html: parse <caption> elements

Pass tests2.dat, test 33: <!DOCTYPE html><table><caption>test TEST</caption><td>test | <!DOCTYPE html> | <html> | <head> | <body> | <table> | <caption> | "test TEST" | <tbody> | <tr> | <td> | "test" R=nigeltao CC=golang-dev https://golang.org/cl/5371099

html: parse <caption> elements
28546ed5 · Andrew Balholm · Nigel Tao · 5e5c5c27 · 28546ed5 · 28546ed5
Commit 28546ed5 authored Nov 16, 2011 by Andrew Balholm Committed by Nigel Tao Nov 16, 2011
Hide whitespace changes
Inline Side-by-side

Showing with 48 additions and 2 deletions

parse.go src/pkg/html/parse.go +47 -1

parse_test.go src/pkg/html/parse_test.go +1 -1

No files found.
--- a/src/pkg/html/parse.go
+++ b/src/pkg/html/parse.go
@@ -298,7 +298,7 @@ func (p *parser) resetInsertionMode() {
 		case "tbody", "thead", "tfoot":
 			p.im = inTableBodyIM
 		case "caption":
-			// TODO: p.im = inCaptionIM
+			p.im = inCaptionIM
 		case "colgroup":
 			p.im = inColumnGroupIM
 		case "table":
@@ -887,6 +887,12 @@ func inTableIM(p *parser) bool {
 		// TODO.
 	case StartTagToken:
 		switch p.tok.Data {
+		case "caption":
+			p.clearStackToContext(tableScopeStopTags)
+			p.afe = append(p.afe, &scopeMarker)
+			p.addElement(p.tok.Data, p.tok.Attr)
+			p.im = inCaptionIM
+			return true
 		case "tbody", "tfoot", "thead":
 			p.clearStackToContext(tableScopeStopTags)
 			p.addElement(p.tok.Data, p.tok.Attr)
@@ -960,6 +966,46 @@ func (p *parser) clearStackToContext(stopTags []string) {
 	}
 }

+// Section 11.2.5.4.11.
+func inCaptionIM(p *parser) bool {
+	switch p.tok.Type {
+	case StartTagToken:
+		switch p.tok.Data {
+		case "caption", "col", "colgroup", "tbody", "td", "tfoot", "thead", "tr":
+			if p.popUntil(tableScopeStopTags, "caption") {
+				p.clearActiveFormattingElements()
+				p.im = inTableIM
+				return false
+			} else {
+				// Ignore the token.
+				return true
+			}
+		}
+	case EndTagToken:
+		switch p.tok.Data {
+		case "caption":
+			if p.popUntil(tableScopeStopTags, "caption") {
+				p.clearActiveFormattingElements()
+				p.im = inTableIM
+			}
+			return true
+		case "table":
+			if p.popUntil(tableScopeStopTags, "caption") {
+				p.clearActiveFormattingElements()
+				p.im = inTableIM
+				return false
+			} else {
+				// Ignore the token.
+				return true
+			}
+		case "body", "col", "colgroup", "html", "tbody", "td", "tfoot", "th", "thead", "tr":
+			// Ignore the token.
+			return true
+		}
+	}
+	return inBodyIM(p)
+}
+
 // Section 11.2.5.4.12.
 func inColumnGroupIM(p *parser) bool {
 	switch p.tok.Type {

--- a/src/pkg/html/parse_test.go
+++ b/src/pkg/html/parse_test.go
@@ -134,7 +134,7 @@ func TestParser(t *testing.T) {
 	}{
 		// TODO(nigeltao): Process all the test cases from all the .dat files.
 		{"tests1.dat", -1},
-		{"tests2.dat", 33},
+		{"tests2.dat", 34},
 		{"tests3.dat", 0},
 	}
 	for _, tf := range testFiles {