[6.5.0] Fix extraction of tar archives containing sparse files. (#20531)

By upgrading the Apache Commons Compress library to 1.20. I'm
deliberately not upgrading to the most recent one (1.24.0) because it
would require an additional JDK module (java.desktop) and significantly
regress the binary size.

Fixes #20090.

Closes #20110.

PiperOrigin-RevId: 580935354
Change-Id: I6c9728ac3fd925432f44a55efaef8f5b52d428c0

Co-authored-by: Tiago Quelhas <tjgq@google.com>

src/test/py/bazel/bazel_external_repository_test.py

@@ -50,8 +50,10 @@ def StopHttpServer(self):
   def setUp(self):
     test_base.TestBase.setUp(self)
     for f in [
-        'six-1.10.0.tar.gz', 'archive_with_symlink.zip',
-        'archive_with_symlink.tar.gz'
+        'six-1.10.0.tar.gz',
+        'archive_with_symlink.zip',
+        'archive_with_symlink.tar.gz',
+        'sparse_archive.tar',
     ]:
       self.CopyFile(self.Rlocation('io_bazel/src/test/py/bazel/testdata/'
                                    'bazel_external_repository_test/' + f), f)
@@ -167,6 +169,63 @@ def testNewHttpTarArchiveWithSymlinks(self):
     ])
     self.assertEqual(exit_code, 0, os.linesep.join(stderr))
 
+  def testNewHttpTarWithSparseFile(self):
+    # Test extraction of tar archives containing sparse files.
+    # The archive under test was produced using GNU tar on Linux:
+    #   truncate -s 1M sparse_file
+    #   tar -c --sparse --format posix -f sparse_archive.tar sparse_file
+
+    ip, port = self._http_server.server_address
+    rule_definition = [
+        'load("@bazel_tools//tools/build_defs/repo:http.bzl", "http_archive")',
+        'http_archive(',
+        '    name = "sparse_archive",',
+        '    urls = ["http://%s:%s/sparse_archive.tar"],' % (ip, port),
+        '    build_file = "@//:sparse_archive.BUILD",',
+        '    sha256 = ',
+        (
+            '  "a1a2b2ce4acd51a8cc1ab80adce6f134ac73e885219911a960a42000e312bb65",'
+            ')'
+        ),
+    ]
+    rule_definition.extend(self.GetDefaultRepoRules())
+    self.ScratchFile('WORKSPACE', rule_definition)
+    self.ScratchFile(
+        'sparse_archive.BUILD',
+        [
+            'exports_files(["sparse_file"])',
+        ],
+    )
+    self.ScratchFile(
+        'test.py',
+        [
+            'import sys',
+            'from tools.python.runfiles import runfiles',
+            'path = runfiles.Create().Rlocation(sys.argv[1])',
+            'if open(path, "rb").read() != b"\\0"*1024*1024:',
+            '  sys.exit(1)',
+        ],
+    )
+    self.ScratchFile(
+        'BUILD',
+        [
+            'py_test(',
+            '    name = "test",',
+            '    srcs = ["test.py"],',
+            (
+                '    args = ["$(rlocationpath @sparse_archive//:sparse_file)"],'
+                '    data = ['
+            ),
+            '        "@sparse_archive//:sparse_file",',
+            '        "@bazel_tools//tools/python/runfiles",',
+            '    ],)',
+        ],
+    )
+    self.RunBazel([
+        'test',
+        '//:test',
+    ])
+
   def _CreatePyWritingStarlarkRule(self, print_string):
     self.ScratchFile('repo/foo.bzl', [
         'def _impl(ctx):',

third_party/BUILD

@@ -140,7 +140,7 @@ distrib_java_import(
 distrib_java_import(
     name = "apache_commons_compress",
     enable_distributions = ["debian"],
-    jars = ["apache_commons_compress/apache-commons-compress-1.19.jar"],
+    jars = ["apache_commons_compress/apache-commons-compress-1.20.jar"],
 )
 
 distrib_java_import(