html: Added ParseOptionIgnoreBOM option

This option treats the UTF-8 BOM, if present, as whitespace to prevent moving comments into the body element. This is mainly intended for use with RenderOptionAllowXMLDeclarations to prevent the XML declaration being moved into the body element (which is invalid). See pgaskin/kepubify#36 for an example of this.
pgaskin · Jan 12, 2020 · fd41993 · fd41993
1 parent 85aecc8
commit fd41993
Show file tree

Hide file tree

Showing 2 changed files with 34 additions and 0 deletions.
diff --git a/html/parse.go b/html/parse.go
@@ -55,6 +55,9 @@ type parser struct {
 	// to be self closing (<whatever ... />). This is mainly for better
 	// compatibility with XHTML found in EPUBs. MOD(geek1011)
 	lenientSelfClosing bool
+	// ignoreBOM skips the BOM at the beginning of the document, if present.
+	// MOD(geek1011)
+	ignoreBOM bool
 }
 
 func (p *parser) top() *Node {
@@ -521,6 +524,11 @@ const whitespace = " \t\r\n\f"
 func initialIM(p *parser) bool {
 	switch p.tok.Type {
 	case TextToken:
+		// MOD(geek1011): Ignore BOM when considering initial data before document.
+		if p.ignoreBOM {
+			p.tok.Data = strings.TrimPrefix(p.tok.Data, "\xEF\xBB\xBF")
+		}
+		// END MOD
 		p.tok.Data = strings.TrimLeft(p.tok.Data, whitespace)
 		if len(p.tok.Data) == 0 {
 			// It was all whitespace, so ignore it.
@@ -2410,6 +2418,17 @@ func ParseOptionLenientSelfClosing(enable bool) ParseOption {
 	}
 }
 
+// ParseOptionIgnoreBOM skips reading the UTF-8 BOM (EF BB BF), if present, at
+// the beginning of the document (technically, it makes the parser consider it
+// as whitespace). This option is mainly intended for use with with
+// RenderOptionAllowXMLDeclarations to prevent the XML declaration being moved
+// into the body element. MOD(geek1011)
+func ParseOptionIgnoreBOM(enable bool) ParseOption {
+	return func(p *parser) {
+		p.ignoreBOM = enable
+	}
+}
+
 // ParseWithOptions is like Parse, with options.
 func ParseWithOptions(r io.Reader, opts ...ParseOption) (*Node, error) {
 	p := &parser{

diff --git a/html/parse_geek1011_test.go b/html/parse_geek1011_test.go
@@ -189,3 +189,18 @@ func TestMod_ParseLenientSelfClosing(t *testing.T) {
 		RenderedB:   `<!DOCTYPE html><html><head><title>Title</title></head><body><p><i>Test &gt;&lt;<span id="test"></span>&gt;<b>&lt; 1<span>test</span></b></i></p><p><i><b>Test 2</b></i></p></body></html>`,
 	}.Test(t)
 }
+
+func TestMod_ParseIgnoreBOM(t *testing.T) {
+	testModCase{
+		What:     `BOM and comment`,
+		Original: "\xEF\xBB\xBF" + `<!-- Comment Text --><!DOCTYPE html><html><head><title>Title</title></head><body><p>Test 1</p></body></html>`,
+
+		ParseOptsA:  []ParseOption{ParseOptionIgnoreBOM(false)},
+		RenderOptsA: nil,
+		RenderedA:   `<html><head></head><body>` + "\xEF\xBB\xBF" + `<!-- Comment Text --><title>Title</title><p>Test 1</p></body></html>`,
+
+		ParseOptsB:  []ParseOption{ParseOptionIgnoreBOM(true)},
+		RenderOptsB: nil,
+		RenderedB:   `<!-- Comment Text --><!DOCTYPE html><html><head><title>Title</title></head><body><p>Test 1</p></body></html>`,
+	}.Test(t)
+}