HTML: support Go templates, fixes #35

go.mod

@@ -8,7 +8,7 @@ require (
 	github.com/fsnotify/fsnotify v1.7.0
 	github.com/matryer/try v0.0.0-20161228173917-9ac251b645a2
 	github.com/tdewolff/argp v0.0.0-20231030173501-fa6c54897951
-	github.com/tdewolff/parse/v2 v2.7.1
+	github.com/tdewolff/parse/v2 v2.7.2
 	github.com/tdewolff/test v1.0.10
 )
 

go.sum

@@ -10,8 +10,8 @@ github.com/matryer/try v0.0.0-20161228173917-9ac251b645a2 h1:JAEbJn3j/FrhdWA9jW8
 github.com/matryer/try v0.0.0-20161228173917-9ac251b645a2/go.mod h1:0KeJpeMD6o+O4hW7qJOT7vyQPKrWmj26uf5wMc/IiIs=
 github.com/tdewolff/argp v0.0.0-20231030173501-fa6c54897951 h1:QiakK8TMHgZfKZ2enwnyYzT3fOtzIwZ6mDkVC3scOrQ=
 github.com/tdewolff/argp v0.0.0-20231030173501-fa6c54897951/go.mod h1:fF+gnKbmf3iMG+ErLiF+orMU/InyZIEnKVVigUjfriw=
-github.com/tdewolff/parse/v2 v2.7.1 h1:gdImkv0sIupYr/cXAu5s+CxfVpxMdYZX2Qr+5Q+RdF8=
-github.com/tdewolff/parse/v2 v2.7.1/go.mod h1:9p2qMIHpjRSTr1qnFxQr+igogyTUTlwvf9awHSm84h8=
+github.com/tdewolff/parse/v2 v2.7.2 h1:9NdxF0nk/+lPI0YADDonSlpiY15hGcVUhXRj9hnK8sM=
+github.com/tdewolff/parse/v2 v2.7.2/go.mod h1:9p2qMIHpjRSTr1qnFxQr+igogyTUTlwvf9awHSm84h8=
 github.com/tdewolff/test v1.0.6/go.mod h1:6DAvZliBAAnD7rhVgwaM7DE5/d9NMOAJ09SqYqeK4QE=
 github.com/tdewolff/test v1.0.10 h1:uWiheaLgLcNFqHcdWveum7PQfMnIUTf9Kl3bFxrIoew=
 github.com/tdewolff/test v1.0.10/go.mod h1:6DAvZliBAAnD7rhVgwaM7DE5/d9NMOAJ09SqYqeK4QE=

html/buffer.go

@@ -8,12 +8,13 @@ import (
 // Token is a single token unit with an attribute value (if given) and hash of the data.
 type Token struct {
 	html.TokenType
-	Hash    Hash
-	Data    []byte
-	Text    []byte
-	AttrVal []byte
-	Traits  traits
-	Offset  int
+	Hash        Hash
+	Data        []byte
+	Text        []byte
+	AttrVal     []byte
+	Traits      traits
+	Offset      int
+	HasTemplate bool
 }
 
 // TokenBuffer is a buffer that allows for token look-ahead.
@@ -40,10 +41,11 @@ func (z *TokenBuffer) read(t *Token) {
 	t.Offset = z.r.Offset()
 	t.TokenType, t.Data = z.l.Next()
 	t.Text = z.l.Text()
+	t.HasTemplate = z.l.HasTemplate()
 	if t.TokenType == html.AttributeToken {
 		t.Offset += 1 + len(t.Text) + 1
 		t.AttrVal = z.l.AttrVal()
-		if len(t.AttrVal) > 1 && (t.AttrVal[0] == '"' || t.AttrVal[0] == '\'') {
+		if 1 < len(t.AttrVal) && (t.AttrVal[0] == '"' || t.AttrVal[0] == '\'') {
 			t.Offset++
 			t.AttrVal = t.AttrVal[1 : len(t.AttrVal)-1] // quotes will be readded in attribute loop if necessary
 		}

html/html.go

@@ -41,6 +41,13 @@ var (
 
 ////////////////////////////////////////////////////////////////
 
+var GoTemplateDelims = [2]string{"{{", "}}"}
+var HandlebarsTemplateDelims = [2]string{"{{", "}}"}
+var MustacheTemplateDelims = [2]string{"{{", "}}"}
+var EJSTemplateDelims = [2]string{"<%", "%>"}
+var ASPTemplateDelims = [2]string{"<%", "%>"}
+var PHPTemplateDelims = [2]string{"<?", "?>"}
+
 // Minifier is an HTML minifier.
 type Minifier struct {
 	KeepComments            bool
@@ -50,6 +57,7 @@ type Minifier struct {
 	KeepEndTags             bool
 	KeepQuotes              bool
 	KeepWhitespace          bool
+	TemplateDelims          [2]string
 }
 
 // Minify minifies HTML data, it reads from r and writes to w.
@@ -71,7 +79,7 @@ func (o *Minifier) Minify(m *minify.M, w io.Writer, r io.Reader, _ map[string]st
 	z := parse.NewInput(r)
 	defer z.Restore()
 
-	l := html.NewLexer(z)
+	l := html.NewTemplateLexer(z, o.TemplateDelims)
 	tb := NewTokenBuffer(z, l)
 	for {
 		t := *tb.Shift()
@@ -127,7 +135,9 @@ func (o *Minifier) Minify(m *minify.M, w io.Writer, r io.Reader, _ map[string]st
 			}
 		case html.TextToken:
 			// CSS and JS minifiers for inline code
-			if rawTagHash != 0 {
+			if t.HasTemplate {
+				w.Write(t.Data)
+			} else if rawTagHash != 0 {
 				if rawTagHash == Style || rawTagHash == Script || rawTagHash == Iframe {
 					var mimetype []byte
 					var params map[string]string
@@ -372,6 +382,9 @@ func (o *Minifier) Minify(m *minify.M, w io.Writer, r io.Reader, _ map[string]st
 						break
 					} else if attr.Text == nil {
 						continue // removed attribute
+					} else if attr.HasTemplate {
+						w.Write(attr.Data)
+						continue // don't minify attributes that contain templates
 					}
 
 					val := attr.AttrVal

html/html_test.go

@@ -26,6 +26,7 @@ func TestHTML(t *testing.T) {
 		expected string
 	}{
 		{`html`, `html`},
+		//{"<title>title</title> <body>", `<title>title</title>`},
 		{`<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML+RDFa 1.0//EN" "http://www.w3.org/MarkUp/DTD/xhtml-rdfa-1.dtd">`, `<!doctype html>`},
 		{`<!-- comment -->`, ``},
 		{`<!--# SSI Tag -->`, `<!--# SSI Tag -->`},
@@ -67,6 +68,7 @@ func TestHTML(t *testing.T) {
 		{`<input type="text" value="">`, `<input>`},
 		{`<a rel="noopener">`, `<a rel=noopener>`},
 		{`<a rel=" noopener  external ">`, `<a rel="noopener external">`},
+		{`<input accept="image/png, image/jpeg">`, `<input accept=image/png,image/jpeg>`},
 
 		// increase coverage
 		{`<script style="css">js</script>`, `<script style=css>js</script>`},
@@ -352,6 +354,30 @@ func TestHTMLURL(t *testing.T) {
 	}
 }
 
+func TestHTMLTemplates(t *testing.T) {
+	htmlTests := []struct {
+		html     string
+		expected string
+	}{
+		{`a<p>  {{ printf "  !  " }}  </p>b`, `a<p>  {{ printf "  !  " }}  </p>b`},
+		{`a<span>  {{ printf "  !  " }}  </span>b`, `a<span>  {{ printf "  !  " }}  </span>b`},
+		{`<a href={{ .Link }} />`, `<a href={{ .Link }}>`},
+		{`<input type="file" accept="{{ .Accept }}, image/jpeg">`, `<input type=file accept="{{ .Accept }}, image/jpeg">`},
+		{`<option value="0" {{ if eq .Type 0 }}selected{{ end }}>Foo</option>`, `<option value=0 {{ if eq .Type 0 }}selected{{ end }}>Foo</option>`},
+	}
+
+	m := minify.New()
+	htmlMinifier := &Minifier{KeepEndTags: true, TemplateDelims: GoTemplateDelims}
+	for _, tt := range htmlTests {
+		t.Run(tt.html, func(t *testing.T) {
+			r := bytes.NewBufferString(tt.html)
+			w := &bytes.Buffer{}
+			err := htmlMinifier.Minify(m, w, r, nil)
+			test.Minify(t, tt.html, err, w.String(), tt.expected)
+		})
+	}
+}
+
 func TestSpecialTagClosing(t *testing.T) {
 	m := minify.New()
 	m.AddFunc("text/html", Minify)