refactor: optimizing regex pettern for html meta matching (#4235)

#### What type of PR is this? /kind improvement /area core /milestone 2.8.x #### What this PR does / why we need it: 优化去除 Html Meta 重复标签的正则表达式 see #4234 for more details. #### Which issue(s) this PR fixes: Fixes #4234 #### Does this PR introduce a user-facing change? ```release-note 优化去除 Html Meta 重复标签的正则表达式 ```
halo-dev · Jul 20, 2023 · 5eb9b68 · 5eb9b68
1 parent 6b70296
commit 5eb9b68
Show file tree

Hide file tree

Showing 3 changed files with 84 additions and 14 deletions.
diff --git a/application/src/main/java/run/halo/app/theme/dialect/DuplicateMetaTagProcessor.java b/application/src/main/java/run/halo/app/theme/dialect/DuplicateMetaTagProcessor.java
@@ -8,7 +8,6 @@
 import java.util.regex.Matcher;
 import java.util.regex.Pattern;
 import lombok.AllArgsConstructor;
-import org.apache.commons.lang3.StringUtils;
 import org.springframework.core.annotation.Order;
 import org.springframework.stereotype.Component;
 import org.thymeleaf.context.ITemplateContext;
@@ -30,7 +29,7 @@
 @Component
 @AllArgsConstructor
 public class DuplicateMetaTagProcessor implements TemplateHeadProcessor {
-    static final Pattern META_PATTERN = Pattern.compile("<meta\\s+name=\"(\\w+)\"(.*?)>");
+    static final Pattern META_PATTERN = Pattern.compile("<meta[^>]+?name=\"([^\"]+)\"[^>]*>\\n*");
 
     @Override
     public Mono<Void> process(ITemplateContext context, IModel model,
@@ -49,15 +48,17 @@ public Mono<Void> process(ITemplateContext context, IModel model,
                 while (matcher.find()) {
                     String tagLine = matcher.group(0);
                     String nameAttribute = matcher.group(1);
-                    IText metaTagNode = context.getModelFactory().createText(tagLine);
+                    // create a new text node to replace the original text node
+                    // replace multiple line breaks with one line break
+                    IText metaTagNode = context.getModelFactory()
+                        .createText(tagLine.replaceAll("\\n+", "\n"));
                     uniqueMetaTags.put(nameAttribute, new IndexedModel(i, metaTagNode));
                     text = text.replace(tagLine, "");
                 }
-                if (StringUtils.isNotBlank(text)) {
-                    IText otherText = context.getModelFactory()
-                        .createText(text);
-                    otherModel.add(new IndexedModel(i, otherText));
-                }
+                // put the rest of the text into the other model
+                IText otherText = context.getModelFactory()
+                    .createText(text);
+                otherModel.add(new IndexedModel(i, otherText));
             } else {
                 otherModel.add(new IndexedModel(i, templateEvent));
             }

diff --git a/application/src/test/java/run/halo/app/theme/dialect/DuplicateMetaTagProcessorTest.java b/application/src/test/java/run/halo/app/theme/dialect/DuplicateMetaTagProcessorTest.java
@@ -0,0 +1,51 @@
+package run.halo.app.theme.dialect;
+
+import static org.assertj.core.api.Assertions.assertThat;
+
+import java.util.regex.Matcher;
+import org.junit.jupiter.api.Test;
+
+/**
+ * Tests for {@link DuplicateMetaTagProcessor}.
+ *
+ * @author guqing
+ * @since 2.8.0
+ */
+class DuplicateMetaTagProcessorTest {
+
+    @Test
+    void extractMetaTag() {
+        // normal
+        String text = "<meta name=\"description\" content=\"a description\"/>";
+        Matcher matcher = DuplicateMetaTagProcessor.META_PATTERN.matcher(text);
+        assertThat(matcher.find()).isTrue();
+        assertThat(matcher.group(1)).isEqualTo("description");
+
+        // name and content are not in the general order
+        text = "<meta content=\"K1,K2\" name=\"keywords\"/>";
+        matcher = DuplicateMetaTagProcessor.META_PATTERN.matcher(text);
+        assertThat(matcher.find()).isTrue();
+        assertThat(matcher.group(1)).isEqualTo("keywords");
+
+        // no closing slash
+        text = "<meta content=\"K1,K2\" name=\"keywords\">";
+        matcher = DuplicateMetaTagProcessor.META_PATTERN.matcher(text);
+        assertThat(matcher.find()).isTrue();
+        assertThat(matcher.group(1)).isEqualTo("keywords");
+
+        // multiple line breaks and other stuff
+        text = """
+            <meta content="全局 Head description" name="description" />
+            
+            <style>
+              .moment .momemt-content pre.notranslate {
+                background: #f3f3f3;
+                color: #444;
+              }
+            </style>
+            """;
+        matcher = DuplicateMetaTagProcessor.META_PATTERN.matcher(text);
+        assertThat(matcher.find()).isTrue();
+        assertThat(matcher.group(1)).isEqualTo("description");
+    }
+}
diff --git a/...ication/src/test/java/run/halo/app/theme/message/ThemeMessageResolverIntegrationTest.java b/...ication/src/test/java/run/halo/app/theme/message/ThemeMessageResolverIntegrationTest.java
@@ -72,7 +72,10 @@ void messageResolverWhenDefaultTheme() {
             .isEqualTo("""
                 <!DOCTYPE html>
                 <html lang="en">
-                <head><meta charset="UTF-8"><title>Title</title></head>
+                <head>
+                    <meta charset="UTF-8">
+                    <title>Title</title>
+                </head>
                 <body>
                 index
                 <div>zh</div>
@@ -93,7 +96,10 @@ void messageResolverForEnLanguageWhenDefaultTheme() {
             .isEqualTo("""
                 <!DOCTYPE html>
                 <html lang="en">
-                <head><meta charset="UTF-8"><title>Title</title></head>
+                <head>
+                    <meta charset="UTF-8">
+                    <title>Title</title>
+                </head>
                 <body>
                 index
                 <div>en</div>
@@ -114,7 +120,10 @@ void shouldUseDefaultWhenLanguageNotSupport() {
             .isEqualTo("""
                 <!DOCTYPE html>
                 <html lang="en">
-                <head><meta charset="UTF-8"><title>Title</title></head>
+                <head>
+                    <meta charset="UTF-8">
+                    <title>Title</title>
+                </head>
                 <body>
                 index
                 <div>foo</div>
@@ -135,7 +144,10 @@ void switchTheme() throws URISyntaxException {
             .isEqualTo("""
                 <!DOCTYPE html>
                 <html lang="en">
-                <head><meta charset="UTF-8"><title>Title</title></head>
+                <head>
+                    <meta charset="UTF-8">
+                    <title>Title</title>
+                </head>
                 <body>
                 index
                 <div>zh</div>
@@ -154,7 +166,10 @@ void switchTheme() throws URISyntaxException {
             .isEqualTo("""
                 <!DOCTYPE html>
                 <html lang="en">
-                <head><meta charset="UTF-8"><title>Other theme title</title></head>
+                <head>
+                    <meta charset="UTF-8">
+                    <title>Other theme title</title>
+                </head>
                 <body>
                 <p>Other 首页</p>
                 </body>
@@ -167,7 +182,10 @@ void switchTheme() throws URISyntaxException {
             .isEqualTo("""
                 <!DOCTYPE html>
                 <html lang="en">
-                <head><meta charset="UTF-8"><title>Other theme title</title></head>
+                <head>
+                    <meta charset="UTF-8">
+                    <title>Other theme title</title>
+                </head>
                 <body>
                 <p>other index</p>
                 </body>