diff options
author | ian <ian@138bc75d-0d04-0410-961f-82ee72b054a4> | 2012-01-25 20:56:26 +0000 |
---|---|---|
committer | ian <ian@138bc75d-0d04-0410-961f-82ee72b054a4> | 2012-01-25 20:56:26 +0000 |
commit | 422eaae5fe0038ad189b8fd28cfd6a7094d67ae1 (patch) | |
tree | c68d6b2a9f5b82a23171b0a488a4b7e5c63ad860 /libgo/go/html | |
parent | e0f3ea3ed4b9d0bce9f4c14762e4257ba62c8fba (diff) | |
download | gcc-422eaae5fe0038ad189b8fd28cfd6a7094d67ae1.tar.gz |
libgo: Update to weekly.2012-01-15.
git-svn-id: svn+ssh://gcc.gnu.org/svn/gcc/trunk@183539 138bc75d-0d04-0410-961f-82ee72b054a4
Diffstat (limited to 'libgo/go/html')
-rw-r--r-- | libgo/go/html/foreign.go | 78 | ||||
-rw-r--r-- | libgo/go/html/node.go | 4 | ||||
-rw-r--r-- | libgo/go/html/parse.go | 309 | ||||
-rw-r--r-- | libgo/go/html/parse_test.go | 19 | ||||
-rw-r--r-- | libgo/go/html/render.go | 8 | ||||
-rw-r--r-- | libgo/go/html/template/escape_test.go | 4 | ||||
-rw-r--r-- | libgo/go/html/token.go | 9 |
7 files changed, 316 insertions, 115 deletions
diff --git a/libgo/go/html/foreign.go b/libgo/go/html/foreign.go index 0f9b4ad560d..3ba81ce4d6f 100644 --- a/libgo/go/html/foreign.go +++ b/libgo/go/html/foreign.go @@ -4,6 +4,42 @@ package html +import ( + "strings" +) + +func adjustForeignAttributes(aa []Attribute) { + for i, a := range aa { + if a.Key == "" || a.Key[0] != 'x' { + continue + } + switch a.Key { + case "xlink:actuate", "xlink:arcrole", "xlink:href", "xlink:role", "xlink:show", + "xlink:title", "xlink:type", "xml:base", "xml:lang", "xml:space", "xmlns:xlink": + j := strings.Index(a.Key, ":") + aa[i].Namespace = a.Key[:j] + aa[i].Key = a.Key[j+1:] + } + } +} + +func htmlIntegrationPoint(n *Node) bool { + if n.Type != ElementNode { + return false + } + switch n.Namespace { + case "math": + // TODO: annotation-xml elements whose start tags have "text/html" or + // "application/xhtml+xml" encodings. + case "svg": + switch n.Data { + case "desc", "foreignObject", "title": + return true + } + } + return false +} + // Section 12.2.5.5. var breakout = map[string]bool{ "b": true, @@ -53,4 +89,44 @@ var breakout = map[string]bool{ "var": true, } -// TODO: add look-up tables for MathML and SVG adjustments. +// Section 12.2.5.5. +var svgTagNameAdjustments = map[string]string{ + "altglyph": "altGlyph", + "altglyphdef": "altGlyphDef", + "altglyphitem": "altGlyphItem", + "animatecolor": "animateColor", + "animatemotion": "animateMotion", + "animatetransform": "animateTransform", + "clippath": "clipPath", + "feblend": "feBlend", + "fecolormatrix": "feColorMatrix", + "fecomponenttransfer": "feComponentTransfer", + "fecomposite": "feComposite", + "feconvolvematrix": "feConvolveMatrix", + "fediffuselighting": "feDiffuseLighting", + "fedisplacementmap": "feDisplacementMap", + "fedistantlight": "feDistantLight", + "feflood": "feFlood", + "fefunca": "feFuncA", + "fefuncb": "feFuncB", + "fefuncg": "feFuncG", + "fefuncr": "feFuncR", + "fegaussianblur": "feGaussianBlur", + "feimage": "feImage", + "femerge": "feMerge", + "femergenode": "feMergeNode", + "femorphology": "feMorphology", + "feoffset": "feOffset", + "fepointlight": "fePointLight", + "fespecularlighting": "feSpecularLighting", + "fespotlight": "feSpotLight", + "fetile": "feTile", + "feturbulence": "feTurbulence", + "foreignobject": "foreignObject", + "glyphref": "glyphRef", + "lineargradient": "linearGradient", + "radialgradient": "radialGradient", + "textpath": "textPath", +} + +// TODO: add look-up tables for MathML and SVG attribute adjustments. diff --git a/libgo/go/html/node.go b/libgo/go/html/node.go index 4ba3f5fb627..83f17308b18 100644 --- a/libgo/go/html/node.go +++ b/libgo/go/html/node.go @@ -26,6 +26,10 @@ var scopeMarker = Node{Type: scopeMarkerNode} // content for text) and are part of a tree of Nodes. Element nodes may also // have a Namespace and contain a slice of Attributes. Data is unescaped, so // that it looks like "a<b" rather than "a<b". +// +// An empty Namespace implies a "http://www.w3.org/1999/xhtml" namespace. +// Similarly, "math" is short for "http://www.w3.org/1998/Math/MathML", and +// "svg" is short for "http://www.w3.org/2000/svg". type Node struct { Parent *Node Child []*Node diff --git a/libgo/go/html/parse.go b/libgo/go/html/parse.go index 6962e643932..43c04727ab8 100644 --- a/libgo/go/html/parse.go +++ b/libgo/go/html/parse.go @@ -51,58 +51,87 @@ func (p *parser) top() *Node { return p.doc } -// stopTags for use in popUntil. These come from section 12.2.3.2. +// Stop tags for use in popUntil. These come from section 12.2.3.2. var ( - defaultScopeStopTags = []string{"applet", "caption", "html", "table", "td", "th", "marquee", "object"} - listItemScopeStopTags = []string{"applet", "caption", "html", "table", "td", "th", "marquee", "object", "ol", "ul"} - buttonScopeStopTags = []string{"applet", "caption", "html", "table", "td", "th", "marquee", "object", "button"} - tableScopeStopTags = []string{"html", "table"} + defaultScopeStopTags = map[string][]string{ + "": {"applet", "caption", "html", "table", "td", "th", "marquee", "object"}, + "math": {"annotation-xml", "mi", "mn", "mo", "ms", "mtext"}, + "svg": {"desc", "foreignObject", "title"}, + } ) -// stopTags for use in clearStackToContext. -var ( - tableRowContextStopTags = []string{"tr", "html"} +type scope int + +const ( + defaultScope scope = iota + listItemScope + buttonScope + tableScope + tableRowScope ) // popUntil pops the stack of open elements at the highest element whose tag -// is in matchTags, provided there is no higher element in stopTags. It returns -// whether or not there was such an element. If there was not, popUntil leaves -// the stack unchanged. +// is in matchTags, provided there is no higher element in the scope's stop +// tags (as defined in section 12.2.3.2). It returns whether or not there was +// such an element. If there was not, popUntil leaves the stack unchanged. // -// For example, if the stack was: +// For example, the set of stop tags for table scope is: "html", "table". If +// the stack was: // ["html", "body", "font", "table", "b", "i", "u"] -// then popUntil([]string{"html, "table"}, "font") would return false, but -// popUntil([]string{"html, "table"}, "i") would return true and the resultant -// stack would be: +// then popUntil(tableScope, "font") would return false, but +// popUntil(tableScope, "i") would return true and the stack would become: // ["html", "body", "font", "table", "b"] // -// If an element's tag is in both stopTags and matchTags, then the stack will -// be popped and the function returns true (provided, of course, there was no -// higher element in the stack that was also in stopTags). For example, -// popUntil([]string{"html, "table"}, "table") would return true and leave: +// If an element's tag is in both the stop tags and matchTags, then the stack +// will be popped and the function returns true (provided, of course, there was +// no higher element in the stack that was also in the stop tags). For example, +// popUntil(tableScope, "table") returns true and leaves: // ["html", "body", "font"] -func (p *parser) popUntil(stopTags []string, matchTags ...string) bool { - if i := p.indexOfElementInScope(stopTags, matchTags...); i != -1 { +func (p *parser) popUntil(s scope, matchTags ...string) bool { + if i := p.indexOfElementInScope(s, matchTags...); i != -1 { p.oe = p.oe[:i] return true } return false } -// indexOfElementInScope returns the index in p.oe of the highest element -// whose tag is in matchTags that is in scope according to stopTags. -// If no matching element is in scope, it returns -1. -func (p *parser) indexOfElementInScope(stopTags []string, matchTags ...string) int { +// indexOfElementInScope returns the index in p.oe of the highest element whose +// tag is in matchTags that is in scope. If no matching element is in scope, it +// returns -1. +func (p *parser) indexOfElementInScope(s scope, matchTags ...string) int { for i := len(p.oe) - 1; i >= 0; i-- { tag := p.oe[i].Data - for _, t := range matchTags { - if t == tag { - return i + if p.oe[i].Namespace == "" { + for _, t := range matchTags { + if t == tag { + return i + } + } + switch s { + case defaultScope: + // No-op. + case listItemScope: + if tag == "ol" || tag == "ul" { + return -1 + } + case buttonScope: + if tag == "button" { + return -1 + } + case tableScope: + if tag == "html" || tag == "table" { + return -1 + } + default: + panic("unreachable") } } - for _, t := range stopTags { - if t == tag { - return -1 + switch s { + case defaultScope, listItemScope, buttonScope: + for _, t := range defaultScopeStopTags[p.oe[i].Namespace] { + if t == tag { + return -1 + } } } } @@ -111,8 +140,30 @@ func (p *parser) indexOfElementInScope(stopTags []string, matchTags ...string) i // elementInScope is like popUntil, except that it doesn't modify the stack of // open elements. -func (p *parser) elementInScope(stopTags []string, matchTags ...string) bool { - return p.indexOfElementInScope(stopTags, matchTags...) != -1 +func (p *parser) elementInScope(s scope, matchTags ...string) bool { + return p.indexOfElementInScope(s, matchTags...) != -1 +} + +// clearStackToContext pops elements off the stack of open elements until a +// scope-defined element is found. +func (p *parser) clearStackToContext(s scope) { + for i := len(p.oe) - 1; i >= 0; i-- { + tag := p.oe[i].Data + switch s { + case tableScope: + if tag == "html" || tag == "table" { + p.oe = p.oe[:i+1] + return + } + case tableRowScope: + if tag == "html" || tag == "tr" { + p.oe = p.oe[:i+1] + return + } + default: + panic("unreachable") + } + } } // addChild adds a child node n to the top element, and pushes n onto the stack @@ -192,10 +243,9 @@ func (p *parser) addText(text string) { // addElement calls addChild with an element node. func (p *parser) addElement(tag string, attr []Attribute) { p.addChild(&Node{ - Type: ElementNode, - Data: tag, - Namespace: p.top().Namespace, - Attr: attr, + Type: ElementNode, + Data: tag, + Attr: attr, }) } @@ -624,10 +674,10 @@ func inBodyIM(p *parser) bool { case "html": copyAttributes(p.oe[0], p.tok) case "address", "article", "aside", "blockquote", "center", "details", "dir", "div", "dl", "fieldset", "figcaption", "figure", "footer", "header", "hgroup", "menu", "nav", "ol", "p", "section", "summary", "ul": - p.popUntil(buttonScopeStopTags, "p") + p.popUntil(buttonScope, "p") p.addElement(p.tok.Data, p.tok.Attr) case "h1", "h2", "h3", "h4", "h5", "h6": - p.popUntil(buttonScopeStopTags, "p") + p.popUntil(buttonScope, "p") switch n := p.top(); n.Data { case "h1", "h2", "h3", "h4", "h5", "h6": p.oe.pop() @@ -649,7 +699,7 @@ func inBodyIM(p *parser) bool { p.addFormattingElement(p.tok.Data, p.tok.Attr) case "nobr": p.reconstructActiveFormattingElements() - if p.elementInScope(defaultScopeStopTags, "nobr") { + if p.elementInScope(defaultScope, "nobr") { p.inBodyEndTagFormatting("nobr") p.reconstructActiveFormattingElements() } @@ -667,14 +717,14 @@ func inBodyIM(p *parser) bool { p.framesetOK = false case "table": if !p.quirks { - p.popUntil(buttonScopeStopTags, "p") + p.popUntil(buttonScope, "p") } p.addElement(p.tok.Data, p.tok.Attr) p.framesetOK = false p.im = inTableIM return true case "hr": - p.popUntil(buttonScopeStopTags, "p") + p.popUntil(buttonScope, "p") p.addElement(p.tok.Data, p.tok.Attr) p.oe.pop() p.acknowledgeSelfClosingTag() @@ -683,12 +733,11 @@ func inBodyIM(p *parser) bool { p.reconstructActiveFormattingElements() p.addElement(p.tok.Data, p.tok.Attr) p.framesetOK = false - // TODO: detect <select> inside a table. p.im = inSelectIM return true case "form": if p.form == nil { - p.popUntil(buttonScopeStopTags, "p") + p.popUntil(buttonScope, "p") p.addElement(p.tok.Data, p.tok.Attr) p.form = p.top() } @@ -698,7 +747,7 @@ func inBodyIM(p *parser) bool { node := p.oe[i] switch node.Data { case "li": - p.popUntil(listItemScopeStopTags, "li") + p.popUntil(listItemScope, "li") case "address", "div", "p": continue default: @@ -708,7 +757,7 @@ func inBodyIM(p *parser) bool { } break } - p.popUntil(buttonScopeStopTags, "p") + p.popUntil(buttonScope, "p") p.addElement(p.tok.Data, p.tok.Attr) case "dd", "dt": p.framesetOK = false @@ -726,13 +775,13 @@ func inBodyIM(p *parser) bool { } break } - p.popUntil(buttonScopeStopTags, "p") + p.popUntil(buttonScope, "p") p.addElement(p.tok.Data, p.tok.Attr) case "plaintext": - p.popUntil(buttonScopeStopTags, "p") + p.popUntil(buttonScope, "p") p.addElement(p.tok.Data, p.tok.Attr) case "button": - p.popUntil(defaultScopeStopTags, "button") + p.popUntil(defaultScope, "button") p.reconstructActiveFormattingElements() p.addElement(p.tok.Data, p.tok.Attr) p.framesetOK = false @@ -750,6 +799,19 @@ func inBodyIM(p *parser) bool { copyAttributes(body, p.tok) } } + case "frameset": + if !p.framesetOK || len(p.oe) < 2 || p.oe[1].Data != "body" { + // Ignore the token. + return true + } + body := p.oe[1] + if body.Parent != nil { + body.Parent.Remove(body) + } + p.oe = p.oe[:1] + p.addElement(p.tok.Data, p.tok.Attr) + p.im = inFramesetIM + return true case "base", "basefont", "bgsound", "command", "link", "meta", "noframes", "script", "style", "title": return inHeadIM(p) case "image": @@ -776,7 +838,7 @@ func inBodyIM(p *parser) bool { } } p.acknowledgeSelfClosingTag() - p.popUntil(buttonScopeStopTags, "p") + p.popUntil(buttonScope, "p") p.addElement("form", nil) p.form = p.top() if action != "" { @@ -794,23 +856,20 @@ func inBodyIM(p *parser) bool { p.oe.pop() p.form = nil case "xmp": - p.popUntil(buttonScopeStopTags, "p") + p.popUntil(buttonScope, "p") p.reconstructActiveFormattingElements() p.framesetOK = false p.addElement(p.tok.Data, p.tok.Attr) case "math", "svg": p.reconstructActiveFormattingElements() - namespace := "" if p.tok.Data == "math" { // TODO: adjust MathML attributes. - namespace = "mathml" } else { // TODO: adjust SVG attributes. - namespace = "svg" } - // TODO: adjust foreign attributes. + adjustForeignAttributes(p.tok.Attr) p.addElement(p.tok.Data, p.tok.Attr) - p.top().Namespace = namespace + p.top().Namespace = p.tok.Data return true case "caption", "col", "colgroup", "frame", "head", "tbody", "td", "tfoot", "th", "thead", "tr": // Ignore the token. @@ -825,16 +884,16 @@ func inBodyIM(p *parser) bool { p.im = afterBodyIM return true case "p": - if !p.elementInScope(buttonScopeStopTags, "p") { + if !p.elementInScope(buttonScope, "p") { p.addElement("p", nil) } - p.popUntil(buttonScopeStopTags, "p") + p.popUntil(buttonScope, "p") case "a", "b", "big", "code", "em", "font", "i", "nobr", "s", "small", "strike", "strong", "tt", "u": p.inBodyEndTagFormatting(p.tok.Data) case "address", "article", "aside", "blockquote", "button", "center", "details", "dir", "div", "dl", "fieldset", "figcaption", "figure", "footer", "header", "hgroup", "listing", "menu", "nav", "ol", "pre", "section", "summary", "ul": - p.popUntil(defaultScopeStopTags, p.tok.Data) + p.popUntil(defaultScope, p.tok.Data) case "applet", "marquee", "object": - if p.popUntil(defaultScopeStopTags, p.tok.Data) { + if p.popUntil(defaultScope, p.tok.Data) { p.clearActiveFormattingElements() } case "br": @@ -883,7 +942,7 @@ func (p *parser) inBodyEndTagFormatting(tag string) { p.afe.remove(formattingElement) return } - if !p.elementInScope(defaultScopeStopTags, tag) { + if !p.elementInScope(defaultScope, tag) { // Ignore the tag. return } @@ -1017,45 +1076,56 @@ func inTableIM(p *parser) bool { case StartTagToken: switch p.tok.Data { case "caption": - p.clearStackToContext(tableScopeStopTags) + p.clearStackToContext(tableScope) 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.clearStackToContext(tableScope) p.addElement(p.tok.Data, p.tok.Attr) p.im = inTableBodyIM return true case "td", "th", "tr": - p.clearStackToContext(tableScopeStopTags) + p.clearStackToContext(tableScope) p.addElement("tbody", nil) p.im = inTableBodyIM return false case "table": - if p.popUntil(tableScopeStopTags, "table") { + if p.popUntil(tableScope, "table") { p.resetInsertionMode() return false } // Ignore the token. return true case "colgroup": - p.clearStackToContext(tableScopeStopTags) + p.clearStackToContext(tableScope) p.addElement(p.tok.Data, p.tok.Attr) p.im = inColumnGroupIM return true case "col": - p.clearStackToContext(tableScopeStopTags) + p.clearStackToContext(tableScope) p.addElement("colgroup", p.tok.Attr) p.im = inColumnGroupIM return false + case "select": + p.reconstructActiveFormattingElements() + switch p.top().Data { + case "table", "tbody", "tfoot", "thead", "tr": + p.fosterParenting = true + } + p.addElement(p.tok.Data, p.tok.Attr) + p.fosterParenting = false + p.framesetOK = false + p.im = inSelectInTableIM + return true default: // TODO. } case EndTagToken: switch p.tok.Data { case "table": - if p.popUntil(tableScopeStopTags, "table") { + if p.popUntil(tableScope, "table") { p.resetInsertionMode() return true } @@ -1082,26 +1152,13 @@ func inTableIM(p *parser) bool { return inBodyIM(p) } -// clearStackToContext pops elements off the stack of open elements -// until an element listed in stopTags is found. -func (p *parser) clearStackToContext(stopTags []string) { - for i := len(p.oe) - 1; i >= 0; i-- { - for _, tag := range stopTags { - if p.oe[i].Data == tag { - p.oe = p.oe[:i+1] - return - } - } - } -} - // Section 12.2.5.4.11. 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") { + if p.popUntil(tableScope, "caption") { p.clearActiveFormattingElements() p.im = inTableIM return false @@ -1109,17 +1166,23 @@ func inCaptionIM(p *parser) bool { // Ignore the token. return true } + case "select": + p.reconstructActiveFormattingElements() + p.addElement(p.tok.Data, p.tok.Attr) + p.framesetOK = false + p.im = inSelectInTableIM + return true } case EndTagToken: switch p.tok.Data { case "caption": - if p.popUntil(tableScopeStopTags, "caption") { + if p.popUntil(tableScope, "caption") { p.clearActiveFormattingElements() p.im = inTableIM } return true case "table": - if p.popUntil(tableScopeStopTags, "caption") { + if p.popUntil(tableScope, "caption") { p.clearActiveFormattingElements() p.im = inTableIM return false @@ -1203,7 +1266,7 @@ func inTableBodyIM(p *parser) bool { data = "tr" consumed = false case "caption", "col", "colgroup", "tbody", "tfoot", "thead": - if !p.popUntil(tableScopeStopTags, "tbody", "thead", "tfoot") { + if !p.popUntil(tableScope, "tbody", "thead", "tfoot") { // Ignore the token. return true } @@ -1215,7 +1278,7 @@ func inTableBodyIM(p *parser) bool { case EndTagToken: switch p.tok.Data { case "table": - if p.popUntil(tableScopeStopTags, "tbody", "thead", "tfoot") { + if p.popUntil(tableScope, "tbody", "thead", "tfoot") { p.im = inTableIM return false } @@ -1251,13 +1314,13 @@ func inRowIM(p *parser) bool { case StartTagToken: switch p.tok.Data { case "td", "th": - p.clearStackToContext(tableRowContextStopTags) + p.clearStackToContext(tableRowScope) p.addElement(p.tok.Data, p.tok.Attr) p.afe = append(p.afe, &scopeMarker) p.im = inCellIM return true case "caption", "col", "colgroup", "tbody", "tfoot", "thead", "tr": - if p.popUntil(tableScopeStopTags, "tr") { + if p.popUntil(tableScope, "tr") { p.im = inTableBodyIM return false } @@ -1269,14 +1332,14 @@ func inRowIM(p *parser) bool { case EndTagToken: switch p.tok.Data { case "tr": - if p.popUntil(tableScopeStopTags, "tr") { + if p.popUntil(tableScope, "tr") { p.im = inTableBodyIM return true } // Ignore the token. return true case "table": - if p.popUntil(tableScopeStopTags, "tr") { + if p.popUntil(tableScope, "tr") { p.im = inTableBodyIM return false } @@ -1311,11 +1374,17 @@ func inCellIM(p *parser) bool { case "caption", "col", "colgroup", "tbody", "td", "tfoot", "th", "thead", "tr": // TODO: check for "td" or "th" in table scope. closeTheCellAndReprocess = true + case "select": + p.reconstructActiveFormattingElements() + p.addElement(p.tok.Data, p.tok.Attr) + p.framesetOK = false + p.im = inSelectInTableIM + return true } case EndTagToken: switch p.tok.Data { case "td", "th": - if !p.popUntil(tableScopeStopTags, p.tok.Data) { + if !p.popUntil(tableScope, p.tok.Data) { // Ignore the token. return true } @@ -1336,7 +1405,7 @@ func inCellIM(p *parser) bool { return true } if closeTheCellAndReprocess { - if p.popUntil(tableScopeStopTags, "td") || p.popUntil(tableScopeStopTags, "th") { + if p.popUntil(tableScope, "td") || p.popUntil(tableScope, "th") { p.clearActiveFormattingElements() p.im = inRowIM return false @@ -1405,21 +1474,40 @@ func inSelectIM(p *parser) bool { }) } if endSelect { - for i := len(p.oe) - 1; i >= 0; i-- { - switch p.oe[i].Data { - case "select": - p.oe = p.oe[:i] - p.resetInsertionMode() - return true - case "option", "optgroup": - continue - default: + p.endSelect() + } + return true +} + +// Section 12.2.5.4.17. +func inSelectInTableIM(p *parser) bool { + switch p.tok.Type { + case StartTagToken, EndTagToken: + switch p.tok.Data { + case "caption", "table", "tbody", "tfoot", "thead", "tr", "td", "th": + if p.tok.Type == StartTagToken || p.elementInScope(tableScope, p.tok.Data) { + p.endSelect() + return false + } else { // Ignore the token. return true } } } - return true + return inSelectIM(p) +} + +func (p *parser) endSelect() { + for i := len(p.oe) - 1; i >= 0; i-- { + switch p.oe[i].Data { + case "option", "optgroup": + continue + case "select": + p.oe = p.oe[:i] + p.resetInsertionMode() + } + return + } } // Section 12.2.5.4.18. @@ -1618,6 +1706,11 @@ func parseForeignContent(p *parser) bool { Data: p.tok.Data, }) case StartTagToken: + if htmlIntegrationPoint(p.top()) { + inBodyIM(p) + p.resetInsertionMode() + return true + } if breakout[p.tok.Data] { for i := len(p.oe) - 1; i >= 0; i-- { // TODO: HTML, MathML integration points. @@ -1629,16 +1722,22 @@ func parseForeignContent(p *parser) bool { return false } switch p.top().Namespace { - case "mathml": + case "math": // TODO: adjust MathML attributes. case "svg": - // TODO: adjust SVG tag names. + // Adjust SVG tag names. The tokenizer lower-cases tag names, but + // SVG wants e.g. "foreignObject" with a capital second "O". + if x := svgTagNameAdjustments[p.tok.Data]; x != "" { + p.tok.Data = x + } // TODO: adjust SVG attributes. default: panic("html: bad parser state: unexpected namespace") } - // TODO: adjust foreign attributes. + adjustForeignAttributes(p.tok.Attr) + namespace := p.top().Namespace p.addElement(p.tok.Data, p.tok.Attr) + p.top().Namespace = namespace case EndTagToken: for i := len(p.oe) - 1; i >= 0; i-- { if p.oe[i].Namespace == "" { diff --git a/libgo/go/html/parse_test.go b/libgo/go/html/parse_test.go index 015b5838f0b..c929c257727 100644 --- a/libgo/go/html/parse_test.go +++ b/libgo/go/html/parse_test.go @@ -103,10 +103,21 @@ func dumpLevel(w io.Writer, n *Node, level int) error { } else { fmt.Fprintf(w, "<%s>", n.Data) } - for _, a := range n.Attr { + attr := n.Attr + if len(attr) == 2 && attr[0].Namespace == "xml" && attr[1].Namespace == "xlink" { + // Some of the test cases in tests10.dat change the order of adjusted + // foreign attributes, but that behavior is not in the spec, and could + // simply be an implementation detail of html5lib's python map ordering. + attr[0], attr[1] = attr[1], attr[0] + } + for _, a := range attr { io.WriteString(w, "\n") dumpIndent(w, level+1) - fmt.Fprintf(w, `%s="%s"`, a.Key, a.Val) + if a.Namespace != "" { + fmt.Fprintf(w, `%s %s="%s"`, a.Namespace, a.Key, a.Val) + } else { + fmt.Fprintf(w, `%s="%s"`, a.Key, a.Val) + } } case TextNode: fmt.Fprintf(w, `"%s"`, n.Data) @@ -172,8 +183,8 @@ func TestParser(t *testing.T) { {"tests3.dat", -1}, {"tests4.dat", -1}, {"tests5.dat", -1}, - {"tests6.dat", 47}, - {"tests10.dat", 16}, + {"tests6.dat", -1}, + {"tests10.dat", 33}, } for _, tf := range testFiles { f, err := os.Open("testdata/webkit/" + tf.filename) diff --git a/libgo/go/html/render.go b/libgo/go/html/render.go index 20751938d9d..07859faa7dd 100644 --- a/libgo/go/html/render.go +++ b/libgo/go/html/render.go @@ -149,6 +149,14 @@ func render1(w writer, n *Node) error { if err := w.WriteByte(' '); err != nil { return err } + if a.Namespace != "" { + if _, err := w.WriteString(a.Namespace); err != nil { + return err + } + if err := w.WriteByte(':'); err != nil { + return err + } + } if _, err := w.WriteString(a.Key); err != nil { return err } diff --git a/libgo/go/html/template/escape_test.go b/libgo/go/html/template/escape_test.go index a57f9826b5b..2ce1fb566a5 100644 --- a/libgo/go/html/template/escape_test.go +++ b/libgo/go/html/template/escape_test.go @@ -302,7 +302,7 @@ func TestEscape(t *testing.T) { }, { "styleObfuscatedExpressionBlocked", - `<p style="width: {{" e\78preS\0Sio/**/n(alert(1337))"}}">`, + `<p style="width: {{" e\\78preS\x00Sio/**/n(alert(1337))"}}">`, `<p style="width: ZgotmplZ">`, }, { @@ -312,7 +312,7 @@ func TestEscape(t *testing.T) { }, { "styleObfuscatedMozBindingBlocked", - `<p style="{{" -mo\7a-B\0I/**/nding(alert(1337))"}}: ...">`, + `<p style="{{" -mo\\7a-B\x00I/**/nding(alert(1337))"}}: ...">`, `<p style="ZgotmplZ: ...">`, }, { diff --git a/libgo/go/html/token.go b/libgo/go/html/token.go index 69af96840c2..5a385a1b5c5 100644 --- a/libgo/go/html/token.go +++ b/libgo/go/html/token.go @@ -52,11 +52,14 @@ func (t TokenType) String() string { return "Invalid(" + strconv.Itoa(int(t)) + ")" } -// An Attribute is an attribute key-value pair. Key is alphabetic (and hence +// An Attribute is an attribute namespace-key-value triple. Namespace is +// non-empty for foreign attributes like xlink, Key is alphabetic (and hence // does not contain escapable characters like '&', '<' or '>'), and Val is // unescaped (it looks like "a<b" rather than "a<b"). +// +// Namespace is only used by the parser, not the tokenizer. type Attribute struct { - Key, Val string + Namespace, Key, Val string } // A Token consists of a TokenType and some Data (tag name for start and end @@ -756,7 +759,7 @@ func (z *Tokenizer) Token() Token { for moreAttr { var key, val []byte key, val, moreAttr = z.TagAttr() - attr = append(attr, Attribute{string(key), string(val)}) + attr = append(attr, Attribute{"", string(key), string(val)}) } t.Data = string(name) t.Attr = attr |