From bbede2de7af805eba70355c56446750b791ebb5b Mon Sep 17 00:00:00 2001
From: Divjot Arora <div.arora@databricks.com>
Date: Tue, 26 May 2026 15:22:29 +0000
Subject: [PATCH] [GH-3587] Fix parquet.thrift.string.size.limit validation

---
 .../java/org/apache/parquet/format/Util.java  |  4 +-
 .../TestParquetFileReaderMaxMessageSize.java  | 48 +++++++++++++++++++
 2 files changed, 50 insertions(+), 2 deletions(-)

diff --git a/parquet-format-structures/src/main/java/org/apache/parquet/format/Util.java b/parquet-format-structures/src/main/java/org/apache/parquet/format/Util.java
index 776fb45576..419d757702 100644
--- a/parquet-format-structures/src/main/java/org/apache/parquet/format/Util.java
+++ b/parquet-format-structures/src/main/java/org/apache/parquet/format/Util.java
@@ -417,8 +417,8 @@ private static InterningProtocol protocol(TIOStreamTransport t, int configuredMa
       // Set to default 100 MB
       maxMessageSize = DEFAULT_MAX_MESSAGE_SIZE;
     }
-    if (configuredMaxMessageSize <= 0) {
-      throw new NumberFormatException("Max message size must be positive: " + configuredMaxMessageSize);
+    if (maxMessageSize <= 0) {
+      throw new NumberFormatException("Max message size must be positive: " + maxMessageSize);
     }
 
     TConfiguration config = t.getConfiguration();
diff --git a/parquet-hadoop/src/test/java/org/apache/parquet/hadoop/TestParquetFileReaderMaxMessageSize.java b/parquet-hadoop/src/test/java/org/apache/parquet/hadoop/TestParquetFileReaderMaxMessageSize.java
index f9f121b998..7b6088e58a 100644
--- a/parquet-hadoop/src/test/java/org/apache/parquet/hadoop/TestParquetFileReaderMaxMessageSize.java
+++ b/parquet-hadoop/src/test/java/org/apache/parquet/hadoop/TestParquetFileReaderMaxMessageSize.java
@@ -143,4 +143,52 @@ public void testInsufficientMaxMessageSizeError() throws IOException {
               || e.getCause().getMessage().contains("MaxMessageSize reached"));
     }
   }
+
+  /**
+   * The -1 sentinel must be honored as "use the default max message size",
+   * not rejected as a non-positive value.
+   */
+  @Test
+  public void testReadAcceptsMinusOneAsDefaultMaxMessageSize() throws IOException {
+    Configuration readConf = new Configuration();
+    readConf.setInt("parquet.thrift.string.size.limit", -1);
+    ParquetReadOptions options = HadoopReadOptions.builder(readConf).build();
+
+    try (ParquetFileReader reader =
+        ParquetFileReader.open(HadoopInputFile.fromPath(TEST_FILE, readConf), options)) {
+      ParquetMetadata metadata = reader.getFooter();
+      assertNotNull(metadata);
+      assertEquals(schema, metadata.getFileMetaData().getSchema());
+    }
+  }
+
+  @Test
+  public void testReadRejectsZeroMaxMessageSize() throws IOException {
+    assertRejectsNonPositiveMaxMessageSize(0);
+  }
+
+  @Test
+  public void testReadRejectsInvalidNegativeMaxMessageSize() throws IOException {
+    assertRejectsNonPositiveMaxMessageSize(-5);
+  }
+
+  private void assertRejectsNonPositiveMaxMessageSize(int maxMessageSize) throws IOException {
+    Configuration readConf = new Configuration();
+    readConf.setInt("parquet.thrift.string.size.limit", maxMessageSize);
+    ParquetReadOptions options = HadoopReadOptions.builder(readConf).build();
+
+    try (ParquetFileReader reader =
+        ParquetFileReader.open(HadoopInputFile.fromPath(TEST_FILE, readConf), options)) {
+      fail("Expected failure for non-positive max message size: " + maxMessageSize);
+    } catch (RuntimeException | IOException e) {
+      String message = e.getMessage() == null ? "" : e.getMessage();
+      String causeMessage = e.getCause() == null || e.getCause().getMessage() == null
+          ? ""
+          : e.getCause().getMessage();
+      assertTrue(
+          "Expected 'Max message size must be positive' in " + message + " / " + causeMessage,
+          message.contains("Max message size must be positive")
+              || causeMessage.contains("Max message size must be positive"));
+    }
+  }
 }