Move components/compression to third_party/zlib/google

We are now implementing CompressibleString in third_party/WebKit/Source/
wtf/text and want to use compression algorithm, but wtf/text can't
depend on components. This CL moves components/compression to
third_party and enables wtf/text to use the compression utility.

The purpose of this CL is similar to crrev.com/14021015.

BUG=574317
TEST=n/a

Review URL: https://codereview.chromium.org/1564773002

Cr-Original-Commit-Position: refs/heads/master@{#369491}
Cr-Mirrored-From: https://chromium.googlesource.com/chromium/src
Cr-Mirrored-Commit: 1e1cc1a7060ab8eb95a62da608f333e31aaba475

BUILD.gn

@@ -149,3 +149,13 @@ static_library("zip") {
     "//base",
   ]
 }
+
+static_library("compression_utils") {
+  sources = [
+    "google/compression_utils.cc",
+    "google/compression_utils.h",
+  ]
+  deps = [
+    ":zlib",
+  ]
+}

google/OWNERS

 hshi@chromium.org
 satorux@chromium.org
+
+# compression_utils*
+asvitkine@chromium.org
+isherman@chromium.org

google/compression_utils.cc

+// Copyright 2014 The Chromium Authors. All rights reserved.
+// Use of this source code is governed by a BSD-style license that can be
+// found in the LICENSE file.
+
+#include "third_party/zlib/google/compression_utils.h"
+
+#include <stddef.h>
+#include <stdint.h>
+#include <string.h>
+
+#include <vector>
+
+#include "base/bit_cast.h"
+#include "base/logging.h"
+#include "base/sys_byteorder.h"
+#include "third_party/zlib/zlib.h"
+
+namespace {
+
+// The difference in bytes between a zlib header and a gzip header.
+const size_t kGzipZlibHeaderDifferenceBytes = 16;
+
+// Pass an integer greater than the following get a gzip header instead of a
+// zlib header when calling deflateInit2() and inflateInit2().
+const int kWindowBitsToGetGzipHeader = 16;
+
+// This describes the amount of memory zlib uses to compress data. It can go
+// from 1 to 9, with 8 being the default. For details, see:
+// http://www.zlib.net/manual.html (search for memLevel).
+const int kZlibMemoryLevel = 8;
+
+// This code is taken almost verbatim from third_party/zlib/compress.c. The only
+// difference is deflateInit2() is called which sets the window bits to be > 16.
+// That causes a gzip header to be emitted rather than a zlib header.
+int GzipCompressHelper(Bytef* dest,
+                       uLongf* dest_length,
+                       const Bytef* source,
+                       uLong source_length) {
+  z_stream stream;
+
+  stream.next_in = bit_cast<Bytef*>(source);
+  stream.avail_in = static_cast<uInt>(source_length);
+  stream.next_out = dest;
+  stream.avail_out = static_cast<uInt>(*dest_length);
+  if (static_cast<uLong>(stream.avail_out) != *dest_length)
+    return Z_BUF_ERROR;
+
+  stream.zalloc = static_cast<alloc_func>(0);
+  stream.zfree = static_cast<free_func>(0);
+  stream.opaque = static_cast<voidpf>(0);
+
+  gz_header gzip_header;
+  memset(&gzip_header, 0, sizeof(gzip_header));
+  int err = deflateInit2(&stream,
+                         Z_DEFAULT_COMPRESSION,
+                         Z_DEFLATED,
+                         MAX_WBITS + kWindowBitsToGetGzipHeader,
+                         kZlibMemoryLevel,
+                         Z_DEFAULT_STRATEGY);
+  if (err != Z_OK)
+    return err;
+
+  err = deflateSetHeader(&stream, &gzip_header);
+  if (err != Z_OK)
+    return err;
+
+  err = deflate(&stream, Z_FINISH);
+  if (err != Z_STREAM_END) {
+    deflateEnd(&stream);
+    return err == Z_OK ? Z_BUF_ERROR : err;
+  }
+  *dest_length = stream.total_out;
+
+  err = deflateEnd(&stream);
+  return err;
+}
+
+// This code is taken almost verbatim from third_party/zlib/uncompr.c. The only
+// difference is inflateInit2() is called which sets the window bits to be > 16.
+// That causes a gzip header to be parsed rather than a zlib header.
+int GzipUncompressHelper(Bytef* dest,
+                         uLongf* dest_length,
+                         const Bytef* source,
+                         uLong source_length) {
+  z_stream stream;
+
+  stream.next_in = bit_cast<Bytef*>(source);
+  stream.avail_in = static_cast<uInt>(source_length);
+  if (static_cast<uLong>(stream.avail_in) != source_length)
+    return Z_BUF_ERROR;
+
+  stream.next_out = dest;
+  stream.avail_out = static_cast<uInt>(*dest_length);
+  if (static_cast<uLong>(stream.avail_out) != *dest_length)
+    return Z_BUF_ERROR;
+
+  stream.zalloc = static_cast<alloc_func>(0);
+  stream.zfree = static_cast<free_func>(0);
+
+  int err = inflateInit2(&stream, MAX_WBITS + kWindowBitsToGetGzipHeader);
+  if (err != Z_OK)
+    return err;
+
+  err = inflate(&stream, Z_FINISH);
+  if (err != Z_STREAM_END) {
+    inflateEnd(&stream);
+    if (err == Z_NEED_DICT || (err == Z_BUF_ERROR && stream.avail_in == 0))
+      return Z_DATA_ERROR;
+    return err;
+  }
+  *dest_length = stream.total_out;
+
+  err = inflateEnd(&stream);
+  return err;
+}
+
+// Returns the uncompressed size from GZIP-compressed |compressed_data|.
+uint32_t GetUncompressedSize(const std::string& compressed_data) {
+  // The uncompressed size is stored in the last 4 bytes of |input| in LE.
+  uint32_t size;
+  if (compressed_data.length() < sizeof(size))
+    return 0;
+  memcpy(&size, &compressed_data[compressed_data.length() - sizeof(size)],
+         sizeof(size));
+  return base::ByteSwapToLE32(size);
+}
+
+}  // namespace
+
+namespace compression {
+
+bool GzipCompress(const std::string& input, std::string* output) {
+  const uLongf input_size = static_cast<uLongf>(input.size());
+  std::vector<Bytef> compressed_data(kGzipZlibHeaderDifferenceBytes +
+                                     compressBound(input_size));
+
+  uLongf compressed_size = static_cast<uLongf>(compressed_data.size());
+  if (GzipCompressHelper(&compressed_data.front(),
+                         &compressed_size,
+                         bit_cast<const Bytef*>(input.data()),
+                         input_size) != Z_OK) {
+    return false;
+  }
+
+  compressed_data.resize(compressed_size);
+  output->assign(compressed_data.begin(), compressed_data.end());
+  DCHECK_EQ(input_size, GetUncompressedSize(*output));
+  return true;
+}
+
+bool GzipUncompress(const std::string& input, std::string* output) {
+  std::string uncompressed_output;
+  uLongf uncompressed_size = static_cast<uLongf>(GetUncompressedSize(input));
+  uncompressed_output.resize(uncompressed_size);
+  if (GzipUncompressHelper(bit_cast<Bytef*>(uncompressed_output.data()),
+                           &uncompressed_size,
+                           bit_cast<const Bytef*>(input.data()),
+                           static_cast<uLongf>(input.length())) == Z_OK) {
+    output->swap(uncompressed_output);
+    return true;
+  }
+  return false;
+}
+
+}  // namespace compression

google/compression_utils.h

+// Copyright 2014 The Chromium Authors. All rights reserved.
+// Use of this source code is governed by a BSD-style license that can be
+// found in the LICENSE file.
+
+#ifndef THIRD_PARTY_ZLIB_GOOGLE_COMPRESSION_UTILS_H_
+#define THIRD_PARTY_ZLIB_GOOGLE_COMPRESSION_UTILS_H_
+
+#include <string>
+
+namespace compression {
+
+// Compresses the data in |input| using gzip, storing the result in |output|.
+// |input| and |output| are allowed to be the same string (in-place operation).
+bool GzipCompress(const std::string& input, std::string* output);
+
+// Uncompresses the data in |input| using gzip, storing the result in |output|.
+// |input| and |output| are allowed to be the same string (in-place operation).
+bool GzipUncompress(const std::string& input, std::string* output);
+
+}  // namespace compression
+
+#endif  // THIRD_PARTY_ZLIB_GOOGLE_COMPRESSION_UTILS_H_

google/compression_utils_unittest.cc

+// Copyright 2014 The Chromium Authors. All rights reserved.
+// Use of this source code is governed by a BSD-style license that can be
+// found in the LICENSE file.
+
+#include "third_party/zlib/google/compression_utils.h"
+
+#include <stddef.h>
+#include <stdint.h>
+
+#include <string>
+
+#include "base/logging.h"
+#include "base/macros.h"
+#include "testing/gtest/include/gtest/gtest.h"
+
+namespace compression {
+
+namespace {
+
+// The data to be compressed by gzip. This is the hex representation of "hello
+// world".
+const uint8_t kData[] = {0x68, 0x65, 0x6c, 0x6c, 0x6f, 0x20,
+                         0x77, 0x6f, 0x72, 0x6c, 0x64};
+
+// This is the string representation of gzip compressed string above. It was
+// obtained by running echo -n "hello world" | gzip -c | hexdump -e '8 1 ",
+// 0x%x"' followed by 0'ing out the OS byte (10th byte) in the header. This is
+// so that the test passes on all platforms (that run various OS'es).
+const uint8_t kCompressedData[] = {
+    0x1f, 0x8b, 0x08, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0xcb,
+    0x48, 0xcd, 0xc9, 0xc9, 0x57, 0x28, 0xcf, 0x2f, 0xca, 0x49, 0x01,
+    0x00, 0x85, 0x11, 0x4a, 0x0d, 0x0b, 0x00, 0x00, 0x00};
+
+}  // namespace
+
+TEST(CompressionUtilsTest, GzipCompression) {
+  std::string data(reinterpret_cast<const char*>(kData), arraysize(kData));
+  std::string compressed_data;
+  EXPECT_TRUE(GzipCompress(data, &compressed_data));
+  std::string golden_compressed_data(
+      reinterpret_cast<const char*>(kCompressedData),
+      arraysize(kCompressedData));
+  EXPECT_EQ(golden_compressed_data, compressed_data);
+}
+
+TEST(CompressionUtilsTest, GzipUncompression) {
+  std::string compressed_data(reinterpret_cast<const char*>(kCompressedData),
+                              arraysize(kCompressedData));
+
+  std::string uncompressed_data;
+  EXPECT_TRUE(GzipUncompress(compressed_data, &uncompressed_data));
+
+  std::string golden_data(reinterpret_cast<const char*>(kData),
+                          arraysize(kData));
+  EXPECT_EQ(golden_data, uncompressed_data);
+}
+
+// Checks that compressing/decompressing input > 256 bytes works as expected.
+TEST(CompressionUtilsTest, LargeInput) {
+  const size_t kSize = 32 * 1024;
+
+  // Generate a data string of |kSize| for testing.
+  std::string data;
+  data.resize(kSize);
+  for (size_t i = 0; i < kSize; ++i)
+    data[i] = static_cast<char>(i & 0xFF);
+
+  std::string compressed_data;
+  EXPECT_TRUE(GzipCompress(data, &compressed_data));
+
+  std::string uncompressed_data;
+  EXPECT_TRUE(GzipUncompress(compressed_data, &uncompressed_data));
+
+  EXPECT_EQ(data, uncompressed_data);
+}
+
+TEST(CompressionUtilsTest, InPlace) {
+  const std::string original_data(reinterpret_cast<const char*>(kData),
+                                  arraysize(kData));
+  const std::string golden_compressed_data(
+      reinterpret_cast<const char*>(kCompressedData),
+      arraysize(kCompressedData));
+
+  std::string data(original_data);
+  EXPECT_TRUE(GzipCompress(data, &data));
+  EXPECT_EQ(golden_compressed_data, data);
+  EXPECT_TRUE(GzipUncompress(data, &data));
+  EXPECT_EQ(original_data, data);
+}
+
+}  // namespace compression

google/zip.gyp

@@ -23,5 +23,19 @@
         'zip_reader.h',
       ],
     },
+    {
+      'target_name': 'compression_utils',
+      'type': 'static_library',
+      'dependencies': [
+        '../zlib.gyp:zlib',
+      ],
+      'include_dirs': [
+        '../../..',
+      ],
+      'sources': [
+        'compression_utils.cc',
+        'compression_utils.h',
+      ],
+    },
   ],
 }