restructure pipelines

README.md

@@ -38,7 +38,7 @@ Convert Markdown to Gemini gemtext[1] markup with Go. Working with streams and p
 
 => https://gemini.circumlunar.space/docs/gemtext.gmi 1: gemtext
 
- See the gemini protocol[1] and the protocol spec[2].
+See the gemini protocol[1] and the protocol spec[2].
 
 => https://gemini.circumlunar.space/ 1: gemini protocol
 => https://gemini.circumlunar.space/docs/specification.gmi 2: protocol spec

main.go

@@ -94,8 +94,15 @@ func main() {
 	source := InputStream(r)
 	sink := OutputStream(w)
 	preproc := NewPreproc()
-	proc := NewProc()
 
 	//sink.Input(preproc.Process(source.Output()))
-	sink.Input(proc.Process(preproc.Process(source.Output())))
+	sink.Input(
+		RemoveComments(
+			FormatHeadings(
+				FormatLinks(
+					preproc.Process(source.Output()),
+				),
+			),
+		),
+	)
 }

preproc.go

@@ -81,7 +81,7 @@ func normal(m *fsm, data []byte) stateFn {
 	if needsFence(data) {
 		m.out <- []byte("```\n")
 		m.out <- append(data[4:], '\n')
-		m.pending = []byte("```")
+		m.pending = []byte("```\n")
 		return toFence
 	}
 	if data[len(data)-1] != '.' {
@@ -105,7 +105,11 @@ func fence(m *fsm, data []byte) stateFn {
 }
 
 func toFence(m *fsm, data []byte) stateFn {
-	m.out <- append(data[4:], '\n')
+	if len(data) >= 3 {
+		m.out <- append(data[4:], '\n')
+	} else {
+		//m.out <- []byte("\n")
+	}
 	if needsFence(data) {
 		return toFence
 	}

proc.go

@@ -6,55 +6,67 @@ import (
 	"regexp"
 )
 
-// state machine
-type proc struct {
-	out chan []byte
-}
-
-func NewProc() *proc {
-	return &proc{}
-}
-
-func (m *proc) Process(in chan []byte) chan []byte {
-	m.out = make(chan []byte)
+func FormatLinks(in chan []byte) chan []byte {
+	out := make(chan []byte)
 	go func() {
 		for b := range in {
-			m.out <- m.process(b)
+			out <- formatLinks(b)
 		}
-		close(m.out)
+		close(out)
 	}()
-	return m.out
+	return out
 }
 
-func (m *proc) process(data []byte) []byte {
+func formatLinks(data []byte) []byte {
 	// find link name and url
 	var buffer []byte
 	re := regexp.MustCompile(`!?\[([^\]*]*)\]\(([^)]*)\)`)
 	for i, match := range re.FindAllSubmatch(data, -1) {
 		replaceWithIndex := append(match[1], fmt.Sprintf("[%d]", i+1)...)
 		data = bytes.Replace(data, match[0], replaceWithIndex, 1)
+		// append entry to buffer to be added later
 		link := fmt.Sprintf("=> %s %d: %s\n", match[2], i+1, match[1])
 		buffer = append(buffer, link...)
 	}
+	// append links to that paragraph
 	if len(buffer) > 0 {
 		data = append(data, []byte("\n")...)
 		data = append(data, buffer...)
 	}
 
-	// remove comments
-	re2 := regexp.MustCompile(`<!--.*-->`)
-	data = re2.ReplaceAll(data, []byte{})
-
-	// collapse headings
-	re3 := regexp.MustCompile(`^[#]{4,}`)
-	data = re3.ReplaceAll(data, []byte("###"))
-
-	// heading without spacing
-	re4 := regexp.MustCompile(`^(#+)[^# ]`)
-	sub := re4.FindSubmatch(data)
-	if len(sub) > 0 {
-		data = bytes.Replace(data, sub[1], append(sub[1], []byte(" ")...), 1)
-	}
-
 	return data
 }
+
+func RemoveComments(in chan []byte) chan []byte {
+	out := make(chan []byte)
+	go func() {
+		re := regexp.MustCompile(`<!--.*-->`)
+		for b := range in {
+			out <- re.ReplaceAll(b, []byte{})
+		}
+		close(out)
+	}()
+	return out
+}
+
+func FormatHeadings(in chan []byte) chan []byte {
+	out := make(chan []byte)
+	go func() {
+		re := regexp.MustCompile(`^[#]{4,}`)
+		re2 := regexp.MustCompile(`^(#+)[^# ]`)
+		for b := range in {
+			// fix up more than 4 levels
+			b = re.ReplaceAll(b, []byte("###"))
+			// ensure we have a space
+			sub := re2.FindSubmatch(b)
+			if len(sub) > 0 {
+				b = bytes.Replace(b, sub[1], append(sub[1], []byte(" ")...), 1)
+			}
+			// writeback
+			out <- b
+
+		}
+		close(out)
+	}()
+	return out
+}