exp/html: adjust inTableIM to match spec

Don't foster-parent text nodes that consist only of whitespace.
(I implemented this entirely in inTableIM instead of creating an
inTableTextIM, because the sole purpose of inTableTextIM seems to be
to combine character tokens into a string, which our tokenizer does
already.)

Use parseImpliedToken to clarify a couple of cases.

Handle <style>, <script>, <input>, and <form>.

Ignore doctype tokens.

Pass 20 additional tests.

R=nigeltao
CC=golang-dev
https://golang.org/cl/6117048

diff --git a/src/pkg/exp/html/parse.go b/src/pkg/exp/html/parse.go
index e3b3c95b591e212ae3b98ea844b9af3b880acb2e..08f029c63eb47ea16f168fb8619a1b137f79171a 100644 (file)
--- a/src/pkg/exp/html/parse.go
+++ b/src/pkg/exp/html/parse.go
@@ -1139,7 +1139,14 @@ func inTableIM(p *parser) bool {
                // Stop parsing.
                return true
        case TextToken:
-               // TODO.
+               p.tok.Data = strings.Replace(p.tok.Data, "\x00", "", -1)
+               switch p.oe.top().Data {
+               case "table", "tbody", "tfoot", "thead", "tr":
+                       if strings.Trim(p.tok.Data, whitespace) == "" {
+                               p.addText(p.tok.Data)
+                               return true
+                       }
+               }
        case StartTagToken:
                switch p.tok.Data {
                case "caption":
@@ -1148,15 +1155,21 @@ func inTableIM(p *parser) bool {
                        p.addElement(p.tok.Data, p.tok.Attr)
                        p.im = inCaptionIM
                        return true
+               case "colgroup":
+                       p.clearStackToContext(tableScope)
+                       p.addElement(p.tok.Data, p.tok.Attr)
+                       p.im = inColumnGroupIM
+                       return true
+               case "col":
+                       p.parseImpliedToken(StartTagToken, "colgroup", nil)
+                       return false
                case "tbody", "tfoot", "thead":
                        p.clearStackToContext(tableScope)
                        p.addElement(p.tok.Data, p.tok.Attr)
                        p.im = inTableBodyIM
                        return true
                case "td", "th", "tr":
-                       p.clearStackToContext(tableScope)
-                       p.addElement("tbody", nil)
-                       p.im = inTableBodyIM
+                       p.parseImpliedToken(StartTagToken, "tbody", nil)
                        return false
                case "table":
                        if p.popUntil(tableScope, "table") {
@@ -1165,16 +1178,24 @@ func inTableIM(p *parser) bool {
                        }
                        // Ignore the token.
                        return true
-               case "colgroup":
-                       p.clearStackToContext(tableScope)
+               case "style", "script":
+                       return inHeadIM(p)
+               case "input":
+                       for _, a := range p.tok.Attr {
+                               if a.Key == "type" && strings.ToLower(a.Val) == "hidden" {
+                                       p.addElement(p.tok.Data, p.tok.Attr)
+                                       p.oe.pop()
+                                       return true
+                               }
+                       }
+                       // Otherwise drop down to the default action.
+               case "form":
+                       if p.form != nil {
+                               // Ignore the token.
+                               return true
+                       }
                        p.addElement(p.tok.Data, p.tok.Attr)
-                       p.im = inColumnGroupIM
-                       return true
-               case "col":
-                       p.clearStackToContext(tableScope)
-                       p.addElement("colgroup", p.tok.Attr)
-                       p.im = inColumnGroupIM
-                       return false
+                       p.form = p.oe.pop()
                case "select":
                        p.reconstructActiveFormattingElements()
                        switch p.top().Data {
@@ -1186,8 +1207,6 @@ func inTableIM(p *parser) bool {
                        p.framesetOK = false
                        p.im = inSelectInTableIM
                        return true
-               default:
-                       // TODO.
                }
        case EndTagToken:
                switch p.tok.Data {
@@ -1208,6 +1227,9 @@ func inTableIM(p *parser) bool {
                        Data: p.tok.Data,
                })
                return true
+       case DoctypeToken:
+               // Ignore the token.
+               return true
        }
 
        switch p.top().Data {

diff --git a/src/pkg/exp/html/testlogs/html5test-com.dat.log b/src/pkg/exp/html/testlogs/html5test-com.dat.log
index 0742940268b575e528ee9b24861e9ee840af91af..5b921de88cdd2a94cce5b7ac4338cf43bf838607 100644 (file)
--- a/src/pkg/exp/html/testlogs/html5test-com.dat.log
+++ b/src/pkg/exp/html/testlogs/html5test-com.dat.log
@@ -17,7 +17,7 @@ PASS "<textarea><!--</textarea>-->"
 PASS "<style><!--</style>--></style>"
 PASS "<style><!--</style>-->"
 PASS "<ul><li>A </li> <li>B</li></ul>"
-FAIL "<table><form><input type=hidden><input></form><div></div></table>"
+PASS "<table><form><input type=hidden><input></form><div></div></table>"
 PASS "<i>A<b>B<p></i>C</b>D"
 PASS "<div></div>"
 PASS "<svg></svg>"

diff --git a/src/pkg/exp/html/testlogs/tables01.dat.log b/src/pkg/exp/html/testlogs/tables01.dat.log
index 16a08c6d92c08d82ffab7c0ad71f38472ae83d9c..a95223cff0e9889b3c5f295bed38d5b257ebbbbb 100644 (file)
--- a/src/pkg/exp/html/testlogs/tables01.dat.log
+++ b/src/pkg/exp/html/testlogs/tables01.dat.log
@@ -1,6 +1,6 @@
 PASS "<table><th>"
 PASS "<table><td>"
-FAIL "<table><col foo='bar'>"
+PASS "<table><col foo='bar'>"
 PASS "<table><colgroup></html>foo"
 PASS "<table></table><p>foo"
 PASS "<table></body></caption></col></colgroup></html></tbody></td></tfoot></th></thead></tr><td>"

diff --git a/src/pkg/exp/html/testlogs/tests15.dat.log b/src/pkg/exp/html/testlogs/tests15.dat.log
index d8ff9dd2050081697e48a64c828360c481e7421a..7129223196008fee5605d7fbddce27cbc23fef23 100644 (file)
--- a/src/pkg/exp/html/testlogs/tests15.dat.log
+++ b/src/pkg/exp/html/testlogs/tests15.dat.log
@@ -8,7 +8,7 @@ PASS "<!doctype html><table> X<meta></table>"
 PASS "<!doctype html><table> x</table>"
 PASS "<!doctype html><table> x </table>"
 PASS "<!doctype html><table><tr> x</table>"
-FAIL "<!doctype html><table>X<style> <tr>x </style> </table>"
-FAIL "<!doctype html><div><table><a>foo</a> <tr><td>bar</td> </tr></table></div>"
+PASS "<!doctype html><table>X<style> <tr>x </style> </table>"
+PASS "<!doctype html><div><table><a>foo</a> <tr><td>bar</td> </tr></table></div>"
 PASS "<frame></frame></frame><frameset><frame><frameset><frame></frameset><noframes></frameset><noframes>"
 PASS "<!DOCTYPE html><object></html>"

diff --git a/src/pkg/exp/html/testlogs/tests16.dat.log b/src/pkg/exp/html/testlogs/tests16.dat.log
index b07eaea5d923d064645635397ffdc83dc1cb9566..902c684841fed34d2cb3dc2d266ea1951e3b5c43 100644 (file)
--- a/src/pkg/exp/html/testlogs/tests16.dat.log
+++ b/src/pkg/exp/html/testlogs/tests16.dat.log
@@ -186,4 +186,4 @@ PASS "<xmp><!--<xmp></xmp>--></xmp>"
 PASS "<noembed><!--<noembed></noembed>--></noembed>"
 FAIL "<!doctype html><table>"
 PASS "<!doctype html><table><td><span><font></span><span>"
-FAIL "<!doctype html><form><table></form><form></table></form>"
+PARSE "<!doctype html><form><table></form><form></table></form>"

diff --git a/src/pkg/exp/html/testlogs/tests18.dat.log b/src/pkg/exp/html/testlogs/tests18.dat.log
index c288ead89a1534a20da1128c14586abeb8a9e174..8fdc01d5159ecd098aed4a24b1b7d755234bf109 100644 (file)
--- a/src/pkg/exp/html/testlogs/tests18.dat.log
+++ b/src/pkg/exp/html/testlogs/tests18.dat.log
@@ -5,8 +5,8 @@ PARSE "<!doctype html><table><tbody><tr><plaintext></plaintext>"
 PARSE "<!doctype html><table><tbody><tr><plaintext></plaintext>"
 PASS "<!doctype html><table><td><plaintext></plaintext>"
 PASS "<!doctype html><table><caption><plaintext></plaintext>"
-FAIL "<!doctype html><table><tr><style></script></style>abc"
-FAIL "<!doctype html><table><tr><script></style></script>abc"
+PASS "<!doctype html><table><tr><style></script></style>abc"
+PASS "<!doctype html><table><tr><script></style></script>abc"
 PASS "<!doctype html><table><caption><style></script></style>abc"
 PASS "<!doctype html><table><td><style></script></style>abc"
 FAIL "<!doctype html><select><script></style></script>abc"

diff --git a/src/pkg/exp/html/testlogs/tests19.dat.log b/src/pkg/exp/html/testlogs/tests19.dat.log
index bca9478848fb50a207b58796490fd2f813debf30..2b9dd1f44eb304107fa42653191bf62d9cfe6875 100644 (file)
--- a/src/pkg/exp/html/testlogs/tests19.dat.log
+++ b/src/pkg/exp/html/testlogs/tests19.dat.log
@@ -23,7 +23,7 @@ PASS "<!doctype html><h1><div><h3><span></h1>foo"
 PASS "<!doctype html><p></h3>foo"
 PASS "<!doctype html><h3><li>abc</h2>foo"
 PASS "<!doctype html><table>abc<!--foo-->"
-FAIL "<!doctype html><table>  <!--foo-->"
+PASS "<!doctype html><table>  <!--foo-->"
 PASS "<!doctype html><table> b <!--foo-->"
 PASS "<!doctype html><select><option><option>"
 PASS "<!doctype html><select><option></optgroup>"

diff --git a/src/pkg/exp/html/testlogs/tests20.dat.log b/src/pkg/exp/html/testlogs/tests20.dat.log
index abf5f57a4757bc2bbb7b9cf95096a5d274f109ae..4f4cb0360ed4bccc7187b04f643bbe1042510cb4 100644 (file)
--- a/src/pkg/exp/html/testlogs/tests20.dat.log
+++ b/src/pkg/exp/html/testlogs/tests20.dat.log
@@ -24,8 +24,8 @@ PASS "<!doctype html><svg>"
 PASS "<!doctype html><p><figcaption>"
 PASS "<!doctype html><p><summary>"
 PASS "<!doctype html><form><table><form>"
-FAIL "<!doctype html><table><form><form>"
-FAIL "<!doctype html><table><form></table><form>"
+PASS "<!doctype html><table><form><form>"
+PASS "<!doctype html><table><form></table><form>"
 PASS "<!doctype html><svg><foreignObject><p>"
 PASS "<!doctype html><svg><title>abc"
 PASS "<option><span><option>"

diff --git a/src/pkg/exp/html/testlogs/tests7.dat.log b/src/pkg/exp/html/testlogs/tests7.dat.log
index 726354141407d3b3208545dadccd706a09157a2f..c18848acb606bac79744cc651c33ad0431abfb21 100644 (file)
--- a/src/pkg/exp/html/testlogs/tests7.dat.log
+++ b/src/pkg/exp/html/testlogs/tests7.dat.log
@@ -3,20 +3,20 @@ PASS "<!doctype html><table><title>X</title></table>"
 FAIL "<!doctype html><head></head><title>X</title>"
 FAIL "<!doctype html></head><title>X</title>"
 PASS "<!doctype html><table><meta></table>"
-FAIL "<!doctype html><table>X<tr><td><table> <meta></table></table>"
+PASS "<!doctype html><table>X<tr><td><table> <meta></table></table>"
 PASS "<!doctype html><html> <head>"
 PASS "<!doctype html> <head>"
-FAIL "<!doctype html><table><style> <tr>x </style> </table>"
-FAIL "<!doctype html><table><TBODY><script> <tr>x </script> </table>"
+PASS "<!doctype html><table><style> <tr>x </style> </table>"
+PASS "<!doctype html><table><TBODY><script> <tr>x </script> </table>"
 PASS "<!doctype html><p><applet><p>X</p></applet>"
 PASS "<!doctype html><listing>\nX</listing>"
 FAIL "<!doctype html><select><input>X"
 PASS "<!doctype html><select><select>X"
-FAIL "<!doctype html><table><input type=hidDEN></table>"
-FAIL "<!doctype html><table>X<input type=hidDEN></table>"
-FAIL "<!doctype html><table>  <input type=hidDEN></table>"
-FAIL "<!doctype html><table>  <input type='hidDEN'></table>"
-FAIL "<!doctype html><table><input type=\" hidden\"><input type=hidDEN></table>"
+PASS "<!doctype html><table><input type=hidDEN></table>"
+PASS "<!doctype html><table>X<input type=hidDEN></table>"
+PASS "<!doctype html><table>  <input type=hidDEN></table>"
+PASS "<!doctype html><table>  <input type='hidDEN'></table>"
+PASS "<!doctype html><table><input type=\" hidden\"><input type=hidDEN></table>"
 PASS "<!doctype html><table><select>X<tr>"
 PASS "<!doctype html><select>X</select>"
 PASS "<!DOCTYPE hTmL><html></html>"

diff --git a/src/pkg/exp/html/testlogs/tricky01.dat.log b/src/pkg/exp/html/testlogs/tricky01.dat.log
index dc411e547a503f8eb8bf19e65fada4f7f928b163..fb0390f818171c87f43aa725e62501584fdd5d91 100644 (file)
--- a/src/pkg/exp/html/testlogs/tricky01.dat.log
+++ b/src/pkg/exp/html/testlogs/tricky01.dat.log
@@ -3,7 +3,7 @@ PASS "<html>\n<font color=red><i>Italic and Red<p>Italic and Red </font> Just it
 PASS "<html><body>\n<p><font size=\"7\">First paragraph.</p>\n<p>Second paragraph.</p></font>\n<b><p><i>Bold and Italic</b> Italic</p>"
 PASS "<html>\n<dl>\n<dt><b>Boo\n<dd>Goo?\n</dl>\n</html>"
 PASS "<html><body>\n<label><a><div>Hello<div>World</div></a></label>  \n</body></html>"
-FAIL "<table><center> <font>a</center> <img> <tr><td> </td> </tr> </table>"
+PASS "<table><center> <font>a</center> <img> <tr><td> </td> </tr> </table>"
 FAIL "<table><tr><p><a><p>You should see this text."
-FAIL "<TABLE>\n<TR>\n<CENTER><CENTER><TD></TD></TR><TR>\n<FONT>\n<TABLE><tr></tr></TABLE>\n</P>\n<a></font><font></a>\nThis page contains an insanely badly-nested tag sequence."
+PASS "<TABLE>\n<TR>\n<CENTER><CENTER><TD></TD></TR><TR>\n<FONT>\n<TABLE><tr></tr></TABLE>\n</P>\n<a></font><font></a>\nThis page contains an insanely badly-nested tag sequence."
 PASS "<html>\n<body>\n<b><nobr><div>This text is in a div inside a nobr</nobr>More text that should not be in the nobr, i.e., the\nnobr should have closed the div inside it implicitly. </b><pre>A pre tag outside everything else.</pre>\n</body>\n</html>"