fixing bug with tabs in log line for MDC

IQSS · Mar 22, 2024 · 2129555 · 2129555
1 parent 63a09cb
commit 2129555
Show file tree

Hide file tree

Showing 2 changed files with 21 additions and 10 deletions.
diff --git a/src/main/java/edu/harvard/iq/dataverse/makedatacount/MakeDataCountLoggingServiceBean.java b/src/main/java/edu/harvard/iq/dataverse/makedatacount/MakeDataCountLoggingServiceBean.java
@@ -46,6 +46,12 @@ public void logEntry(MakeDataCountEntry entry) {
     public String getLogFileName() {
         return "counter_"+new SimpleDateFormat("yyyy-MM-dd").format(new Timestamp(new Date().getTime()))+".log";
     }
+
+    // Sanitize the values to a safe string for the log file
+    static String sanitize(String in) {
+        // Log lines are tab delimited so tabs must be replaced. Replacing escape sequences with a space.
+        return in != null ? in.replaceAll("\\s+", " ") : null;
+    }
 
     public static class MakeDataCountEntry {
 
@@ -367,7 +373,7 @@ public String getTitle() {
          * @param title the title to set
          */
         public final void setTitle(String title) {
-            this.title = title;
+            this.title = sanitize(title);
         }
 
         /**
@@ -384,7 +390,7 @@ public String getPublisher() {
          * @param publisher the publisher to set
          */
         public final void setPublisher(String publisher) {
-            this.publisher = publisher;
+            this.publisher = sanitize(publisher);
         }
 
         /**
@@ -401,7 +407,7 @@ public String getPublisherId() {
          * @param publisherId the publisherId to set
          */
         public final void setPublisherId(String publisherId) {
-            this.publisherId = publisherId;
+            this.publisherId = sanitize(publisherId);
         }
 
         /**
@@ -418,7 +424,7 @@ public String getAuthors() {
          * @param authors the authors to set
          */
         public final void setAuthors(String authors) {
-            this.authors = authors;
+            this.authors = sanitize(authors);
         }
 
         /**
@@ -452,7 +458,7 @@ public String getVersion() {
          * @param version the version to set
          */
         public final void setVersion(String version) {
-            this.version = version;
+            this.version = sanitize(version);
         }
 
         /**
@@ -469,7 +475,7 @@ public String getOtherId() {
          * @param otherId the otherId to set
          */
         public void setOtherId(String otherId) {
-            this.otherId = otherId;
+            this.otherId = sanitize(otherId);
         }
 
         /**

diff --git a/...test/java/edu/harvard/iq/dataverse/makedatacount/MakeDataCountLoggingServiceBeanTest.java b/...test/java/edu/harvard/iq/dataverse/makedatacount/MakeDataCountLoggingServiceBeanTest.java
@@ -21,7 +21,6 @@
 import static org.hamcrest.CoreMatchers.not;
 import static org.hamcrest.MatcherAssert.assertThat;
 
-import org.hamcrest.MatcherAssert;
 import org.junit.jupiter.api.Test;
 
 /**
@@ -45,8 +44,8 @@ public void testMainAndFileConstructor() {
         GlobalId id = dataset.getGlobalId();
         dataset.setGlobalId(id);
         dvVersion.setDataset(dataset);
-        dvVersion.setAuthorsStr("OneAuthor;TwoAuthor");
-        dvVersion.setTitle("Title");
+        dvVersion.setAuthorsStr("OneAuthor;\tTwoAuthor");
+        dvVersion.setTitle("Title\tWith Tab");
         dvVersion.setVersionNumber(1L);
         dvVersion.setReleaseTime(new Date());
 
@@ -64,7 +63,13 @@ public void testMainAndFileConstructor() {
 
         //lastly setting attributes we don't actually use currently in our logging/constructors, just in case
         entry.setUserCookieId("UserCookId");
-        entry.setOtherId("OtherId");
+        entry.setOtherId(null); // null pointer check for sanitize method
+        assertThat(entry.getOtherId(), is("-"));
+        entry.setOtherId("OtherId\t\r\nX");
+        // escape sequences get replaced with a space in sanitize method
+        assertThat(entry.getOtherId(), is("OtherId X"));
+        // check other replacements for author list ";" becomes "|"
+        assertThat(entry.getAuthors(), is("OneAuthor| TwoAuthor"));
 
         //And test. "-" is the default
         assertThat(entry.getEventTime(), is(not("-")));