Proxy archive interfaces through buffered input

Internal bazel use of compressed archives, exemplified through
repository inspection, is utilizing a single-byte read syscall over
all, possibly extremely large, archives, with substantial syscall
roundtrip penalties. Time spent fetching all archives for a workspace
was observed to be <50% of the time without the fix (performing single-
byte reads).

Fixes #3397

Closes #3870.

PiperOrigin-RevId: 179921752

3 files changed, 16 insertions, 10 deletions

diff --git a/src/main/java/com/google/devtools/build/lib/bazel/repository/TarBz2Function.java b/src/main/java/com/google/devtools/build/lib/bazel/repository/TarBz2Function.java
index 8566afd3b4..8d25d7d0d6 100644
--- a/src/main/java/com/google/devtools/build/lib/bazel/repository/TarBz2Function.java
+++ b/src/main/java/com/google/devtools/build/lib/bazel/repository/TarBz2Function.java
@@ -15,18 +15,18 @@
 package com.google.devtools.build.lib.bazel.repository;
 
 import com.google.devtools.build.lib.bazel.repository.DecompressorValue.Decompressor;
-
-import org.apache.commons.compress.compressors.bzip2.BZip2CompressorInputStream;
-
+import java.io.BufferedInputStream;
 import java.io.FileInputStream;
 import java.io.IOException;
 import java.io.InputStream;
+import org.apache.commons.compress.compressors.bzip2.BZip2CompressorInputStream;
 
 /**
  * Creates a repository by unarchiving a .tar.bz2 file.
  */
 public class TarBz2Function extends CompressedTarFunction {
   public static final Decompressor INSTANCE = new TarBz2Function();
+  private static final int BUFFER_SIZE = 32 * 1024;
 
   private TarBz2Function() {
   }
@@ -35,6 +35,7 @@ public class TarBz2Function extends CompressedTarFunction {
   protected InputStream getDecompressorStream(DecompressorDescriptor descriptor)
       throws IOException {
     return new BZip2CompressorInputStream(
-        new FileInputStream(descriptor.archivePath().getPathFile()));
+        new BufferedInputStream(
+            new FileInputStream(descriptor.archivePath().getPathFile()), BUFFER_SIZE));
   }
 }
diff --git a/src/main/java/com/google/devtools/build/lib/bazel/repository/TarGzFunction.java b/src/main/java/com/google/devtools/build/lib/bazel/repository/TarGzFunction.java
index 4a52684efa..e1e1204d7d 100644
--- a/src/main/java/com/google/devtools/build/lib/bazel/repository/TarGzFunction.java
+++ b/src/main/java/com/google/devtools/build/lib/bazel/repository/TarGzFunction.java
@@ -15,7 +15,7 @@
 package com.google.devtools.build.lib.bazel.repository;
 
 import com.google.devtools.build.lib.bazel.repository.DecompressorValue.Decompressor;
-
+import java.io.BufferedInputStream;
 import java.io.FileInputStream;
 import java.io.IOException;
 import java.io.InputStream;
@@ -26,6 +26,7 @@ import java.util.zip.GZIPInputStream;
  */
 public class TarGzFunction extends CompressedTarFunction {
   public static final Decompressor INSTANCE = new TarGzFunction();
+  private static final int BUFFER_SIZE = 32 * 1024;
 
   private TarGzFunction() {
   }
@@ -33,6 +34,8 @@ public class TarGzFunction extends CompressedTarFunction {
   @Override
   protected InputStream getDecompressorStream(DecompressorDescriptor descriptor)
       throws IOException {
-    return new GZIPInputStream(new FileInputStream(descriptor.archivePath().getPathFile()));
+    return new GZIPInputStream(
+        new BufferedInputStream(
+            new FileInputStream(descriptor.archivePath().getPathFile()), BUFFER_SIZE));
   }
 }
diff --git a/src/main/java/com/google/devtools/build/lib/bazel/repository/TarXzFunction.java b/src/main/java/com/google/devtools/build/lib/bazel/repository/TarXzFunction.java
index 7e0b014302..568f85864f 100644
--- a/src/main/java/com/google/devtools/build/lib/bazel/repository/TarXzFunction.java
+++ b/src/main/java/com/google/devtools/build/lib/bazel/repository/TarXzFunction.java
@@ -15,18 +15,18 @@
 package com.google.devtools.build.lib.bazel.repository;
 
 import com.google.devtools.build.lib.bazel.repository.DecompressorValue.Decompressor;
-
-import org.tukaani.xz.XZInputStream;
-
+import java.io.BufferedInputStream;
 import java.io.FileInputStream;
 import java.io.IOException;
 import java.io.InputStream;
+import org.tukaani.xz.XZInputStream;
 
 /**
  * Creates a repository by unarchiving a .tar.xz file.
  */
 class TarXzFunction extends CompressedTarFunction {
   public static final Decompressor INSTANCE = new TarXzFunction();
+  private static final int BUFFER_SIZE = 32 * 1024;
 
   private TarXzFunction() {
   }
@@ -34,6 +34,8 @@ class TarXzFunction extends CompressedTarFunction {
   @Override
   protected InputStream getDecompressorStream(DecompressorDescriptor descriptor)
       throws IOException {
-    return new XZInputStream(new FileInputStream(descriptor.archivePath().getPathFile()));
+    return new XZInputStream(
+        new BufferedInputStream(
+            new FileInputStream(descriptor.archivePath().getPathFile()), BUFFER_SIZE));
   }
 }