Fix #31: Prettify leading and trailing spaces for text nodes

sibprogrammer · sibprogrammer · commit 8aed2a2eaf28 · 2023-04-21T18:23:59.000+03:00
diff --git a/internal/utils/utils.go b/internal/utils/utils.go
@@ -14,6 +14,7 @@ import (
 	"io"
 	"os"
 	"os/exec"
+	"regexp"
 	"strings"
 )
 
@@ -101,10 +102,7 @@ func FormatXml(reader io.Reader, writer io.Writer, indent string, colors int) er
 			startTagClosed = false
 			level++
 		case xml.CharData:
-			str := string(typedToken)
-			if strings.TrimSpace(str) == "" {
-				str = ""
-			}
+			str := normalizeSpaces(string(typedToken), indent, level)
 			hasContent = str != ""
 			if hasContent && !startTagClosed {
 				_, _ = fmt.Fprint(writer, tagColor(">"))
@@ -247,10 +245,7 @@ func FormatHtml(reader io.Reader, writer io.Writer, indent string, colors int) e
 
 		switch token {
 		case html.TextToken:
-			str := string(tokenizer.Text())
-			if strings.TrimSpace(str) == "" {
-				str = ""
-			}
+			str := normalizeSpaces(string(tokenizer.Text()), indent, level)
 			hasContent = str != ""
 			_, _ = fmt.Fprint(writer, str)
 		case html.StartTagToken, html.SelfClosingTagToken:
@@ -397,3 +392,23 @@ func escapeText(input string) (string, error) {
 
 	return result, nil
 }
+
+func normalizeSpaces(input string, indent string, level int) string {
+	if strings.TrimSpace(input) == "" {
+		input = ""
+	}
+
+	regexpHead, _ := regexp.Compile("^ *\n +")
+	if regexpHead.MatchString(input) {
+		input = strings.TrimLeft(input, " \n")
+		input = "\n" + strings.Repeat(indent, level) + input
+	}
+
+	regexpTail, _ := regexp.Compile("\n +$")
+	if regexpTail.MatchString(input) {
+		input = strings.TrimRight(input, " \n")
+		input += "\n" + strings.Repeat(indent, level-1)
+	}
+
+	return input
+}
diff --git a/internal/utils/utils_test.go b/internal/utils/utils_test.go
@@ -32,6 +32,7 @@ func TestFormatXml(t *testing.T) {
 		"unformatted8.xml":  "formatted8.xml",
 		"unformatted9.xml":  "formatted9.xml",
 		"unformatted10.xml": "formatted10.xml",
+		"unformatted11.xml": "formatted11.xml",
 	}
 
 	for unformattedFile, expectedFile := range files {
@@ -55,6 +56,7 @@ func TestFormatHtml(t *testing.T) {
 		"unformatted3.html": "formatted3.html",
 		"unformatted4.html": "formatted4.html",
 		"unformatted5.html": "formatted5.html",
+		"unformatted6.html": "formatted6.html",
 		"unformatted.xml":   "formatted.xml",
 	}
 
diff --git a/test/data/html/formatted6.html b/test/data/html/formatted6.html
@@ -0,0 +1,8 @@
+<body>
+  <h1>
+    Welcome
+  </h1>
+  <p>
+    Here is no content, yet.
+  </p>
+</body>
diff --git a/test/data/html/unformatted6.html b/test/data/html/unformatted6.html
@@ -0,0 +1,8 @@
+<body>
+    <h1>
+        Welcome
+    </h1>
+    <p>
+        Here is no content, yet.
+    </p>
+</body>
diff --git a/test/data/xml/formatted11.xml b/test/data/xml/formatted11.xml
@@ -0,0 +1,9 @@
+<?xml version="1.0"?>
+<body>
+  <h1>
+    Welcome
+  </h1>
+  <p>
+    Here is no content, yet.
+  </p>
+</body>
diff --git a/test/data/xml/unformatted11.xml b/test/data/xml/unformatted11.xml
@@ -0,0 +1,9 @@
+<?xml version="1.0"?>
+<body>
+    <h1>
+        Welcome
+    </h1>
+    <p>
+        Here is no content, yet.
+    </p>
+</body>

Original file line number	Diff line number	Diff line change
`@@ -32,6 +32,7 @@ func TestFormatXml(t *testing.T) {`
`32`	`32`	`"unformatted8.xml": "formatted8.xml",`
`33`	`33`	`"unformatted9.xml": "formatted9.xml",`
`34`	`34`	`"unformatted10.xml": "formatted10.xml",`
	`35`	`+ "unformatted11.xml": "formatted11.xml",`
`35`	`36`	`}`
`36`	`37`
`37`	`38`	`for unformattedFile, expectedFile := range files {`
`@@ -55,6 +56,7 @@ func TestFormatHtml(t *testing.T) {`
`55`	`56`	`"unformatted3.html": "formatted3.html",`
`56`	`57`	`"unformatted4.html": "formatted4.html",`
`57`	`58`	`"unformatted5.html": "formatted5.html",`
	`59`	`+ "unformatted6.html": "formatted6.html",`
`58`	`60`	`"unformatted.xml": "formatted.xml",`
`59`	`61`	`}`
`60`	`62`
-Original file line number
+Diff line change
@@ @@ -0,0 +1,8 @@ @@
 +<body>
 +  <h1>
 +    Welcome
 +  </h1>
 +  <p>
 +    Here is no content, yet.
 +  </p>
 +</body>