handing <br> and paragraphs

Mario Hros · Mario Hros · commit 514e0f424729 · 2017-05-06T16:34:27.000+02:00
diff --git a/html2text.go b/html2text.go
@@ -53,6 +53,9 @@ func HTMLEntitiesToText(htmlEntsText string) string {
 
 	for i, r := range htmlEntsText {
 		switch {
+		case r == '\n', r == '\r': // skip new lines
+			continue
+
 		case r == ';' && inEnt:
 			inEnt = false
 			continue
@@ -96,21 +99,30 @@ func HTMLEntitiesToText(htmlEntsText string) string {
 
 // HTML2Text converts html into a text form
 func HTML2Text(html string) string {
+	inLen := len(html)
 	tagStart := 0
 	inEnt := false
 	badTagStackDepth := 0 // if == 1 it means we are inside <head>...</head>
 	shouldOutput := true
+	// new line cannot be printed at the beginning or
+	// for <p> after a new line created by previous <p></p>
+	canPrintNewline := false
 
 	outBuf := bytes.NewBufferString("")
 
 	for i, r := range html {
+		if inLen > 0 && i == inLen-1 {
+			// prevent new line at the end of the document
+			canPrintNewline = false
+		}
+
 		switch {
-		case r == ';' && inEnt:
+		case r == ';' && inEnt: // end of html entity
 			inEnt = false
 			shouldOutput = true
 			continue
 
-		case r == '&' && shouldOutput: //possible html entity
+		case r == '&' && shouldOutput: // possible html entity
 			entName := ""
 			isEnt := false
 
@@ -139,15 +151,25 @@ func HTML2Text(html string) string {
 				}
 			}
 
-		case r == '<':
+		case r == '<': // start of a tag
 			tagStart = i + 1
 			shouldOutput = false
 			continue
-		case r == '>':
+
+		case r == '>': // end of a tag
 			shouldOutput = true
 			tagName := strings.ToLower(html[tagStart:i])
 
-			if badTagnamesRE.MatchString(tagName) {
+			if tagName == "br" || tagName == "br/" {
+				// new line
+				outBuf.WriteString("\r\n")
+			} else if tagName == "p" || tagName == "/p" {
+				if canPrintNewline {
+					outBuf.WriteString("\r\n")
+				}
+				canPrintNewline = false
+			} else if badTagnamesRE.MatchString(tagName) {
+				// unwanted block
 				badTagStackDepth++
 
 				// parse link href
@@ -162,16 +184,17 @@ func HTML2Text(html string) string {
 						outBuf.WriteString(HTMLEntitiesToText(link))
 					}
 				}
-
 			} else if len(tagName) > 0 && tagName[0] == '/' &&
 				badTagnamesRE.MatchString(tagName[1:]) {
+				// end of unwanted block
 				badTagStackDepth--
 			}
-
 			continue
-		}
+
+		} // switch end
 
 		if shouldOutput && badTagStackDepth == 0 && !inEnt {
+			canPrintNewline = true
 			outBuf.WriteRune(r)
 		}
 	}
diff --git a/html2text_test.go b/html2text_test.go
@@ -20,6 +20,11 @@ func TestHTML2Text(t *testing.T) {
 			So(HTML2Text(`some <div id="a" class="b">div</div>`), ShouldEqual, "some div")
 		})
 
+		Convey("Line breaks", func() {
+			So(HTML2Text(`two<br>line<br/>breaks`), ShouldEqual, "two\r\nline\r\nbreaks")
+			So(HTML2Text(`<p>two</p><p>paragraphs</p>`), ShouldEqual, "two\r\nparagraphs")
+		})
+
 		Convey("HTML entities", func() {
 			So(HTML2Text(`two&nbsp;&nbsp;spaces`), ShouldEqual, "two  spaces")
 			So(HTML2Text(`&copy; 2017 K3A`), ShouldEqual, "© 2017 K3A")
@@ -31,6 +36,7 @@ func TestHTML2Text(t *testing.T) {
 		})
 
 		Convey("Full HTML structure", func() {
+			So(HTML2Text(``), ShouldEqual, "")
 			So(HTML2Text(`<html><head><title>Good</title></head><body>x</body>`), ShouldEqual, "x")
 			So(HTML2Text(`we are not <script type="javascript"></script>interested in scripts`),
 				ShouldEqual, "we are not interested in scripts")
