Update commons compress

wellcomecollection · Oct 7, 2024 · 6bd6974 · 6bd6974
1 parent 36346e9
commit 6bd6974
Show file tree

Hide file tree

Showing 4 changed files with 11 additions and 17 deletions.
diff --git a/bag_unpacker/src/main/scala/weco/storage_service/bag_unpacker/services/Unpacker.scala b/bag_unpacker/src/main/scala/weco/storage_service/bag_unpacker/services/Unpacker.scala
@@ -190,7 +190,7 @@ trait Unpacker[
       case err: IOException
           if err.getMessage == "Error detected parsing the header" =>
         UnpackerUnarchiverError(UnexpectedUnarchiverError(err))
-      case err: IOException if err.getMessage.startsWith("unexpected EOF") =>
+      case err: IOException if err.getMessage.startsWith("unexpected EOF") || err.getMessage.startsWith("Truncated TAR archive") =>
         UnpackerEOFError(err)
       case err => UnpackerUnexpectedError(err)
     }

diff --git a/bag_unpacker/src/main/scala/weco/storage_service/bag_unpacker/storage/Unarchiver.scala b/bag_unpacker/src/main/scala/weco/storage_service/bag_unpacker/storage/Unarchiver.scala
@@ -50,7 +50,7 @@ object Unarchiver {
     } yield iterator
 
   private def createIterator(
-    archiveInputStream: ArchiveInputStream
+    archiveInputStream: ArchiveInputStream[ArchiveEntry]
   ): Iterator[(ArchiveEntry, InputStream)] =
     new Iterator[(ArchiveEntry, InputStream)] {
       private var latest: ArchiveEntry = _
@@ -90,13 +90,13 @@ object Unarchiver {
     }
   private def extract(
     inputStream: InputStream
-  ): Either[UnarchiverError, ArchiveInputStream] =
+  ): Either[UnarchiverError, ArchiveInputStream[ArchiveEntry]] =
     Try {
       // We have to wrap in a BufferedInputStream because this method
       // only takes InputStreams that support the `mark()` method.
       new ArchiveStreamFactory()
         .createArchiveInputStream(new BufferedInputStream(inputStream))
-        .asInstanceOf[ArchiveInputStream]
+        .asInstanceOf[ArchiveInputStream[ArchiveEntry]]
     } match {
       case Success(stream)                => Right(stream)
       case Failure(err: ArchiveException) => Left(ArchiveFormatError(err))

diff --git a/bag_unpacker/src/test/scala/weco/storage_service/bag_unpacker/fixtures/CompressFixture.scala b/bag_unpacker/src/test/scala/weco/storage_service/bag_unpacker/fixtures/CompressFixture.scala
@@ -2,16 +2,9 @@ package weco.storage_service.bag_unpacker.fixtures
 
 import java.io.{File, _}
 import grizzled.slf4j.Logging
-import org.apache.commons.compress.archivers.{
-  ArchiveEntry,
-  ArchiveOutputStream,
-  ArchiveStreamFactory
-}
-import org.apache.commons.compress.compressors.{
-  CompressorOutputStream,
-  CompressorStreamFactory
-}
 import org.apache.commons.io.IOUtils
+import org.apache.commons.compress.archivers.{ArchiveEntry, ArchiveOutputStream, ArchiveStreamFactory}
+import org.apache.commons.compress.compressors.{CompressorOutputStream, CompressorStreamFactory}
 import org.scalatest.matchers.should.Matchers
 import weco.fixtures.TestWith
 import weco.storage_service.generators.StorageRandomGenerators
@@ -146,6 +139,7 @@ trait CompressFixture[BagLocation <: Location, Namespace]
     def finish(): Unit = {
       archiveOutputStream.flush()
       archiveOutputStream.finish()
+      archiveOutputStream.close()
       compressorOutputStream.flush()
       compressorOutputStream.close()
     }
@@ -182,7 +176,7 @@ trait CompressFixture[BagLocation <: Location, Namespace]
       compressorName: String
     )(
       outputStream: OutputStream
-    ): CompressorOutputStream = {
+    ): CompressorOutputStream[BufferedOutputStream] = {
 
       val compressorStreamFactory =
         new CompressorStreamFactory()
@@ -195,13 +189,13 @@ trait CompressFixture[BagLocation <: Location, Namespace]
           compressorName,
           bufferedOutputStream
         )
-    }
+    }.asInstanceOf[CompressorOutputStream[BufferedOutputStream]]
 
     private def packer(
       archiverName: String
     )(
       outputStream: OutputStream
-    ): ArchiveOutputStream = {
+    ): ArchiveOutputStream[ArchiveEntry] = {
 
       val archiveStreamFactory =
         new ArchiveStreamFactory()

diff --git a/project/Dependencies.scala b/project/Dependencies.scala
@@ -83,7 +83,7 @@ object ExternalDependencies {
   lazy val versions = new {
 
     val azure = "12.25.4"
-    val commonsCompress = "1.5"
+    val commonsCompress = "1.27.1"
     val commonsIO = "2.17.0"
     val mockito = "5.13.0"
     val scalatest = "3.2.19"