Parse deflates in gzip files.

The diff for NOTICE.xml.gz is reduced from 200+K to 17K.

Also we are now honoring puffin::LocateDeflatesInGzip() failures,
since if won't complain about padding data anymore, and
puffin::LocateDeflateSubBlocksInZipArchive() is updated to
puffin::LocateDeflatesInZipArchive().

Test: generate a delta payload extracting deflates in gzip

Change-Id: I44b816cff8ac99ce78f758eef5875fa99e36e369

3 files changed, 43 insertions, 38 deletions

diff --git a/payload_generator/boot_img_filesystem.cc b/payload_generator/boot_img_filesystem.cc
index 90be9669..19de4106 100644
--- a/payload_generator/boot_img_filesystem.cc
+++ b/payload_generator/boot_img_filesystem.cc
@@ -70,15 +70,10 @@ FilesystemInterface::File BootImgFilesystem::GetFile(const string& name,
     // Check GZip header magic.
     if (data.size() > kGZipHeaderSize && data[0] == 0x1F && data[1] == 0x8B) {
       if (!puffin::LocateDeflatesInGzip(data, &file.deflates)) {
-        // We still use the deflates found even if LocateDeflatesInGzip() fails,
-        // if any deflates are returned, they should be correct, it's possible
-        // something went wrong later but it shouldn't stop us from using the
-        // previous deflates. Another common case is if there's more data after
-        // the gzip, the function will try to parse that as another gzip and
-        // will fail, but we still want the deflates from the first gzip.
-        LOG(WARNING) << "Error occurred parsing gzip " << name << " at offset "
-                     << offset << " of " << filename_ << ", found "
-                     << file.deflates.size() << " deflates.";
+        LOG(ERROR) << "Error occurred parsing gzip " << name << " at offset "
+                   << offset << " of " << filename_ << ", found "
+                   << file.deflates.size() << " deflates.";
+        return file;
       }
       for (auto& deflate : file.deflates) {
         deflate.offset += offset * 8;
diff --git a/payload_generator/deflate_utils.cc b/payload_generator/deflate_utils.cc
index 2719048c..6f474888 100644
--- a/payload_generator/deflate_utils.cc
+++ b/payload_generator/deflate_utils.cc
@@ -102,6 +102,15 @@ bool IsBitExtentInExtent(const Extent& extent, const BitExtent& bit_extent) {
              ((extent.start_block() + extent.num_blocks()) * kBlockSize);
 }
 
+// Returns whether the given file |name| has an extension listed in
+// |extensions|.
+bool IsFileExtensions(const string& name,
+                      const std::initializer_list<string>& extensions) {
+  return any_of(extensions.begin(), extensions.end(), [&name](const auto& ext) {
+    return base::EndsWith(name, ext, base::CompareCase::INSENSITIVE_ASCII);
+  });
+}
+
 }  // namespace
 
 ByteExtent ExpandToByteExtent(const BitExtent& extent) {
@@ -286,35 +295,35 @@ bool PreprocessParitionFiles(const PartitionConfig& part,
       }
     }
 
-    // Search for deflates if the file is in zip format.
-    // .zvoice files may eventually move out of rootfs. If that happens, remove
-    // ".zvoice" (crbug.com/782918).
-    const string zip_file_extensions[] = {".apk", ".zip", ".jar", ".zvoice"};
-    bool is_zip =
-        any_of(zip_file_extensions,
-               std::end(zip_file_extensions),
-               [&file](const string& ext) {
-                 return base::EndsWith(
-                     file.name, ext, base::CompareCase::INSENSITIVE_ASCII);
-               });
-
-    if (is_zip && extract_deflates) {
-      brillo::Blob data;
-      TEST_AND_RETURN_FALSE(
-          utils::ReadExtents(part.path,
-                             file.extents,
-                             &data,
-                             kBlockSize * utils::BlocksInExtents(file.extents),
-                             kBlockSize));
-      std::vector<puffin::BitExtent> deflates_sub_blocks;
-      TEST_AND_RETURN_FALSE(puffin::LocateDeflateSubBlocksInZipArchive(
-          data, &deflates_sub_blocks));
-      // Shift the deflate's extent to the offset starting from the beginning
-      // of the current partition; and the delta processor will align the
-      // extents in a continuous buffer later.
-      TEST_AND_RETURN_FALSE(
-          ShiftBitExtentsOverExtents(file.extents, &deflates_sub_blocks));
-      file.deflates = std::move(deflates_sub_blocks);
+    if (extract_deflates) {
+      // Search for deflates if the file is in zip or gzip format.
+      // .zvoice files may eventually move out of rootfs. If that happens,
+      // remove ".zvoice" (crbug.com/782918).
+      bool is_zip =
+          IsFileExtensions(file.name, {".apk", ".zip", ".jar", ".zvoice"});
+      bool is_gzip = IsFileExtensions(file.name, {".gz", ".gzip", ".tgz"});
+      if (is_zip || is_gzip) {
+        brillo::Blob data;
+        TEST_AND_RETURN_FALSE(utils::ReadExtents(
+            part.path,
+            file.extents,
+            &data,
+            kBlockSize * utils::BlocksInExtents(file.extents),
+            kBlockSize));
+        vector<puffin::BitExtent> deflates;
+        if (is_zip) {
+          TEST_AND_RETURN_FALSE(
+              puffin::LocateDeflatesInZipArchive(data, &deflates));
+        } else if (is_gzip) {
+          TEST_AND_RETURN_FALSE(puffin::LocateDeflatesInGzip(data, &deflates));
+        }
+        // Shift the deflate's extent to the offset starting from the beginning
+        // of the current partition; and the delta processor will align the
+        // extents in a continuous buffer later.
+        TEST_AND_RETURN_FALSE(
+            ShiftBitExtentsOverExtents(file.extents, &deflates));
+        file.deflates = std::move(deflates);
+      }
     }
 
     result_files->push_back(file);
diff --git a/payload_generator/deflate_utils.h b/payload_generator/deflate_utils.h
index 798ce253..9b40a474 100644
--- a/payload_generator/deflate_utils.h
+++ b/payload_generator/deflate_utils.h
@@ -29,6 +29,7 @@ namespace deflate_utils {
 // Gets the files from the partition and processes all its files. Processing
 // includes:
 //  - splitting large Squashfs containers into its smaller files.
+//  - extracting deflates in zip and gzip files.
 bool PreprocessParitionFiles(const PartitionConfig& part,
                              std::vector<FilesystemInterface::File>* result,
                              bool extract_deflates);