From 636d50caa0a2d4ac35d528f75e2ac51a99a4288a Mon Sep 17 00:00:00 2001
From: Nikita Keba <nikita.keba@gmail.com>
Date: Mon, 29 May 2023 19:35:24 +0000
Subject: [PATCH] fix cmake + add unit tests for TarArchiveReader

---
 contrib/libarchive-cmake/CMakeLists.txt       |  2 +-
 contrib/libarchive-cmake/config.h             |  2 +-
 src/IO/Archives/util/tar_archive_writer.h     | 27 ++++++++++
 .../tests/gtest_archive_reader_and_writer.cpp | 50 ++++++++++++++++++-
 4 files changed, 78 insertions(+), 3 deletions(-)
 create mode 100644 src/IO/Archives/util/tar_archive_writer.h

diff --git a/contrib/libarchive-cmake/CMakeLists.txt b/contrib/libarchive-cmake/CMakeLists.txt
index 0bf53e737db..4593f7f96c8 100644
--- a/contrib/libarchive-cmake/CMakeLists.txt
+++ b/contrib/libarchive-cmake/CMakeLists.txt
@@ -127,7 +127,7 @@ set(SRCS
     "${LIBRARY_DIR}/libarchive/archive_write_set_format_ustar.c"
     "${LIBRARY_DIR}/libarchive/archive_write_set_format_v7tar.c"
     "${LIBRARY_DIR}/libarchive/archive_write_set_format_warc.c"
-    # "${LIBRARY_DIR}/libarchive/archive_write_set_format_xar.c"
+    "${LIBRARY_DIR}/libarchive/archive_write_set_format_xar.c"
     "${LIBRARY_DIR}/libarchive/archive_write_set_format_zip.c"
     "${LIBRARY_DIR}/libarchive/archive_write_set_options.c"
     "${LIBRARY_DIR}/libarchive/archive_write_set_passphrase.c"
diff --git a/contrib/libarchive-cmake/config.h b/contrib/libarchive-cmake/config.h
index 0b54bf63ddd..2fa1bb8945d 100644
--- a/contrib/libarchive-cmake/config.h
+++ b/contrib/libarchive-cmake/config.h
@@ -730,7 +730,7 @@ typedef uint64_t uintmax_t;
 #define HAVE_LIBXML_XMLREADER_H 0
 
 /* Define to 1 if you have the <libxml/xmlwriter.h> header file. */
-#define HAVE_LIBXML_XMLWRITER_H 1
+// #define HAVE_LIBXML_XMLWRITER_H 1
 
 /* Define to 1 if you have the `z' library (-lz). */
 /* #undef HAVE_LIBZ */
diff --git a/src/IO/Archives/util/tar_archive_writer.h b/src/IO/Archives/util/tar_archive_writer.h
new file mode 100644
index 00000000000..138bcb036a6
--- /dev/null
+++ b/src/IO/Archives/util/tar_archive_writer.h
@@ -0,0 +1,27 @@
+#include <archive.h>
+#include <archive_entry.h>
+
+bool create_tar_with_file(const std::string &archivename, std::map<std::string, std::string> files) {
+    struct archive *a;
+    struct archive_entry *entry;
+
+    a = archive_write_new();
+    archive_write_set_format_pax_restricted(a);
+    archive_write_open_filename(a, archivename.c_str());
+
+    for (auto &[filename, content] : files) {
+        entry = archive_entry_new();
+        archive_entry_set_pathname(entry, filename.c_str());
+        archive_entry_set_size(entry, content.size());
+        archive_entry_set_mode(entry, S_IFREG | 0644); // regular file with rw-r--r-- permissions
+        archive_entry_set_mtime(entry, time(nullptr), 0);
+        archive_write_header(a, entry);
+        archive_write_data(a, content.c_str(), content.size());
+        archive_entry_free(entry);
+    }
+    
+    archive_write_close(a);
+    archive_write_free(a);
+
+    return true;
+}
diff --git a/src/IO/tests/gtest_archive_reader_and_writer.cpp b/src/IO/tests/gtest_archive_reader_and_writer.cpp
index 3bc9d670f05..186fdb89532 100644
--- a/src/IO/tests/gtest_archive_reader_and_writer.cpp
+++ b/src/IO/tests/gtest_archive_reader_and_writer.cpp
@@ -5,6 +5,7 @@
 #include <IO/Archives/IArchiveWriter.h>
 #include <IO/Archives/createArchiveReader.h>
 #include <IO/Archives/createArchiveWriter.h>
+#include <IO/Archives/util/tar_archive_writer.h>
 #include <IO/ReadBufferFromFileBase.h>
 #include <IO/ReadBufferFromString.h>
 #include <IO/ReadHelpers.h>
@@ -327,6 +328,53 @@ TEST_P(ArchiveReaderAndWriterTest, ArchiveNotExist)
                     [&]{ createArchiveReader(getPathToArchive()); });
 }
 
+TEST(TarArchiveReaderTest, FileExists) {
+    String archive_path = "archive.tar";
+    String filename = "file.txt";
+    String contents = "test";
+    bool created = create_tar_with_file(archive_path, {{filename, contents}});
+    EXPECT_EQ(created, true);
+    auto reader = createArchiveReader(archive_path);
+    EXPECT_EQ(reader->fileExists(filename), true);
+    fs::remove(archive_path);
+}
+
+TEST(TarArchiveReaderTest, ReadFile) {
+    String archive_path = "archive.tar";
+    String filename = "file.txt";
+    String contents = "test";
+    bool created = create_tar_with_file(archive_path, {{filename, contents}});
+    EXPECT_EQ(created, true);
+    auto reader = createArchiveReader(archive_path);
+    auto in = reader->readFile(filename);
+    String str;
+    readStringUntilEOF(str, *in);
+    EXPECT_EQ(str, contents);
+    fs::remove(archive_path);
+}
+
+TEST(TarArchiveReaderTest, ReadTwoFiles) {
+    String archive_path = "archive.tar";
+    String file1 = "file1.txt";
+    String contents1 = "test1";
+    String file2 = "file2.txt";
+    String contents2 = "test2";
+    bool created = create_tar_with_file(archive_path, {{file1, contents1}, {file2, contents2}});
+    EXPECT_EQ(created, true);
+    auto reader = createArchiveReader(archive_path);
+    EXPECT_EQ(reader->fileExists(file1), true);
+    EXPECT_EQ(reader->fileExists(file2), true);
+    auto in = reader->readFile(file1);
+    String str;
+    readStringUntilEOF(str, *in);
+    EXPECT_EQ(str, contents1);
+    in = reader->readFile(file2);
+    
+    readStringUntilEOF(str, *in);
+    EXPECT_EQ(str, contents2);
+    fs::remove(archive_path);
+}
+
 
 #if USE_MINIZIP
 
@@ -334,7 +382,7 @@ namespace
 {
     const char * supported_archive_file_exts[] =
     {
-        ".zip",
+        ".zip"
     };
 }