2014-08-29 14:10:04 -07:00
|
|
|
/*
|
New implementation of AssetManager/ResTable
The multiwindow model and Resources-per-activity
model that came in N puts greater demands on AssetManagers.
They are created whenever window dimensions change, which
can be frequently. There is a need to be able to cheaply
create a new AssetManager for each Activity, which shares
a lot of underlying state.
In order to make the creation of AssetManagers cheap,
we need a new implementation of the native AssetManager
and ResTable to support immutable representations of
APKs. This new data structure/class is ApkAssets.
ApkAssets have the same functionality of an AssetManager, except
that they operate on a single APK, and they do not do any caching.
Once loaded, they are immutable.
ApkAssets will be exposed as a Java object, with its implementation in
native code. The existing Java StringBlock will be owned by ApkAssets,
which means that Strings can be shared across AssetManagers.
ApkAssets can be cached by the ResourcesManager. Creating an AssetManager
requires only a list of ApkAssets and a configuration.
AssetManager2 (named with the suffix '2' for now while transitioning
to the new implementation) caches bags that are accessed.
Since ApkAssets are expected to be kept around longer, they do more validation
of the resource table, which cause slower load times. Measured on an angler-userdebug,
loading the framework assets takes 11ms with ApkAssets, and 2ms with the old
AssetManager implementation.
The tradeoff is that there does not need to be any security checks once an ApkAssets
is loaded, and regular resource retrieval is faster. Measured on an angler-userdebug,
accessing resource (android:string/ok) with many locales takes 18us with AssetManager2,
and 19us with AssetManager (this is per resource, so these add up).
Test: make libandroidfw_tests
Change-Id: Id0e57ee828f17008891fe3741935a9be8830b01d
2016-10-28 16:39:15 -07:00
|
|
|
* Copyright (C) 2016 The Android Open Source Project
|
2014-08-29 14:10:04 -07:00
|
|
|
*
|
|
|
|
* Licensed under the Apache License, Version 2.0 (the "License");
|
|
|
|
* you may not use this file except in compliance with the License.
|
|
|
|
* You may obtain a copy of the License at
|
|
|
|
*
|
|
|
|
* http://www.apache.org/licenses/LICENSE-2.0
|
|
|
|
*
|
|
|
|
* Unless required by applicable law or agreed to in writing, software
|
|
|
|
* distributed under the License is distributed on an "AS IS" BASIS,
|
|
|
|
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
|
|
|
|
* See the License for the specific language governing permissions and
|
|
|
|
* limitations under the License.
|
|
|
|
*/
|
|
|
|
|
New implementation of AssetManager/ResTable
The multiwindow model and Resources-per-activity
model that came in N puts greater demands on AssetManagers.
They are created whenever window dimensions change, which
can be frequently. There is a need to be able to cheaply
create a new AssetManager for each Activity, which shares
a lot of underlying state.
In order to make the creation of AssetManagers cheap,
we need a new implementation of the native AssetManager
and ResTable to support immutable representations of
APKs. This new data structure/class is ApkAssets.
ApkAssets have the same functionality of an AssetManager, except
that they operate on a single APK, and they do not do any caching.
Once loaded, they are immutable.
ApkAssets will be exposed as a Java object, with its implementation in
native code. The existing Java StringBlock will be owned by ApkAssets,
which means that Strings can be shared across AssetManagers.
ApkAssets can be cached by the ResourcesManager. Creating an AssetManager
requires only a list of ApkAssets and a configuration.
AssetManager2 (named with the suffix '2' for now while transitioning
to the new implementation) caches bags that are accessed.
Since ApkAssets are expected to be kept around longer, they do more validation
of the resource table, which cause slower load times. Measured on an angler-userdebug,
loading the framework assets takes 11ms with ApkAssets, and 2ms with the old
AssetManager implementation.
The tradeoff is that there does not need to be any security checks once an ApkAssets
is loaded, and regular resource retrieval is faster. Measured on an angler-userdebug,
accessing resource (android:string/ok) with many locales takes 18us with AssetManager2,
and 19us with AssetManager (this is per resource, so these add up).
Test: make libandroidfw_tests
Change-Id: Id0e57ee828f17008891fe3741935a9be8830b01d
2016-10-28 16:39:15 -07:00
|
|
|
#include "androidfw/AssetManager2.h"
|
2016-11-10 16:43:59 -08:00
|
|
|
|
New implementation of AssetManager/ResTable
The multiwindow model and Resources-per-activity
model that came in N puts greater demands on AssetManagers.
They are created whenever window dimensions change, which
can be frequently. There is a need to be able to cheaply
create a new AssetManager for each Activity, which shares
a lot of underlying state.
In order to make the creation of AssetManagers cheap,
we need a new implementation of the native AssetManager
and ResTable to support immutable representations of
APKs. This new data structure/class is ApkAssets.
ApkAssets have the same functionality of an AssetManager, except
that they operate on a single APK, and they do not do any caching.
Once loaded, they are immutable.
ApkAssets will be exposed as a Java object, with its implementation in
native code. The existing Java StringBlock will be owned by ApkAssets,
which means that Strings can be shared across AssetManagers.
ApkAssets can be cached by the ResourcesManager. Creating an AssetManager
requires only a list of ApkAssets and a configuration.
AssetManager2 (named with the suffix '2' for now while transitioning
to the new implementation) caches bags that are accessed.
Since ApkAssets are expected to be kept around longer, they do more validation
of the resource table, which cause slower load times. Measured on an angler-userdebug,
loading the framework assets takes 11ms with ApkAssets, and 2ms with the old
AssetManager implementation.
The tradeoff is that there does not need to be any security checks once an ApkAssets
is loaded, and regular resource retrieval is faster. Measured on an angler-userdebug,
accessing resource (android:string/ok) with many locales takes 18us with AssetManager2,
and 19us with AssetManager (this is per resource, so these add up).
Test: make libandroidfw_tests
Change-Id: Id0e57ee828f17008891fe3741935a9be8830b01d
2016-10-28 16:39:15 -07:00
|
|
|
#include "android-base/logging.h"
|
2014-08-29 14:10:04 -07:00
|
|
|
|
|
|
|
#include "TestHelpers.h"
|
2017-01-16 19:11:19 -08:00
|
|
|
#include "androidfw/ResourceUtils.h"
|
2016-12-29 16:08:16 -05:00
|
|
|
#include "data/lib_one/R.h"
|
|
|
|
#include "data/libclient/R.h"
|
New implementation of AssetManager/ResTable
The multiwindow model and Resources-per-activity
model that came in N puts greater demands on AssetManagers.
They are created whenever window dimensions change, which
can be frequently. There is a need to be able to cheaply
create a new AssetManager for each Activity, which shares
a lot of underlying state.
In order to make the creation of AssetManagers cheap,
we need a new implementation of the native AssetManager
and ResTable to support immutable representations of
APKs. This new data structure/class is ApkAssets.
ApkAssets have the same functionality of an AssetManager, except
that they operate on a single APK, and they do not do any caching.
Once loaded, they are immutable.
ApkAssets will be exposed as a Java object, with its implementation in
native code. The existing Java StringBlock will be owned by ApkAssets,
which means that Strings can be shared across AssetManagers.
ApkAssets can be cached by the ResourcesManager. Creating an AssetManager
requires only a list of ApkAssets and a configuration.
AssetManager2 (named with the suffix '2' for now while transitioning
to the new implementation) caches bags that are accessed.
Since ApkAssets are expected to be kept around longer, they do more validation
of the resource table, which cause slower load times. Measured on an angler-userdebug,
loading the framework assets takes 11ms with ApkAssets, and 2ms with the old
AssetManager implementation.
The tradeoff is that there does not need to be any security checks once an ApkAssets
is loaded, and regular resource retrieval is faster. Measured on an angler-userdebug,
accessing resource (android:string/ok) with many locales takes 18us with AssetManager2,
and 19us with AssetManager (this is per resource, so these add up).
Test: make libandroidfw_tests
Change-Id: Id0e57ee828f17008891fe3741935a9be8830b01d
2016-10-28 16:39:15 -07:00
|
|
|
#include "data/styles/R.h"
|
2017-09-25 13:10:14 -07:00
|
|
|
#include "data/system/R.h"
|
2014-08-29 14:10:04 -07:00
|
|
|
|
2016-11-10 16:43:59 -08:00
|
|
|
namespace app = com::android::app;
|
2016-12-29 16:08:16 -05:00
|
|
|
namespace lib_one = com::android::lib_one;
|
|
|
|
namespace libclient = com::android::libclient;
|
2014-08-29 14:10:04 -07:00
|
|
|
|
2016-11-10 16:43:59 -08:00
|
|
|
namespace android {
|
2014-08-29 14:10:04 -07:00
|
|
|
|
New implementation of AssetManager/ResTable
The multiwindow model and Resources-per-activity
model that came in N puts greater demands on AssetManagers.
They are created whenever window dimensions change, which
can be frequently. There is a need to be able to cheaply
create a new AssetManager for each Activity, which shares
a lot of underlying state.
In order to make the creation of AssetManagers cheap,
we need a new implementation of the native AssetManager
and ResTable to support immutable representations of
APKs. This new data structure/class is ApkAssets.
ApkAssets have the same functionality of an AssetManager, except
that they operate on a single APK, and they do not do any caching.
Once loaded, they are immutable.
ApkAssets will be exposed as a Java object, with its implementation in
native code. The existing Java StringBlock will be owned by ApkAssets,
which means that Strings can be shared across AssetManagers.
ApkAssets can be cached by the ResourcesManager. Creating an AssetManager
requires only a list of ApkAssets and a configuration.
AssetManager2 (named with the suffix '2' for now while transitioning
to the new implementation) caches bags that are accessed.
Since ApkAssets are expected to be kept around longer, they do more validation
of the resource table, which cause slower load times. Measured on an angler-userdebug,
loading the framework assets takes 11ms with ApkAssets, and 2ms with the old
AssetManager implementation.
The tradeoff is that there does not need to be any security checks once an ApkAssets
is loaded, and regular resource retrieval is faster. Measured on an angler-userdebug,
accessing resource (android:string/ok) with many locales takes 18us with AssetManager2,
and 19us with AssetManager (this is per resource, so these add up).
Test: make libandroidfw_tests
Change-Id: Id0e57ee828f17008891fe3741935a9be8830b01d
2016-10-28 16:39:15 -07:00
|
|
|
class ThemeTest : public ::testing::Test {
|
|
|
|
public:
|
|
|
|
void SetUp() override {
|
2017-09-25 13:10:14 -07:00
|
|
|
system_assets_ = ApkAssets::Load(GetTestDataPath() + "/system/system.apk", true /*system*/);
|
|
|
|
ASSERT_NE(nullptr, system_assets_);
|
|
|
|
|
New implementation of AssetManager/ResTable
The multiwindow model and Resources-per-activity
model that came in N puts greater demands on AssetManagers.
They are created whenever window dimensions change, which
can be frequently. There is a need to be able to cheaply
create a new AssetManager for each Activity, which shares
a lot of underlying state.
In order to make the creation of AssetManagers cheap,
we need a new implementation of the native AssetManager
and ResTable to support immutable representations of
APKs. This new data structure/class is ApkAssets.
ApkAssets have the same functionality of an AssetManager, except
that they operate on a single APK, and they do not do any caching.
Once loaded, they are immutable.
ApkAssets will be exposed as a Java object, with its implementation in
native code. The existing Java StringBlock will be owned by ApkAssets,
which means that Strings can be shared across AssetManagers.
ApkAssets can be cached by the ResourcesManager. Creating an AssetManager
requires only a list of ApkAssets and a configuration.
AssetManager2 (named with the suffix '2' for now while transitioning
to the new implementation) caches bags that are accessed.
Since ApkAssets are expected to be kept around longer, they do more validation
of the resource table, which cause slower load times. Measured on an angler-userdebug,
loading the framework assets takes 11ms with ApkAssets, and 2ms with the old
AssetManager implementation.
The tradeoff is that there does not need to be any security checks once an ApkAssets
is loaded, and regular resource retrieval is faster. Measured on an angler-userdebug,
accessing resource (android:string/ok) with many locales takes 18us with AssetManager2,
and 19us with AssetManager (this is per resource, so these add up).
Test: make libandroidfw_tests
Change-Id: Id0e57ee828f17008891fe3741935a9be8830b01d
2016-10-28 16:39:15 -07:00
|
|
|
style_assets_ = ApkAssets::Load(GetTestDataPath() + "/styles/styles.apk");
|
|
|
|
ASSERT_NE(nullptr, style_assets_);
|
2016-12-29 16:08:16 -05:00
|
|
|
|
|
|
|
libclient_assets_ = ApkAssets::Load(GetTestDataPath() + "/libclient/libclient.apk");
|
|
|
|
ASSERT_NE(nullptr, libclient_assets_);
|
|
|
|
|
|
|
|
lib_one_assets_ = ApkAssets::Load(GetTestDataPath() + "/lib_one/lib_one.apk");
|
|
|
|
ASSERT_NE(nullptr, lib_one_assets_);
|
|
|
|
|
|
|
|
lib_two_assets_ = ApkAssets::Load(GetTestDataPath() + "/lib_two/lib_two.apk");
|
|
|
|
ASSERT_NE(nullptr, lib_two_assets_);
|
New implementation of AssetManager/ResTable
The multiwindow model and Resources-per-activity
model that came in N puts greater demands on AssetManagers.
They are created whenever window dimensions change, which
can be frequently. There is a need to be able to cheaply
create a new AssetManager for each Activity, which shares
a lot of underlying state.
In order to make the creation of AssetManagers cheap,
we need a new implementation of the native AssetManager
and ResTable to support immutable representations of
APKs. This new data structure/class is ApkAssets.
ApkAssets have the same functionality of an AssetManager, except
that they operate on a single APK, and they do not do any caching.
Once loaded, they are immutable.
ApkAssets will be exposed as a Java object, with its implementation in
native code. The existing Java StringBlock will be owned by ApkAssets,
which means that Strings can be shared across AssetManagers.
ApkAssets can be cached by the ResourcesManager. Creating an AssetManager
requires only a list of ApkAssets and a configuration.
AssetManager2 (named with the suffix '2' for now while transitioning
to the new implementation) caches bags that are accessed.
Since ApkAssets are expected to be kept around longer, they do more validation
of the resource table, which cause slower load times. Measured on an angler-userdebug,
loading the framework assets takes 11ms with ApkAssets, and 2ms with the old
AssetManager implementation.
The tradeoff is that there does not need to be any security checks once an ApkAssets
is loaded, and regular resource retrieval is faster. Measured on an angler-userdebug,
accessing resource (android:string/ok) with many locales takes 18us with AssetManager2,
and 19us with AssetManager (this is per resource, so these add up).
Test: make libandroidfw_tests
Change-Id: Id0e57ee828f17008891fe3741935a9be8830b01d
2016-10-28 16:39:15 -07:00
|
|
|
}
|
|
|
|
|
|
|
|
protected:
|
2017-09-25 13:10:14 -07:00
|
|
|
std::unique_ptr<const ApkAssets> system_assets_;
|
2017-01-13 20:47:26 -08:00
|
|
|
std::unique_ptr<const ApkAssets> style_assets_;
|
|
|
|
std::unique_ptr<const ApkAssets> libclient_assets_;
|
|
|
|
std::unique_ptr<const ApkAssets> lib_one_assets_;
|
|
|
|
std::unique_ptr<const ApkAssets> lib_two_assets_;
|
New implementation of AssetManager/ResTable
The multiwindow model and Resources-per-activity
model that came in N puts greater demands on AssetManagers.
They are created whenever window dimensions change, which
can be frequently. There is a need to be able to cheaply
create a new AssetManager for each Activity, which shares
a lot of underlying state.
In order to make the creation of AssetManagers cheap,
we need a new implementation of the native AssetManager
and ResTable to support immutable representations of
APKs. This new data structure/class is ApkAssets.
ApkAssets have the same functionality of an AssetManager, except
that they operate on a single APK, and they do not do any caching.
Once loaded, they are immutable.
ApkAssets will be exposed as a Java object, with its implementation in
native code. The existing Java StringBlock will be owned by ApkAssets,
which means that Strings can be shared across AssetManagers.
ApkAssets can be cached by the ResourcesManager. Creating an AssetManager
requires only a list of ApkAssets and a configuration.
AssetManager2 (named with the suffix '2' for now while transitioning
to the new implementation) caches bags that are accessed.
Since ApkAssets are expected to be kept around longer, they do more validation
of the resource table, which cause slower load times. Measured on an angler-userdebug,
loading the framework assets takes 11ms with ApkAssets, and 2ms with the old
AssetManager implementation.
The tradeoff is that there does not need to be any security checks once an ApkAssets
is loaded, and regular resource retrieval is faster. Measured on an angler-userdebug,
accessing resource (android:string/ok) with many locales takes 18us with AssetManager2,
and 19us with AssetManager (this is per resource, so these add up).
Test: make libandroidfw_tests
Change-Id: Id0e57ee828f17008891fe3741935a9be8830b01d
2016-10-28 16:39:15 -07:00
|
|
|
};
|
|
|
|
|
|
|
|
TEST_F(ThemeTest, EmptyTheme) {
|
|
|
|
AssetManager2 assetmanager;
|
|
|
|
assetmanager.SetApkAssets({style_assets_.get()});
|
|
|
|
|
|
|
|
std::unique_ptr<Theme> theme = assetmanager.NewTheme();
|
|
|
|
EXPECT_EQ(0u, theme->GetChangingConfigurations());
|
|
|
|
EXPECT_EQ(&assetmanager, theme->GetAssetManager());
|
|
|
|
|
|
|
|
Res_value value;
|
|
|
|
uint32_t flags;
|
|
|
|
EXPECT_EQ(kInvalidCookie, theme->GetAttribute(app::R::attr::attr_one, &value, &flags));
|
|
|
|
}
|
|
|
|
|
|
|
|
TEST_F(ThemeTest, SingleThemeNoParent) {
|
|
|
|
AssetManager2 assetmanager;
|
|
|
|
assetmanager.SetApkAssets({style_assets_.get()});
|
|
|
|
|
|
|
|
std::unique_ptr<Theme> theme = assetmanager.NewTheme();
|
|
|
|
ASSERT_TRUE(theme->ApplyStyle(app::R::style::StyleOne));
|
|
|
|
|
|
|
|
Res_value value;
|
|
|
|
uint32_t flags;
|
|
|
|
ApkAssetsCookie cookie;
|
|
|
|
|
|
|
|
cookie = theme->GetAttribute(app::R::attr::attr_one, &value, &flags);
|
|
|
|
ASSERT_NE(kInvalidCookie, cookie);
|
|
|
|
EXPECT_EQ(Res_value::TYPE_INT_DEC, value.dataType);
|
|
|
|
EXPECT_EQ(1u, value.data);
|
|
|
|
EXPECT_EQ(static_cast<uint32_t>(ResTable_typeSpec::SPEC_PUBLIC), flags);
|
|
|
|
|
|
|
|
cookie = theme->GetAttribute(app::R::attr::attr_two, &value, &flags);
|
|
|
|
ASSERT_NE(kInvalidCookie, cookie);
|
|
|
|
EXPECT_EQ(Res_value::TYPE_INT_DEC, value.dataType);
|
|
|
|
EXPECT_EQ(2u, value.data);
|
|
|
|
EXPECT_EQ(static_cast<uint32_t>(ResTable_typeSpec::SPEC_PUBLIC), flags);
|
|
|
|
}
|
|
|
|
|
|
|
|
TEST_F(ThemeTest, SingleThemeWithParent) {
|
|
|
|
AssetManager2 assetmanager;
|
|
|
|
assetmanager.SetApkAssets({style_assets_.get()});
|
|
|
|
|
|
|
|
std::unique_ptr<Theme> theme = assetmanager.NewTheme();
|
|
|
|
ASSERT_TRUE(theme->ApplyStyle(app::R::style::StyleTwo));
|
|
|
|
|
|
|
|
Res_value value;
|
|
|
|
uint32_t flags;
|
|
|
|
ApkAssetsCookie cookie;
|
|
|
|
|
|
|
|
cookie = theme->GetAttribute(app::R::attr::attr_one, &value, &flags);
|
|
|
|
ASSERT_NE(kInvalidCookie, cookie);
|
|
|
|
EXPECT_EQ(Res_value::TYPE_INT_DEC, value.dataType);
|
|
|
|
EXPECT_EQ(1u, value.data);
|
|
|
|
EXPECT_EQ(static_cast<uint32_t>(ResTable_typeSpec::SPEC_PUBLIC), flags);
|
|
|
|
|
|
|
|
cookie = theme->GetAttribute(app::R::attr::attr_two, &value, &flags);
|
|
|
|
ASSERT_NE(kInvalidCookie, cookie);
|
|
|
|
EXPECT_EQ(Res_value::TYPE_STRING, value.dataType);
|
|
|
|
EXPECT_EQ(0, cookie);
|
|
|
|
EXPECT_EQ(std::string("string"),
|
|
|
|
GetStringFromPool(assetmanager.GetStringPoolForCookie(0), value.data));
|
|
|
|
EXPECT_EQ(static_cast<uint32_t>(ResTable_typeSpec::SPEC_PUBLIC), flags);
|
|
|
|
|
|
|
|
// This attribute should point to an attr_indirect, so the result should be 3.
|
|
|
|
cookie = theme->GetAttribute(app::R::attr::attr_three, &value, &flags);
|
|
|
|
ASSERT_NE(kInvalidCookie, cookie);
|
|
|
|
EXPECT_EQ(Res_value::TYPE_INT_DEC, value.dataType);
|
|
|
|
EXPECT_EQ(3u, value.data);
|
|
|
|
EXPECT_EQ(static_cast<uint32_t>(ResTable_typeSpec::SPEC_PUBLIC), flags);
|
|
|
|
}
|
|
|
|
|
|
|
|
TEST_F(ThemeTest, MultipleThemesOverlaidNotForce) {
|
|
|
|
AssetManager2 assetmanager;
|
|
|
|
assetmanager.SetApkAssets({style_assets_.get()});
|
|
|
|
|
|
|
|
std::unique_ptr<Theme> theme = assetmanager.NewTheme();
|
|
|
|
ASSERT_TRUE(theme->ApplyStyle(app::R::style::StyleTwo));
|
|
|
|
ASSERT_TRUE(theme->ApplyStyle(app::R::style::StyleThree));
|
|
|
|
|
|
|
|
Res_value value;
|
|
|
|
uint32_t flags;
|
|
|
|
ApkAssetsCookie cookie;
|
|
|
|
|
|
|
|
// attr_one is still here from the base.
|
|
|
|
cookie = theme->GetAttribute(app::R::attr::attr_one, &value, &flags);
|
|
|
|
ASSERT_NE(kInvalidCookie, cookie);
|
|
|
|
EXPECT_EQ(Res_value::TYPE_INT_DEC, value.dataType);
|
|
|
|
EXPECT_EQ(1u, value.data);
|
|
|
|
EXPECT_EQ(static_cast<uint32_t>(ResTable_typeSpec::SPEC_PUBLIC), flags);
|
|
|
|
|
|
|
|
// check for the new attr_six
|
|
|
|
cookie = theme->GetAttribute(app::R::attr::attr_six, &value, &flags);
|
|
|
|
ASSERT_NE(kInvalidCookie, cookie);
|
|
|
|
EXPECT_EQ(Res_value::TYPE_INT_DEC, value.dataType);
|
|
|
|
EXPECT_EQ(6u, value.data);
|
|
|
|
EXPECT_EQ(static_cast<uint32_t>(ResTable_typeSpec::SPEC_PUBLIC), flags);
|
|
|
|
|
|
|
|
// check for the old attr_five (force=true was not used).
|
|
|
|
cookie = theme->GetAttribute(app::R::attr::attr_five, &value, &flags);
|
|
|
|
ASSERT_NE(kInvalidCookie, cookie);
|
|
|
|
EXPECT_EQ(Res_value::TYPE_REFERENCE, value.dataType);
|
|
|
|
EXPECT_EQ(app::R::string::string_one, value.data);
|
|
|
|
EXPECT_EQ(static_cast<uint32_t>(ResTable_typeSpec::SPEC_PUBLIC), flags);
|
|
|
|
}
|
|
|
|
|
|
|
|
TEST_F(ThemeTest, MultipleThemesOverlaidForced) {
|
|
|
|
AssetManager2 assetmanager;
|
|
|
|
assetmanager.SetApkAssets({style_assets_.get()});
|
|
|
|
|
|
|
|
std::unique_ptr<Theme> theme = assetmanager.NewTheme();
|
|
|
|
ASSERT_TRUE(theme->ApplyStyle(app::R::style::StyleTwo));
|
|
|
|
ASSERT_TRUE(theme->ApplyStyle(app::R::style::StyleThree, true /* force */));
|
|
|
|
|
|
|
|
Res_value value;
|
|
|
|
uint32_t flags;
|
|
|
|
ApkAssetsCookie cookie;
|
|
|
|
|
|
|
|
// attr_one is still here from the base.
|
|
|
|
cookie = theme->GetAttribute(app::R::attr::attr_one, &value, &flags);
|
|
|
|
ASSERT_NE(kInvalidCookie, cookie);
|
|
|
|
EXPECT_EQ(Res_value::TYPE_INT_DEC, value.dataType);
|
|
|
|
EXPECT_EQ(1u, value.data);
|
|
|
|
EXPECT_EQ(static_cast<uint32_t>(ResTable_typeSpec::SPEC_PUBLIC), flags);
|
|
|
|
|
|
|
|
// check for the new attr_six
|
|
|
|
cookie = theme->GetAttribute(app::R::attr::attr_six, &value, &flags);
|
|
|
|
ASSERT_NE(kInvalidCookie, cookie);
|
|
|
|
EXPECT_EQ(Res_value::TYPE_INT_DEC, value.dataType);
|
|
|
|
EXPECT_EQ(6u, value.data);
|
|
|
|
EXPECT_EQ(static_cast<uint32_t>(ResTable_typeSpec::SPEC_PUBLIC), flags);
|
|
|
|
|
|
|
|
// check for the new attr_five (force=true was used).
|
|
|
|
cookie = theme->GetAttribute(app::R::attr::attr_five, &value, &flags);
|
|
|
|
ASSERT_NE(kInvalidCookie, cookie);
|
|
|
|
EXPECT_EQ(Res_value::TYPE_INT_DEC, value.dataType);
|
|
|
|
EXPECT_EQ(5u, value.data);
|
|
|
|
EXPECT_EQ(static_cast<uint32_t>(ResTable_typeSpec::SPEC_PUBLIC), flags);
|
2016-12-29 16:08:16 -05:00
|
|
|
}
|
|
|
|
|
|
|
|
TEST_F(ThemeTest, ResolveDynamicAttributesAndReferencesToSharedLibrary) {
|
|
|
|
AssetManager2 assetmanager;
|
|
|
|
assetmanager.SetApkAssets(
|
|
|
|
{lib_two_assets_.get(), lib_one_assets_.get(), libclient_assets_.get()});
|
|
|
|
|
|
|
|
std::unique_ptr<Theme> theme = assetmanager.NewTheme();
|
|
|
|
ASSERT_TRUE(theme->ApplyStyle(libclient::R::style::Theme, false /*force*/));
|
|
|
|
|
|
|
|
Res_value value;
|
|
|
|
uint32_t flags;
|
|
|
|
ApkAssetsCookie cookie;
|
|
|
|
|
|
|
|
// The attribute should be resolved to the final value.
|
|
|
|
cookie = theme->GetAttribute(libclient::R::attr::foo, &value, &flags);
|
|
|
|
ASSERT_NE(kInvalidCookie, cookie);
|
|
|
|
EXPECT_EQ(Res_value::TYPE_INT_DEC, value.dataType);
|
|
|
|
EXPECT_EQ(700u, value.data);
|
|
|
|
EXPECT_EQ(static_cast<uint32_t>(ResTable_typeSpec::SPEC_PUBLIC), flags);
|
|
|
|
|
|
|
|
// The reference should be resolved to a TYPE_REFERENCE.
|
|
|
|
cookie = theme->GetAttribute(libclient::R::attr::bar, &value, &flags);
|
|
|
|
ASSERT_NE(kInvalidCookie, cookie);
|
|
|
|
EXPECT_EQ(Res_value::TYPE_REFERENCE, value.dataType);
|
|
|
|
|
|
|
|
// lib_one is assigned package ID 0x03.
|
2017-01-16 19:11:19 -08:00
|
|
|
EXPECT_EQ(3u, get_package_id(value.data));
|
|
|
|
EXPECT_EQ(get_type_id(lib_one::R::string::foo), get_type_id(value.data));
|
|
|
|
EXPECT_EQ(get_entry_id(lib_one::R::string::foo), get_entry_id(value.data));
|
New implementation of AssetManager/ResTable
The multiwindow model and Resources-per-activity
model that came in N puts greater demands on AssetManagers.
They are created whenever window dimensions change, which
can be frequently. There is a need to be able to cheaply
create a new AssetManager for each Activity, which shares
a lot of underlying state.
In order to make the creation of AssetManagers cheap,
we need a new implementation of the native AssetManager
and ResTable to support immutable representations of
APKs. This new data structure/class is ApkAssets.
ApkAssets have the same functionality of an AssetManager, except
that they operate on a single APK, and they do not do any caching.
Once loaded, they are immutable.
ApkAssets will be exposed as a Java object, with its implementation in
native code. The existing Java StringBlock will be owned by ApkAssets,
which means that Strings can be shared across AssetManagers.
ApkAssets can be cached by the ResourcesManager. Creating an AssetManager
requires only a list of ApkAssets and a configuration.
AssetManager2 (named with the suffix '2' for now while transitioning
to the new implementation) caches bags that are accessed.
Since ApkAssets are expected to be kept around longer, they do more validation
of the resource table, which cause slower load times. Measured on an angler-userdebug,
loading the framework assets takes 11ms with ApkAssets, and 2ms with the old
AssetManager implementation.
The tradeoff is that there does not need to be any security checks once an ApkAssets
is loaded, and regular resource retrieval is faster. Measured on an angler-userdebug,
accessing resource (android:string/ok) with many locales takes 18us with AssetManager2,
and 19us with AssetManager (this is per resource, so these add up).
Test: make libandroidfw_tests
Change-Id: Id0e57ee828f17008891fe3741935a9be8830b01d
2016-10-28 16:39:15 -07:00
|
|
|
}
|
|
|
|
|
|
|
|
TEST_F(ThemeTest, CopyThemeSameAssetManager) {
|
|
|
|
AssetManager2 assetmanager;
|
|
|
|
assetmanager.SetApkAssets({style_assets_.get()});
|
|
|
|
|
|
|
|
std::unique_ptr<Theme> theme_one = assetmanager.NewTheme();
|
|
|
|
ASSERT_TRUE(theme_one->ApplyStyle(app::R::style::StyleOne));
|
|
|
|
|
|
|
|
Res_value value;
|
|
|
|
uint32_t flags;
|
|
|
|
ApkAssetsCookie cookie;
|
|
|
|
|
|
|
|
// attr_one is still here from the base.
|
|
|
|
cookie = theme_one->GetAttribute(app::R::attr::attr_one, &value, &flags);
|
|
|
|
ASSERT_NE(kInvalidCookie, cookie);
|
|
|
|
EXPECT_EQ(Res_value::TYPE_INT_DEC, value.dataType);
|
|
|
|
EXPECT_EQ(1u, value.data);
|
|
|
|
EXPECT_EQ(static_cast<uint32_t>(ResTable_typeSpec::SPEC_PUBLIC), flags);
|
|
|
|
|
|
|
|
// attr_six is not here.
|
|
|
|
EXPECT_EQ(kInvalidCookie, theme_one->GetAttribute(app::R::attr::attr_six, &value, &flags));
|
|
|
|
|
|
|
|
std::unique_ptr<Theme> theme_two = assetmanager.NewTheme();
|
|
|
|
ASSERT_TRUE(theme_two->ApplyStyle(app::R::style::StyleThree));
|
|
|
|
|
|
|
|
// Copy the theme to theme_one.
|
|
|
|
ASSERT_TRUE(theme_one->SetTo(*theme_two));
|
|
|
|
|
|
|
|
// Clear theme_two to make sure we test that there WAS a copy.
|
|
|
|
theme_two->Clear();
|
|
|
|
|
|
|
|
// attr_one is now not here.
|
|
|
|
EXPECT_EQ(kInvalidCookie, theme_one->GetAttribute(app::R::attr::attr_one, &value, &flags));
|
|
|
|
|
|
|
|
// attr_six is now here because it was copied.
|
|
|
|
cookie = theme_one->GetAttribute(app::R::attr::attr_six, &value, &flags);
|
|
|
|
ASSERT_NE(kInvalidCookie, cookie);
|
|
|
|
EXPECT_EQ(Res_value::TYPE_INT_DEC, value.dataType);
|
|
|
|
EXPECT_EQ(6u, value.data);
|
|
|
|
EXPECT_EQ(static_cast<uint32_t>(ResTable_typeSpec::SPEC_PUBLIC), flags);
|
|
|
|
}
|
|
|
|
|
2017-09-25 13:10:14 -07:00
|
|
|
TEST_F(ThemeTest, OnlyCopySystemThemeWhenAssetManagersDiffer) {
|
New implementation of AssetManager/ResTable
The multiwindow model and Resources-per-activity
model that came in N puts greater demands on AssetManagers.
They are created whenever window dimensions change, which
can be frequently. There is a need to be able to cheaply
create a new AssetManager for each Activity, which shares
a lot of underlying state.
In order to make the creation of AssetManagers cheap,
we need a new implementation of the native AssetManager
and ResTable to support immutable representations of
APKs. This new data structure/class is ApkAssets.
ApkAssets have the same functionality of an AssetManager, except
that they operate on a single APK, and they do not do any caching.
Once loaded, they are immutable.
ApkAssets will be exposed as a Java object, with its implementation in
native code. The existing Java StringBlock will be owned by ApkAssets,
which means that Strings can be shared across AssetManagers.
ApkAssets can be cached by the ResourcesManager. Creating an AssetManager
requires only a list of ApkAssets and a configuration.
AssetManager2 (named with the suffix '2' for now while transitioning
to the new implementation) caches bags that are accessed.
Since ApkAssets are expected to be kept around longer, they do more validation
of the resource table, which cause slower load times. Measured on an angler-userdebug,
loading the framework assets takes 11ms with ApkAssets, and 2ms with the old
AssetManager implementation.
The tradeoff is that there does not need to be any security checks once an ApkAssets
is loaded, and regular resource retrieval is faster. Measured on an angler-userdebug,
accessing resource (android:string/ok) with many locales takes 18us with AssetManager2,
and 19us with AssetManager (this is per resource, so these add up).
Test: make libandroidfw_tests
Change-Id: Id0e57ee828f17008891fe3741935a9be8830b01d
2016-10-28 16:39:15 -07:00
|
|
|
AssetManager2 assetmanager_one;
|
2017-09-25 13:10:14 -07:00
|
|
|
assetmanager_one.SetApkAssets({system_assets_.get(), style_assets_.get()});
|
New implementation of AssetManager/ResTable
The multiwindow model and Resources-per-activity
model that came in N puts greater demands on AssetManagers.
They are created whenever window dimensions change, which
can be frequently. There is a need to be able to cheaply
create a new AssetManager for each Activity, which shares
a lot of underlying state.
In order to make the creation of AssetManagers cheap,
we need a new implementation of the native AssetManager
and ResTable to support immutable representations of
APKs. This new data structure/class is ApkAssets.
ApkAssets have the same functionality of an AssetManager, except
that they operate on a single APK, and they do not do any caching.
Once loaded, they are immutable.
ApkAssets will be exposed as a Java object, with its implementation in
native code. The existing Java StringBlock will be owned by ApkAssets,
which means that Strings can be shared across AssetManagers.
ApkAssets can be cached by the ResourcesManager. Creating an AssetManager
requires only a list of ApkAssets and a configuration.
AssetManager2 (named with the suffix '2' for now while transitioning
to the new implementation) caches bags that are accessed.
Since ApkAssets are expected to be kept around longer, they do more validation
of the resource table, which cause slower load times. Measured on an angler-userdebug,
loading the framework assets takes 11ms with ApkAssets, and 2ms with the old
AssetManager implementation.
The tradeoff is that there does not need to be any security checks once an ApkAssets
is loaded, and regular resource retrieval is faster. Measured on an angler-userdebug,
accessing resource (android:string/ok) with many locales takes 18us with AssetManager2,
and 19us with AssetManager (this is per resource, so these add up).
Test: make libandroidfw_tests
Change-Id: Id0e57ee828f17008891fe3741935a9be8830b01d
2016-10-28 16:39:15 -07:00
|
|
|
|
|
|
|
AssetManager2 assetmanager_two;
|
2017-09-25 13:10:14 -07:00
|
|
|
assetmanager_two.SetApkAssets({system_assets_.get(), style_assets_.get()});
|
New implementation of AssetManager/ResTable
The multiwindow model and Resources-per-activity
model that came in N puts greater demands on AssetManagers.
They are created whenever window dimensions change, which
can be frequently. There is a need to be able to cheaply
create a new AssetManager for each Activity, which shares
a lot of underlying state.
In order to make the creation of AssetManagers cheap,
we need a new implementation of the native AssetManager
and ResTable to support immutable representations of
APKs. This new data structure/class is ApkAssets.
ApkAssets have the same functionality of an AssetManager, except
that they operate on a single APK, and they do not do any caching.
Once loaded, they are immutable.
ApkAssets will be exposed as a Java object, with its implementation in
native code. The existing Java StringBlock will be owned by ApkAssets,
which means that Strings can be shared across AssetManagers.
ApkAssets can be cached by the ResourcesManager. Creating an AssetManager
requires only a list of ApkAssets and a configuration.
AssetManager2 (named with the suffix '2' for now while transitioning
to the new implementation) caches bags that are accessed.
Since ApkAssets are expected to be kept around longer, they do more validation
of the resource table, which cause slower load times. Measured on an angler-userdebug,
loading the framework assets takes 11ms with ApkAssets, and 2ms with the old
AssetManager implementation.
The tradeoff is that there does not need to be any security checks once an ApkAssets
is loaded, and regular resource retrieval is faster. Measured on an angler-userdebug,
accessing resource (android:string/ok) with many locales takes 18us with AssetManager2,
and 19us with AssetManager (this is per resource, so these add up).
Test: make libandroidfw_tests
Change-Id: Id0e57ee828f17008891fe3741935a9be8830b01d
2016-10-28 16:39:15 -07:00
|
|
|
|
|
|
|
auto theme_one = assetmanager_one.NewTheme();
|
|
|
|
ASSERT_TRUE(theme_one->ApplyStyle(app::R::style::StyleOne));
|
|
|
|
|
|
|
|
auto theme_two = assetmanager_two.NewTheme();
|
2017-09-25 13:10:14 -07:00
|
|
|
ASSERT_TRUE(theme_two->ApplyStyle(R::style::Theme_One));
|
New implementation of AssetManager/ResTable
The multiwindow model and Resources-per-activity
model that came in N puts greater demands on AssetManagers.
They are created whenever window dimensions change, which
can be frequently. There is a need to be able to cheaply
create a new AssetManager for each Activity, which shares
a lot of underlying state.
In order to make the creation of AssetManagers cheap,
we need a new implementation of the native AssetManager
and ResTable to support immutable representations of
APKs. This new data structure/class is ApkAssets.
ApkAssets have the same functionality of an AssetManager, except
that they operate on a single APK, and they do not do any caching.
Once loaded, they are immutable.
ApkAssets will be exposed as a Java object, with its implementation in
native code. The existing Java StringBlock will be owned by ApkAssets,
which means that Strings can be shared across AssetManagers.
ApkAssets can be cached by the ResourcesManager. Creating an AssetManager
requires only a list of ApkAssets and a configuration.
AssetManager2 (named with the suffix '2' for now while transitioning
to the new implementation) caches bags that are accessed.
Since ApkAssets are expected to be kept around longer, they do more validation
of the resource table, which cause slower load times. Measured on an angler-userdebug,
loading the framework assets takes 11ms with ApkAssets, and 2ms with the old
AssetManager implementation.
The tradeoff is that there does not need to be any security checks once an ApkAssets
is loaded, and regular resource retrieval is faster. Measured on an angler-userdebug,
accessing resource (android:string/ok) with many locales takes 18us with AssetManager2,
and 19us with AssetManager (this is per resource, so these add up).
Test: make libandroidfw_tests
Change-Id: Id0e57ee828f17008891fe3741935a9be8830b01d
2016-10-28 16:39:15 -07:00
|
|
|
ASSERT_TRUE(theme_two->ApplyStyle(app::R::style::StyleTwo));
|
|
|
|
|
2017-09-25 13:10:14 -07:00
|
|
|
EXPECT_TRUE(theme_one->SetTo(*theme_two));
|
|
|
|
|
|
|
|
Res_value value;
|
|
|
|
uint32_t flags;
|
|
|
|
|
|
|
|
// No app resources.
|
|
|
|
EXPECT_EQ(kInvalidCookie, theme_one->GetAttribute(app::R::attr::attr_one, &value, &flags));
|
|
|
|
|
|
|
|
// Only system.
|
|
|
|
EXPECT_NE(kInvalidCookie, theme_one->GetAttribute(R::attr::foreground, &value, &flags));
|
2014-08-29 14:10:04 -07:00
|
|
|
}
|
2016-11-10 16:43:59 -08:00
|
|
|
|
|
|
|
} // namespace android
|