Support "dash" in HtmlTagIntegrityChecker

box · Sep 3, 2024 · a3ffeb3 · a3ffeb3
1 parent a87c783
commit a3ffeb3
Show file tree

Hide file tree

Showing 3 changed files with 13 additions and 3 deletions.
diff --git a/...x/l10n/mojito/service/assetintegritychecker/integritychecker/HtmlTagIntegrityChecker.java b/...x/l10n/mojito/service/assetintegritychecker/integritychecker/HtmlTagIntegrityChecker.java
@@ -21,7 +21,7 @@ public class HtmlTagIntegrityChecker extends RegexIntegrityChecker {
 
   @Override
   public String getRegex() {
-    return "(<\\w+(\\s+\\w+(\\s*=\\s*('([^']*?)'|\"([^\"]*?)\"))?)*\\s*/?>|</\\w+>)";
+    return "(<[a-zA-Z][\\w-]*(\\s+\\w+(\\s*=\\s*('([^']*?)'|\"([^\"]*?)\"))?)*\\s*/?>|<\\/[a-zA-Z][\\w-]*>)";
   }
 
   @Override

diff --git a/...box/l10n/mojito/service/assetintegritychecker/integritychecker/RegexIntegrityChecker.java b/...box/l10n/mojito/service/assetintegritychecker/integritychecker/RegexIntegrityChecker.java
@@ -1,6 +1,6 @@
 package com.box.l10n.mojito.service.assetintegritychecker.integritychecker;
 
-import java.util.HashSet;
+import java.util.LinkedHashSet;
 import java.util.Set;
 import java.util.regex.Matcher;
 import java.util.regex.Pattern;
@@ -68,7 +68,7 @@ Pattern getPattern() {
    */
   Set<String> getPlaceholders(String string) {
 
-    Set<String> placeholders = new HashSet<>();
+    Set<String> placeholders = new LinkedHashSet<>();
 
     if (string != null) {
 

diff --git a/...0n/mojito/service/assetintegritychecker/integritychecker/HtmlTagIntegrityCheckerTest.java b/...0n/mojito/service/assetintegritychecker/integritychecker/HtmlTagIntegrityCheckerTest.java
@@ -1,5 +1,7 @@
 package com.box.l10n.mojito.service.assetintegritychecker.integritychecker;
 
+import java.util.Set;
+import org.assertj.core.api.Assertions;
 import org.junit.Test;
 import org.slf4j.Logger;
 import org.slf4j.LoggerFactory;
@@ -191,4 +193,12 @@ public void testHtmlTagCheckNonTagLessThanDoesntConfuseThings() {
 
     checker.check(source, target);
   }
+
+  @Test
+  public void testHtmlTagCheckWithDash() {
+    String source = "<sign-up>Mag-sign up</sign-up> o <login> upang </login>";
+    Set<String> placeholders = checker.getPlaceholders(source);
+    Assertions.assertThat(placeholders)
+        .containsExactly("<sign-up>", "</sign-up>", "<login>", "</login>");
+  }
 }