output validates, command-line tool has useful options

README.md

@@ -53,18 +53,14 @@ All features of upskirt are supported, including:
 
 *   Good performance. I have not done rigorous benchmarking, but
     informal testing suggests it is around 3.5x slower than upskirt.
-    This is an ugly, direct translation from the C code, so
-    the difference is unlikely to be related to differences in
-    coding style. There is a lot of bounds checking that is
-    duplicated (by user code for the application and again by code
-    the compiler generates) and there is some additional memory
-    management overhead, since I allocate and garbage collect
-    buffers instead of explicitly managing them as upskirt does.
 
 *   Minimal dependencies. blackfriday only depends on standard
     library packages in Go. The source code is pretty
     self-contained, so it is easy to add to any project.
 
+*   Output successfully validates using the W3C validation tool for
+    HTML 4.01 and XHTML 1.0 Transitional.
+
 
 Extensions
 ----------
@@ -83,15 +79,12 @@ LaTeX Output
 ------------
 
 A rudimentary LaTeX rendering backend is also included. To see an
-example of its usage, comment out this link in `main.go`:
+example of its usage, see `main.go`:
 
-    renderer := blackfriday.HtmlRenderer(html_flags)
-
-and uncomment this line:
-
-    renderer := blackfriday.LatexRenderer(0)
-
-It renders some basic documents, but is only experimental at this point.
+It renders some basic documents, but is only experimental at this
+point. In particular, it does not do any inline escaping, so input
+that happens to look like LaTeX code will be passed through without
+modification.
 
 
 Todo

example/main.go

@@ -13,6 +13,7 @@
 package main
 
 import (
+	"flag"
 	"fmt"
 	"io/ioutil"
 	"github.com/russross/blackfriday"
@@ -20,22 +21,55 @@ import (
 )
 
 func main() {
+	// parse command-line options
+	var page, xhtml, latex, smartypants bool
+	var css string
+	var repeat int
+	flag.BoolVar(&page, "page", false,
+		"Generate a standalone HTML page (implies -latex=false)")
+	flag.BoolVar(&xhtml, "xhtml", true,
+		"Use XHTML-style tags in HTML output")
+	flag.BoolVar(&latex, "latex", false,
+		"Generate LaTeX output instead of HTML")
+	flag.BoolVar(&smartypants, "smartypants", false,
+		"Apply smartypants-style substitutions")
+	flag.StringVar(&css, "css", "",
+		"Link to a CSS stylesheet (implies -page)")
+	flag.IntVar(&repeat, "repeat", 1,
+		"Process the input multiple times (for benchmarking)")
+	flag.Usage = func() {
+		fmt.Fprintf(os.Stderr, "Usage:\n"+
+			"  %s [options] [inputfile [outputfile]]\n\n"+
+			"Options:\n", os.Args[0])
+		flag.PrintDefaults()
+	}
+	flag.Parse()
+
+	// enforce implied options
+	if css != "" {
+		page = true
+	}
+	if page {
+		latex = false
+	}
+
 	// read the input
 	var input []byte
 	var err os.Error
-	switch len(os.Args) {
-	case 1:
+	args := flag.Args()
+	switch len(args) {
+	case 0:
 		if input, err = ioutil.ReadAll(os.Stdin); err != nil {
 			fmt.Fprintln(os.Stderr, "Error reading from Stdin:", err)
 			os.Exit(-1)
 		}
-	case 2, 3:
-		if input, err = ioutil.ReadFile(os.Args[1]); err != nil {
-			fmt.Fprintln(os.Stderr, "Error reading from", os.Args[1], ":", err)
+	case 1, 2:
+		if input, err = ioutil.ReadFile(args[0]); err != nil {
+			fmt.Fprintln(os.Stderr, "Error reading from", args[0], ":", err)
 			os.Exit(-1)
 		}
 	default:
-		fmt.Fprintln(os.Stderr, "Usage:", os.Args[0], "[inputfile [outputfile]]")
+		flag.Usage()
 		os.Exit(-1)
 	}
 
@@ -48,33 +82,70 @@ func main() {
 	extensions |= blackfriday.EXTENSION_STRIKETHROUGH
 	extensions |= blackfriday.EXTENSION_SPACE_HEADERS
 
-	html_flags := 0
-	html_flags |= blackfriday.HTML_USE_XHTML
-	// note: uncomment the following line to enable smartypants
-	// it is commented out by default so that markdown
-	// compatibility tests pass
-	//html_flags |= blackfriday.HTML_USE_SMARTYPANTS
-	html_flags |= blackfriday.HTML_SMARTYPANTS_FRACTIONS
-	html_flags |= blackfriday.HTML_SMARTYPANTS_LATEX_DASHES
+	var renderer *blackfriday.Renderer
+	if latex {
+		// render the data into LaTeX
+		renderer = blackfriday.LatexRenderer(0)
+	} else {
+		// render the data into HTML
+		html_flags := 0
+		if xhtml {
+			html_flags |= blackfriday.HTML_USE_XHTML
+		}
+		if smartypants {
+			html_flags |= blackfriday.HTML_USE_SMARTYPANTS
+			html_flags |= blackfriday.HTML_SMARTYPANTS_FRACTIONS
+			html_flags |= blackfriday.HTML_SMARTYPANTS_LATEX_DASHES
+		}
+		renderer = blackfriday.HtmlRenderer(html_flags)
+	}
 
-	// render the data into HTML (comment this out to deselect HTML)
-	renderer := blackfriday.HtmlRenderer(html_flags)
-
-	// render the data into LaTeX (uncomment to select LaTeX)
-	//renderer := blackfriday.LatexRenderer(0)
-
-	output := blackfriday.Markdown(input, renderer, extensions)
+	// parse and render
+	var output []byte
+	for i := 0; i < repeat; i++ {
+		output = blackfriday.Markdown(input, renderer, extensions)
+	}
 
 	// output the result
-	if len(os.Args) == 3 {
-		if err = ioutil.WriteFile(os.Args[2], output, 0644); err != nil {
-			fmt.Fprintln(os.Stderr, "Error writing to", os.Args[2], ":", err)
+	var out *os.File
+	if len(args) == 2 {
+		if out, err = os.Create(args[1]); err != nil {
+			fmt.Fprintf(os.Stderr, "Error creating %s: %v", args[1], err)
 			os.Exit(-1)
 		}
+		defer out.Close()
 	} else {
-		if _, err = os.Stdout.Write(output); err != nil {
-			fmt.Fprintln(os.Stderr, "Error writing to Stdout:", err)
-			os.Exit(-1)
+		out = os.Stdout
+	}
+
+	if page {
+		ending := ""
+		if xhtml {
+			fmt.Fprint(out, "<!DOCTYPE html PUBLIC \"-//W3C//DTD XHTML 1.0 Transitional//EN\" ")
+			fmt.Fprintln(out, "\"http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd\">")
+			fmt.Fprintln(out, "<html xmlns=\"http://www.w3.org/1999/xhtml\">")
+			ending = " /"
+		} else {
+			fmt.Fprint(out, "<!DOCTYPE html PUBLIC \"-//W3C//DTD HTML 4.01//EN\" ")
+			fmt.Fprintln(out, "\"http://www.w3.org/TR/html4/strict.dtd\">")
+			fmt.Fprintln(out, "<html>")
 		}
+		fmt.Fprintln(out, "<head>")
+		fmt.Fprintln(out, "  <title></title>")
+		fmt.Fprintf(out, "  <meta name=\"GENERATOR\" content=\"Blackfriday markdown processor\"%s>\n", ending)
+		fmt.Fprintf(out, "  <meta http-equiv=\"Content-Type\" content=\"text/html; charset=utf-8\"%s>\n", ending)
+		if css != "" {
+			fmt.Fprintf(out, "  <link rel=\"stylesheet\" type=\"text/css\" href=\"%s\" />\n", css)
+		}
+		fmt.Fprintln(out, "</head>")
+		fmt.Fprintln(out, "<body>")
+	}
+	if _, err = out.Write(output); err != nil {
+		fmt.Fprintln(os.Stderr, "Error writing output:", err)
+		os.Exit(-1)
+	}
+	if page {
+		fmt.Fprintln(out, "</body>")
+		fmt.Fprintln(out, "</html>")
 	}
 }

html.go

@@ -419,7 +419,7 @@ func htmlAutoLink(out *bytes.Buffer, link []byte, kind int, opaque interface{})
 	if kind == LINK_TYPE_EMAIL {
 		out.WriteString("mailto:")
 	}
-	out.Write(link)
+	attrEscape(out, link)
 	out.WriteString("\">")
 
 	/*
@@ -504,17 +504,13 @@ func htmlLink(out *bytes.Buffer, link []byte, title []byte, content []byte, opaq
 	}
 
 	out.WriteString("<a href=\"")
-	if len(link) > 0 {
-		out.Write(link)
-	}
+	attrEscape(out, link)
 	if len(title) > 0 {
 		out.WriteString("\" title=\"")
 		attrEscape(out, title)
 	}
 	out.WriteString("\">")
-	if len(content) > 0 {
-		out.Write(content)
-	}
+	out.Write(content)
 	out.WriteString("</a>")
 	return 1
 }

inline.go

@@ -124,7 +124,7 @@ func inlineCodeSpan(out *bytes.Buffer, rndr *render, data []byte, offset int) in
 	}
 
 	if i < nb && end >= len(data) {
-        out.WriteByte('`')
+		out.WriteByte('`')
 		return 0 // no matching delimiter
 	}
 
@@ -185,6 +185,11 @@ func inlineLineBreak(out *bytes.Buffer, rndr *render, data []byte, offset int) i
 
 // '[': parse a link or an image
 func inlineLink(out *bytes.Buffer, rndr *render, data []byte, offset int) int {
+	// no links allowed inside other links
+	if rndr.insideLink {
+		return 0
+	}
+
 	isImg := offset > 0 && data[offset-1] == '!'
 
 	data = data[offset:]
@@ -410,7 +415,11 @@ func inlineLink(out *bytes.Buffer, rndr *render, data []byte, offset int) int {
 		if isImg {
 			content.Write(data[1:txt_e])
 		} else {
+			// links cannot contain other links, so turn off link parsing temporarily
+			insideLink := rndr.insideLink
+			rndr.insideLink = true
 			parseInline(&content, rndr, data[1:txt_e])
+			rndr.insideLink = insideLink
 		}
 	}
 
@@ -539,7 +548,7 @@ func inlineEntity(out *bytes.Buffer, rndr *render, data []byte, offset int) int
 
 func inlineAutoLink(out *bytes.Buffer, rndr *render, data []byte, offset int) int {
 	// quick check to rule out most false hits on ':'
-	if len(data) < offset+3 || data[offset+1] != '/' || data[offset+2] != '/' {
+	if rndr.insideLink || len(data) < offset+3 || data[offset+1] != '/' || data[offset+2] != '/' {
 		return 0
 	}
 

markdown.go

@@ -141,6 +141,7 @@ type render struct {
 	flags      uint32
 	nesting    int
 	maxNesting int
+	insideLink bool
 }
 
 
@@ -165,6 +166,7 @@ func Markdown(input []byte, renderer *Renderer, extensions uint32) []byte {
 	rndr.flags = extensions
 	rndr.refs = make(map[string]*reference)
 	rndr.maxNesting = 16
+	rndr.insideLink = false
 
 	// register inline parsers
 	if rndr.mk.Emphasis != nil || rndr.mk.DoubleEmphasis != nil || rndr.mk.TripleEmphasis != nil {
@@ -464,7 +466,7 @@ func expandTabs(out *bytes.Buffer, line []byte) {
 	// the slow case: we need to count runes to figure out how
 	// many spaces to insert for each tab
 	column := 0
-    i = 0
+	i = 0
 	for i < len(line) {
 		start := i
 		for i < len(line) && line[i] != '\t' {

upskirtref/Amps and angle encoding.html

@@ -8,10 +8,10 @@
 
 <p>6 &gt; 5.</p>
 
-<p>Here's a <a href="http://example.com/?foo=1&bar=2">link</a> with an ampersand in the URL.</p>
+<p>Here's a <a href="http://example.com/?foo=1&amp;bar=2">link</a> with an ampersand in the URL.</p>
 
 <p>Here's a link with an amersand in the link text: <a href="http://att.com/" title="AT&amp;T">AT&amp;T</a>.</p>
 
-<p>Here's an inline <a href="/script?foo=1&bar=2">link</a>.</p>
+<p>Here's an inline <a href="/script?foo=1&amp;bar=2">link</a>.</p>
 
-<p>Here's an inline <a href="/script?foo=1&bar=2">link</a>.</p>
+<p>Here's an inline <a href="/script?foo=1&amp;bar=2">link</a>.</p>

upskirtref/Auto links.html

@@ -1,6 +1,6 @@
 <p>Link: <a href="http://example.com/">http://example.com/</a>.</p>
 
-<p>With an ampersand: <a href="http://example.com/?foo=1&bar=2">http://example.com/?foo=1&amp;bar=2</a></p>
+<p>With an ampersand: <a href="http://example.com/?foo=1&amp;bar=2">http://example.com/?foo=1&amp;bar=2</a></p>
 
 <ul>
 <li>In a list?</li>