                       // TODO.
+               case "script":
+                       // TODO.
+               default:
+                       // Ignore the token.
+               }
+       case EndTagToken:
+               switch p.tok.Data {
+               case "option":
+                       // TODO.
+               case "optgroup":
+                       // TODO.
+               case "select":
+                       endSelect = true
+               default:
+                       // Ignore the token.
+               }
+       case CommentToken:
+               p.doc.Add(&Node{
+                       Type: CommentNode,
+                       Data: p.tok.Data,
+               })
+       }
+       if endSelect {
+               for i := len(p.oe) - 1; i >= 0; i-- {
+                       switch p.oe[i].Data {
+                       case "select":
+                               p.oe = p.oe[:i]
+                               return p.resetInsertionMode(), true
+                       case "option", "optgroup":
+                               continue
+                       default:
+                               // Ignore the token.
+                               return inSelectIM, true
+                       }
+               }
+       }
+       return inSelectIM, true
+}
+
 // Section 11.2.5.4.18.
 func afterBodyIM(p *parser) (insertionMode, bool) {
        switch p.tok.Type {

diff --git a/src/pkg/html/parse_test.go b/src/pkg/html/parse_test.go
index 18b70bf5100518072f7b049e18cf1e682a3ed283..2c56ffd6ad9b9d7a829c63a3db1f2fee64c1b63b 100644 (file)
--- a/src/pkg/html/parse_test.go
+++ b/src/pkg/html/parse_test.go
@@ -123,7 +123,7 @@ func TestParser(t *testing.T) {
                rc := make(chan io.Reader)
                go readDat(filename, rc)
                // TODO(nigeltao): Process all test cases, not just a subset.
-               for i := 0; i < 29; i++ {
+               for i := 0; i < 30; i++ {
                        // Parse the #data section.
                        b, err := ioutil.ReadAll(<-rc)
                        if err != nil {