Improved performance

README.md

@@ -1,6 +1,6 @@
 # markdown
 
-Markdown renderer.
+A markdown renderer that supports only a subset of the CommonMark spec in order to make the rendering more efficient and the syntax more consistent.
 
 ## Features
 
@@ -43,7 +43,9 @@ coverage: 100.0% of statements
 ## Benchmarks
 
 ```
-BenchmarkSmall-12        2421213               494.2 ns/op           248 B/op          5 allocs/op
+BenchmarkSmall-12        7223979               164.2 ns/op            64 B/op          2 allocs/op
+BenchmarkMedium-12        832531              1310 ns/op             992 B/op          2 allocs/op
+BenchmarkLarge-12         295946              3732 ns/op            3712 B/op          3 allocs/op
 ```
 
 ## License

Render.go

@@ -10,8 +10,9 @@ var (
 	headerEnd   = []string{"</h1>", "</h2>", "</h3>", "</h4>", "</h5>", "</h6>"}
 )
 
+// renderer represents a Markdown to HTML renderer.
 type renderer struct {
-	out                strings.Builder
+	out                []byte
 	paragraphLevel     int
 	quoteLevel         int
 	listLevel          int
@@ -29,15 +30,17 @@ func Render(markdown string) string {
 		lineStart = 0
 	)
 
+	r.out = make([]byte, 0, nextPowerOf2(uint32(len(markdown)+4)))
+
 	for {
 		if i > len(markdown) {
 			r.closeAll()
 
 			for range r.quoteLevel {
-				r.out.WriteString("</blockquote>")
+				r.WriteString("</blockquote>")
 			}
 
-			return r.out.String()
+			return string(r.out)
 		}
 
 		if i != len(markdown) && markdown[i] != '\n' {
@@ -56,15 +59,15 @@ func Render(markdown string) string {
 func (r *renderer) processLine(line string) {
 	if r.inCodeBlock {
 		if strings.HasPrefix(line, "```") {
-			r.out.WriteString("</code></pre>")
+			r.WriteString("</code></pre>")
 			r.inCodeBlock = false
 			r.codeLines = 0
 		} else {
 			if r.codeLines != 0 {
-				r.out.WriteByte('\n')
+				r.WriteByte('\n')
 			}
 
-			r.out.WriteString(html.EscapeString(line))
+			r.WriteString(html.EscapeString(line))
 			r.codeLines++
 		}
 
@@ -82,13 +85,13 @@ func (r *renderer) processLine(line string) {
 		r.closeParagraphs()
 
 		for range newQuoteLevel - r.quoteLevel {
-			r.out.WriteString("<blockquote>")
+			r.WriteString("<blockquote>")
 		}
 	} else if newQuoteLevel < r.quoteLevel {
 		r.closeParagraphs()
 
 		for range r.quoteLevel - newQuoteLevel {
-			r.out.WriteString("</blockquote>")
+			r.WriteString("</blockquote>")
 		}
 	}
 
@@ -105,9 +108,9 @@ func (r *renderer) processLine(line string) {
 		space := strings.IndexByte(line, ' ')
 
 		if space > 0 && space <= 6 {
-			r.out.WriteString(headerStart[space-1])
+			r.WriteString(headerStart[space-1])
 			r.writeText(line[space+1:])
-			r.out.WriteString(headerEnd[space-1])
+			r.WriteString(headerEnd[space-1])
 		}
 
 		return
@@ -117,13 +120,13 @@ func (r *renderer) processLine(line string) {
 		line = strings.TrimSpace(line[1:])
 
 		if r.listLevel == 0 {
-			r.out.WriteString("<ul>")
+			r.WriteString("<ul>")
 			r.listLevel++
 		}
 
-		r.out.WriteString("<li>")
+		r.WriteString("<li>")
 		r.writeText(line)
-		r.out.WriteString("</li>")
+		r.WriteString("</li>")
 		return
 
 	case '`':
@@ -132,11 +135,11 @@ func (r *renderer) processLine(line string) {
 
 			if !r.inCodeBlock {
 				if language != "" {
-					r.out.WriteString("<pre><code class=\"language-")
+					r.WriteString("<pre><code class=\"language-")
-					r.out.WriteString(html.EscapeString(language))
+					r.WriteString(html.EscapeString(language))
-					r.out.WriteString("\">")
+					r.WriteString("\">")
 				} else {
-					r.out.WriteString("<pre><code>")
+					r.WriteString("<pre><code>")
 				}
 
 				r.inCodeBlock = true
@@ -150,7 +153,7 @@ func (r *renderer) processLine(line string) {
 		line = line[1:]
 
 		if r.tableLevel == 0 {
-			r.out.WriteString("<table><thead>")
+			r.WriteString("<table><thead>")
 			r.tableLevel++
 		}
 
@@ -160,30 +163,30 @@ func (r *renderer) processLine(line string) {
 			pipe := strings.IndexByte(line, '|')
 
 			if pipe == -1 {
-				r.out.WriteString("</tr>")
+				r.WriteString("</tr>")
 				return
 			}
 
 			content := strings.TrimSpace(line[:pipe])
 
 			if strings.HasPrefix(content, "---") {
-				r.out.WriteString("</thead><tbody>")
+				r.WriteString("</thead><tbody>")
 				r.tableHeaderWritten = true
 				return
 			}
 
 			if column == 0 {
-				r.out.WriteString("<tr>")
+				r.WriteString("<tr>")
 			}
 
 			if r.tableHeaderWritten {
-				r.out.WriteString("<td>")
+				r.WriteString("<td>")
 				r.writeText(content)
-				r.out.WriteString("</td>")
+				r.WriteString("</td>")
 			} else {
-				r.out.WriteString("<th>")
+				r.WriteString("<th>")
 				r.writeText(content)
-				r.out.WriteString("</th>")
+				r.WriteString("</th>")
 			}
 
 			line = line[pipe+1:]
@@ -192,13 +195,13 @@ func (r *renderer) processLine(line string) {
 	}
 
 	if r.paragraphLevel == 0 {
-		r.out.WriteString("<p>")
+		r.WriteString("<p>")
 		r.paragraphLevel++
 		r.writeText(line)
 		return
 	}
 
-	r.out.WriteByte(' ')
+	r.WriteByte(' ')
 	r.writeText(line)
 }
 
@@ -212,7 +215,7 @@ func (r *renderer) closeAll() {
 // closeParagraphs closes open paragraphs.
 func (r *renderer) closeParagraphs() {
 	for range r.paragraphLevel {
-		r.out.WriteString("</p>")
+		r.WriteString("</p>")
 	}
 
 	r.paragraphLevel = 0
@@ -221,7 +224,7 @@ func (r *renderer) closeParagraphs() {
 // closeLists closes open lists.
 func (r *renderer) closeLists() {
 	for range r.listLevel {
-		r.out.WriteString("</ul>")
+		r.WriteString("</ul>")
 	}
 
 	r.listLevel = 0
@@ -230,7 +233,7 @@ func (r *renderer) closeLists() {
 // closeTables closes open tables.
 func (r *renderer) closeTables() {
 	for range r.tableLevel {
-		r.out.WriteString("</tbody></table>")
+		r.WriteString("</tbody></table>")
 	}
 
 	r.tableLevel = 0
@@ -253,7 +256,7 @@ func (r *renderer) writeText(markdown string) {
 
 	for {
 		if i == len(markdown) {
-			r.out.WriteString(html.EscapeString(markdown[tokenStart:]))
+			r.WriteString(html.EscapeString(markdown[tokenStart:]))
 			return
 		}
 
@@ -261,7 +264,7 @@ func (r *renderer) writeText(markdown string) {
 
 		switch c {
 		case '[':
-			r.out.WriteString(html.EscapeString(markdown[tokenStart:i]))
+			r.WriteString(html.EscapeString(markdown[tokenStart:i]))
 			tokenStart = i
 			textStart = i
 		case ']':
@@ -279,11 +282,11 @@ func (r *renderer) writeText(markdown string) {
 				linkText := markdown[textStart+1 : textEnd]
 				linkURL := markdown[urlStart+1 : i]
 
-				r.out.WriteString("<a href=\"")
+				r.WriteString("<a href=\"")
-				r.out.WriteString(sanitizeURL(linkURL))
+				r.WriteString(sanitizeURL(linkURL))
-				r.out.WriteString("\">")
+				r.WriteString("\">")
-				r.out.WriteString(html.EscapeString(linkText))
+				r.WriteString(html.EscapeString(linkText))
-				r.out.WriteString("</a>")
+				r.WriteString("</a>")
 
 				textStart = -1
 				textEnd = -1
@@ -307,3 +310,27 @@ func sanitizeURL(linkURL string) string {
 
 	return html.EscapeString(linkURL)
 }
+
+// WriteByte adds a single byte to the output.
+func (r *renderer) WriteByte(b byte) error {
+	r.out = append(r.out, b)
+	return nil
+}
+
+// WriteString adds a string to the output.
+func (r *renderer) WriteString(text string) (int, error) {
+	r.out = append(r.out, text...)
+	return len(text), nil
+}
+
+// nextPowerOf2 calculates the next 32-bit power of 2.
+func nextPowerOf2(x uint32) uint32 {
+	x--
+	x |= x >> 1
+	x |= x >> 2
+	x |= x >> 4
+	x |= x >> 8
+	x |= x >> 16
+	x++
+	return x
+}

benchmarks_test.go

@@ -1,13 +1,39 @@
 package markdown_test
 
 import (
+	"os"
 	"testing"
 
+	"git.akyoto.dev/go/assert"
 	"git.akyoto.dev/go/markdown"
 )
 
 func BenchmarkSmall(b *testing.B) {
+	small, err := os.ReadFile("testdata/small.md")
+	assert.Nil(b, err)
+	input := string(small)
+
 	for range b.N {
-		markdown.Render("# Header\nText.\nText.\n# Header\nText.\nText.")
+		markdown.Render(input)
+	}
+}
+
+func BenchmarkMedium(b *testing.B) {
+	medium, err := os.ReadFile("testdata/medium.md")
+	assert.Nil(b, err)
+	input := string(medium)
+
+	for range b.N {
+		markdown.Render(input)
+	}
+}
+
+func BenchmarkLarge(b *testing.B) {
+	small, err := os.ReadFile("testdata/large.md")
+	assert.Nil(b, err)
+	input := string(small)
+
+	for range b.N {
+		markdown.Render(input)
 	}
 }

testdata/large.md

@@ -0,0 +1,37 @@
+# Lorem Ipsum
+
+Lorem ipsum dolor sit amet, consectetur adipiscing elit, sed do eiusmod tempor incididunt ut labore et dolore magna aliqua. Ut enim ad minim veniam, quis nostrud exercitation ullamco laboris nisi ut aliquip ex ea commodo consequat. Duis aute irure dolor in reprehenderit in voluptate velit esse cillum dolore eu fugiat nulla pariatur. Excepteur sint occaecat cupidatat non proident, sunt in culpa qui officia deserunt mollit anim id est laborum.
+
+# Code
+
+```shell
+mkdir example
+cd example
+touch example.txt
+```
+
+# Formatting
+
+*Italic*, **bold**, `monospace` and [links](https://example.com/).
+
+# Lists
+
+- List entry
+- List entry
+- List entry
+
+# Tables
+
+|  X  |  Y  |
+| --- | --- |
+|  0  |  0  |
+|  1  |  2  |
+|  2  |  4  |
+|  3  |  6  |
+|  4  |  8  |
+
+# Quotes
+
+> Blockquote
+> Blockquote
+> Blockquote

testdata/medium.md

@@ -0,0 +1,3 @@
+# Lorem Ipsum
+
+Lorem ipsum dolor sit amet, consectetur adipiscing elit, sed do eiusmod tempor incididunt ut labore et dolore magna aliqua. Ut enim ad minim veniam, quis nostrud exercitation ullamco laboris nisi ut aliquip ex ea commodo consequat. Duis aute irure dolor in reprehenderit in voluptate velit esse cillum dolore eu fugiat nulla pariatur. Excepteur sint occaecat cupidatat non proident, sunt in culpa qui officia deserunt mollit anim id est laborum.

testdata/small.md

@@ -0,0 +1,3 @@
+# Title
+
+Text.