zipalign的使用

zipalign

是一个存档对齐工具，可为Android应用程序（.apk）文件提供重要的优化。目的是确保所有未压缩数据以相对于文件开头的特定对齐开始。具体来说，它会导致.apk中的所有未压缩数据（如图像或原始文件）在4字节边界上对齐。这允许直接访问所有部分，mmap()即使它们包含具有对齐限制的二进制数据。其好处是能够减少应用程序的RAM内存资源消耗。

在将.apk文件分发给最终用户之前，应始终使用此工具来对齐.apk文件。Android构建工具可以为您处理此问题。将Eclipse与ADT插件一起使用时，导出向导会在您使用私钥对其进行签名后自动为.apk设置zipalign。在使用Ant编译应用程序时使用的构建脚本也将对.apk进行zipalign，只要您提供了密钥库的路径和项目ant.properties文件中的密钥别名，以便构建工具可以首先对包进行签名。

警告：只有在使用您的私钥对.apk文件进行签名后才能执行zipalign 。如果在签名之前执行zipalign，则签名过程将撤消对齐。此外，不要对对齐的包进行更改。对存档的更改（例如重命名或删除条目）可能会破坏已修改条目和所有后续条目的对齐。添加到“对齐”存档的任何文件都不会对齐。

通过更改zip本地文件头部分中“额外”字段的大小来进行调整。“额外”字段中的现有数据可以通过该过程改变。

有关如何在构建应用程序时使用zipalign的更多信息，请阅读签署您的应用程序。

1、对infile.apk进行对齐并且保存为outfile.apk

zipalign [-f] [-v] infile.apk outfile.apk

2、检查apk是否进行了对齐

zipalign – c – v existing.apk

infile.apk表示需要存档对齐的apk
outfile.apk表示存档对齐后的apk
alignment：表示指定的对应字节数，是一个整数且必须指定为4。

转载请标明出处：http://77blogs.com/?p=288

原文：google开发文档：http://www.android-doc.com/tools/help/zipalign.html

验证apk签名方式（V1 || V2）

转载请标明出处：http://77blogs.com/?p=289

进入SDK\build-tools\28.0.2目录（或者其他版本），该目录有apksigner.bar脚本，我们可以利用它来验证。

在此目录打开命令行。

命令为：apksigner verify -v apk

回车便可以看到信息。

Gradle Task顺序

原文链接：https://blog.csdn.net/lzyzsd/article/details/46935405

我注意到我在使用Gradle的时候遇到的大多数问题都是和task的执行顺序有关的。很明显如果我的构建会工作的更好如果我的task都是在正确的时候执行。下面我们就深入了解一下如何更改task的执行顺序。

dependsOn

我认为最直接的方式来说明的你task的执行时依赖别的task的方法就是使用dependsOn方法。
比如下面的场景，已经存在task A，我们要添加一个task B，它的执行必须要在A执行完之后：
B->A
这是一个很简单的场景，假定A和B的定义如下：

task A &lt;&lt; {println 'Hello from A'}
task B &lt;&lt; {println 'Hello from B'}

1 2	task A << {println 'Hello from A'} task B << {println 'Hello from B'}

只需要简单的调用B.dependsOn A，就可以了。
这意味着，只要我执行task B，task A都会先执行。

<span style="font-size: 15px; color: #993300;">paveldudka$ gradle B
:A
Hello from A
:B
Hello from B</span>

paveldudka$ gradle B

Hello from A

Hello from B

另外，你也可以在task的配置区中来声明它的依赖：

task A &lt;&lt; {println 'Hello from A'}
task B {
    dependsOn A
    doLast {
        println 'Hello from B'  
    }
}

task A << {println 'Hello from A'}

task B {

dependsOn A

doLast {

println 'Hello from B'

}

如果我们想要在已经存在的task依赖中插入我们的task该怎么做呢？
insert task
过程和刚才类似。假定已经存在如下的task依赖：

task A &lt;&lt; {println 'Hello from A'}
task B &lt;&lt; {println 'Hello from B'}
task C &lt;&lt; {println 'Hello from C'}

B.dependsOn A
C.dependsOn B

task A << {println 'Hello from A'}

task B << {println 'Hello from B'}

task C << {println 'Hello from C'}

B.dependsOn A

C.dependsOn B

加入我们的新的task

task B1 &lt;&lt; {println 'Hello from B1'}
B1.dependsOn B
C.dependsOn B1

task B1 << {println 'Hello from B1'}

B1.dependsOn B

C.dependsOn B1

输出：

<span style="font-size: 15px; color: #993300;"><code class="language-shell hljs ruby has-numbering">paveldudka<span class="hljs-variable">$ gradle <span class="hljs-constant">C
<span class="hljs-symbol">:A
<span class="hljs-constant">Hello from <span class="hljs-constant">A
<span class="hljs-symbol">:B
<span class="hljs-constant">Hello from <span class="hljs-constant">B
<span class="hljs-symbol">:B1
<span class="hljs-constant">Hello from <span class="hljs-constant">B1
<span class="hljs-symbol">:C
<span class="hljs-constant">Hello from <span class="hljs-constant">C</span></span></span></span></span></span></span></span></span></span></span></span></span></span></code></span>

<code class="language-shell hljs ruby has-numbering">paveldudka$ gradle C

:A

Hello from A

:B

Hello from B

:B1

Hello from B1

:C

Hello from C</code>

注意dependsOn把task添加到依赖的集合中，所以依赖多个task是没有问题的。
multi depends

task B1 &lt;&lt; {println 'Hello from B1'}
B1.dependsOn B
B1.dependsOn Q

task B1 << {println 'Hello from B1'}

B1.dependsOn B

B1.dependsOn Q

输出：

<span style="font-size: 15px; color: #993300;"><code class="language-shell hljs ruby has-numbering">paveldudka<span class="hljs-variable">$ gradle <span class="hljs-constant">B1
<span class="hljs-symbol">:A
<span class="hljs-constant">Hello from <span class="hljs-constant">A
<span class="hljs-symbol">:B
<span class="hljs-constant">Hello from <span class="hljs-constant">B
<span class="hljs-symbol">:Q
<span class="hljs-constant">Hello from <span class="hljs-constant">Q
<span class="hljs-symbol">:B1
<span class="hljs-constant">Hello from <span class="hljs-constant">B1</span></span></span></span></span></span></span></span></span></span></span></span></span></span></code></span>

<code class="language-shell hljs ruby has-numbering">paveldudka$ gradle B1

:A

Hello from A

:B

Hello from B

:Q

Hello from Q

:B1

Hello from B1</code>

mustRunAfter

现在假定我又一个task，它依赖于其他两个task。这里我使用一个真实的场景，我有两个task，一个单元测试的task，一个是UI测试的task。另外还有一个task是跑所有的测试的，它依赖于前面的两个task。
all tests

task unit &lt;&lt; {println 'Hello from unit tests'}
task ui &lt;&lt; {println 'Hello from UI tests'}
task tests &lt;&lt; {println 'Hello from all tests!'}

tests.dependsOn unit
tests.dependsOn ui

task unit << {println 'Hello from unit tests'}

task ui << {println 'Hello from UI tests'}

task tests << {println 'Hello from all tests!'}

tests.dependsOn unit

tests.dependsOn ui

输出：

<span style="font-size: 15px; color: #993300;"><code class="language-shell hljs oxygene has-numbering">paveldudka$ gradle tests
:ui
Hello <span class="hljs-keyword">from UI tests
:<span class="hljs-keyword">unit
Hello <span class="hljs-keyword">from <span class="hljs-keyword">unit tests
:tests
Hello <span class="hljs-keyword">from all tests!</span></span></span></span></span></code></span>

<code class="language-shell hljs oxygene has-numbering">paveldudka$ gradle tests

:ui

Hello from UI tests

:unit

Hello from unit tests

:tests

Hello from all tests!</code>

尽管unitest和UI test会子啊test task之前执行，但是unit和ui这两个task的执行顺序是不能保证的。虽然现在来看是按照字母表的顺序执行，但这是依赖于Gradle的实现的，你的代码中绝对不能依赖这种顺序。
由于UI测试时间远比unit test时间长，因此我希望unit test先执行。一个解决办法就是让ui task依赖于unit task。
ui->unit

task unit &lt;&lt; {println 'Hello from unit tests'}
task ui &lt;&lt; {println 'Hello from UI tests'}
task tests &lt;&lt; {println 'Hello from all tests!'}

tests.dependsOn unit
tests.dependsOn ui
ui.dependsOn unit // &lt;-- I added this dependency

task unit << {println 'Hello from unit tests'}

task ui << {println 'Hello from UI tests'}

task tests << {println 'Hello from all tests!'}

tests.dependsOn unit

tests.dependsOn ui

ui.dependsOn unit // <-- I added this dependency

输出:

<span style="font-size: 15px; color: #993300;"><code class="language-shell hljs oxygene has-numbering">paveldudka$ gradle tests
:<span class="hljs-keyword">unit
Hello <span class="hljs-keyword">from <span class="hljs-keyword">unit tests
:ui
Hello <span class="hljs-keyword">from UI tests
:tests
Hello <span class="hljs-keyword">from all tests!</span></span></span></span></span></code></span>

<code class="language-shell hljs oxygene has-numbering">paveldudka$ gradle tests

:unit

Hello from unit tests

:ui

Hello from UI tests

:tests

Hello from all tests!</code>

现在unit test会在ui test之前执行了。
但是这里有个很恶心的问题，我的ui测试其实并不依赖于unit test。我希望能够单独的执行ui test，但是这里每次我执行ui test，都会先执行unit test。
这里就要用到mustRunAfter了。mustRunAfter并不会添加依赖，它只是告诉Gradle执行的优先级如果两个task同时存在。比如我们这里就可以指定ui.mustRunAfter unit，这样如果ui task和unit task同时存在，Gradle会先执行unit test，而如果只执行gradle ui，并不会去执行unit task。

task unit &lt;&lt; {println 'Hello from unit tests'}
task ui &lt;&lt; {println 'Hello from UI tests'}
task tests &lt;&lt; {println 'Hello from all tests!'}

tests.dependsOn unit
tests.dependsOn ui
ui.mustRunAfter unit

task unit << {println 'Hello from unit tests'}

task ui << {println 'Hello from UI tests'}

task tests << {println 'Hello from all tests!'}

tests.dependsOn unit

tests.dependsOn ui

ui.mustRunAfter unit

输出：

<span style="font-size: 15px; color: #993300;"><code class="language-groovy hljs oxygene has-numbering">paveldudka$ gradle tests
:<span class="hljs-keyword">unit
Hello <span class="hljs-keyword">from <span class="hljs-keyword">unit tests
:ui
Hello <span class="hljs-keyword">from UI tests
:tests
Hello <span class="hljs-keyword">from all tests!</span></span></span></span></span></code></span>

<code class="language-groovy hljs oxygene has-numbering">paveldudka$ gradle tests

:unit

Hello from unit tests

:ui

Hello from UI tests

:tests

Hello from all tests!</code>

依赖关系如下图:
mustRunAfter

mustRunAfter在Gradle2.4中目前还是实验性的功能。

finalizedBy

现在我们已经有两个task，unit和ui，假定这两个task都会输出测试报告，现在我想把这两个测试报告合并成一个:
merge

task unit &lt;&lt; {println 'Hello from unit tests'}
task ui &lt;&lt; {println 'Hello from UI tests'}
task tests &lt;&lt; {println 'Hello from all tests!'}
task mergeReports &lt;&lt; {println 'Merging test reports'}

tests.dependsOn unit
tests.dependsOn ui
ui.mustRunAfter unit
mergeReports.dependsOn tests

task unit << {println 'Hello from unit tests'}

task ui << {println 'Hello from UI tests'}

task tests << {println 'Hello from all tests!'}

task mergeReports << {println 'Merging test reports'}

tests.dependsOn unit

tests.dependsOn ui

ui.mustRunAfter unit

mergeReports.dependsOn tests

现在如果我想获得ui和unit的测试报告，执行task mergeReports就可以了。

<span style="font-size: 15px; color: #993300;"><code class="language-shell hljs ruby has-numbering">paveldudka<span class="hljs-variable">$ gradle mergeReports
<span class="hljs-symbol">:unit
<span class="hljs-constant">Hello from unit tests
<span class="hljs-symbol">:ui
<span class="hljs-constant">Hello from <span class="hljs-constant">UI tests
<span class="hljs-symbol">:tests
<span class="hljs-constant">Hello from all tests!
<span class="hljs-symbol">:mergeReports
<span class="hljs-constant">Merging test reports</span></span></span></span></span></span></span></span></span></span></code></span>

<code class="language-shell hljs ruby has-numbering">paveldudka$ gradle mergeReports

:unit

Hello from unit tests

:ui

Hello from UI tests

:tests

Hello from all tests!

:mergeReports

Merging test reports</code>

这个task是能工作，但是看起来好笨啊。mergeReports从用户的角度来看感觉不是特别好。我希望执行tests task就可以获得测试报告，而不必知道mergeReports的存在。当然我可以把merge的逻辑挪到tests task中，但我不想把tests task搞的太臃肿，我还是继续把merge的逻辑放在mergeReports task中。
finalizeBy来救场了。顾名思义，finalizeBy就是在task执行完之后要执行的task。修改我们的脚本如下：

task unit &lt;&lt; {println 'Hello from unit tests'}
task ui &lt;&lt; {println 'Hello from UI tests'}
task tests &lt;&lt; {println 'Hello from all tests!'}
task mergeReports &lt;&lt; {println 'Merging test reports'}

tests.dependsOn unit
tests.dependsOn ui
ui.mustRunAfter unit
mergeReports.dependsOn tests

tests.finalizedBy mergeReports

task unit << {println 'Hello from unit tests'}

task ui << {println 'Hello from UI tests'}

task tests << {println 'Hello from all tests!'}

task mergeReports << {println 'Merging test reports'}

tests.dependsOn unit

tests.dependsOn ui

ui.mustRunAfter unit

mergeReports.dependsOn tests

tests.finalizedBy mergeReports

现在执行tests task就可以拿到测试报告了：

<span style="font-size: 15px; color: #993300;"><code class="language-shell hljs ruby has-numbering">paveldudka<span class="hljs-variable">$ gradle tests
<span class="hljs-symbol">:unit
<span class="hljs-constant">Hello from unit tests
<span class="hljs-symbol">:ui
<span class="hljs-constant">Hello from <span class="hljs-constant">UI tests
<span class="hljs-symbol">:tests
<span class="hljs-constant">Hello from all tests!
<span class="hljs-symbol">:mergeReports
<span class="hljs-constant">Merging test reports</span></span></span></span></span></span></span></span></span></span></code></span>

<code class="language-shell hljs ruby has-numbering">paveldudka$ gradle tests

:unit

Hello from unit tests

:ui

Hello from UI tests

:tests

Hello from all tests!

:mergeReports

Merging test reports</code>

注意，finalizedBy也是Gradle2.4的实验性功能

shrinkResources去除无用资源原理

开启shrinkResources后，打包过程会新增task transformClassesWithShrinkResFor{variant}，gradle1.5之后只需要注册一个tranform就会产生一个对应的task，查看源码发现对应的tranfrom在com.android.build.gradle.internal.transforms.ShrinkResourcesTransform，此类中调用com.android.build.gradle.tasks.ResourceUsageAnalyzer的analyze方法进行分析无用资源。

ResourceUsageAnalyzer：

该类的开头有这么一段话：

<span style="color: #008000;">/**</span><span style="color: #008000;">
 * Class responsible for searching through a Gradle built tree (after resource merging, compilation
 * and shrinking has been completed, but before final .apk assembly), which figures out which
 * resources if any are unused, and removes them.
 *
 * <p>It does this by examining
 *
 * <ul>
 *   <li>The merged manifest, to find root resource references (such as drawables used for activity
 *       icons)
 *   <li>The merged R class (to find the actual integer constants assigned to resources)
 *   <li>The ProGuard mapping files (to find the mapping from original symbol names to short names)*
 *   <li>The merged resources (to find which resources reference other resources, e.g. drawable
 *       state lists including other drawables, or layouts including other layouts, or styles
 *       referencing other drawables, or menus items including action layouts, etc.)
 *   <li>The shrinked output classes (to find resource references in code that are actually
 *       reachable)
 * </ul>
 *
 * From all this, it builds up a reference graph, and based on the root references (e.g. from the
 * manifest and from the remaining code) it computes which resources are actually reachable in the
 * app, and anything that is not reachable is then marked for deletion.
 *
 * <p>A resource is referenced in code if either the field R.type.name is referenced (which is the
 * case for non-final resource references, e.g. in libraries), or if the corresponding int value is
 * referenced (for final resource values). We check this by looking at the shrinked output classes
 * with an ASM visitor. One complication is that code can also call {</span><span style="color: #808080;">@code</span><span style="color: #008000;">
 * Resources#getIdentifier(String,String,String)} where they can pass in the names of resources to
 * look up. To handle this scenario, we use the ClassVisitor to see if there are any calls to the
 * specific {</span><span style="color: #808080;">@code</span><span style="color: #008000;"> Resources#getIdentifier} method. If not, great, the usage analysis is completely
 * accurate. If we <b>do</b> find one, we check <b>all</b> the string constants found anywhere in
 * the app, and look to see if any look relevant. For example, if we find the string "string/foo" or
 * "my.pkg:string/foo", we will then mark the string resource named foo (if any) as potentially
 * used. Similarly, if we find just "foo" or "/foo", we will mark <b>all</b> resources named "foo"
 * as potentially used. However, if the string is "bar/foo" or " foo " these strings are ignored.
 * This means we can potentially miss resources usages where the resource name is completed computed
 * (e.g. by concatenating individual characters or taking substrings of strings that do not look
 * like resource names), but that seems extremely unlikely to be a real-world scenario.
 *
 * <p>Analyzing dex files is also supported. It follows the same rules as analyzing class files.
 *
 * <p>For now, for reasons detailed in the code, this only applies to file-based resources like
 * layouts, menus and drawables, not value-based resources like strings and dimensions.
 </span><span style="color: #008000;">*/</span>

/**

* Class responsible for searching through a Gradle built tree (after resource merging, compilation

* and shrinking has been completed, but before final .apk assembly), which figures out which

* resources if any are unused, and removes them.

* It does this by examining

* <ul>

* <li>The merged manifest, to find root resource references (such as drawables used for activity

* icons)

* <li>The merged R class (to find the actual integer constants assigned to resources)

* <li>The ProGuard mapping files (to find the mapping from original symbol names to short names)*

* <li>The merged resources (to find which resources reference other resources, e.g. drawable

* state lists including other drawables, or layouts including other layouts, or styles

* referencing other drawables, or menus items including action layouts, etc.)

* <li>The shrinked output classes (to find resource references in code that are actually

* reachable)

* </ul>

* From all this, it builds up a reference graph, and based on the root references (e.g. from the

* manifest and from the remaining code) it computes which resources are actually reachable in the

* app, and anything that is not reachable is then marked for deletion.

* A resource is referenced in code if either the field R.type.name is referenced (which is the

* case for non-final resource references, e.g. in libraries), or if the corresponding int value is

* referenced (for final resource values). We check this by looking at the shrinked output classes

* with an ASM visitor. One complication is that code can also call {@code

* Resources#getIdentifier(String,String,String)} where they can pass in the names of resources to

* look up. To handle this scenario, we use the ClassVisitor to see if there are any calls to the

* specific {@code Resources#getIdentifier} method. If not, great, the usage analysis is completely

* accurate. If we do find one, we check all the string constants found anywhere in

* the app, and look to see if any look relevant. For example, if we find the string "string/foo" or

* "my.pkg:string/foo", we will then mark the string resource named foo (if any) as potentially

* used. Similarly, if we find just "foo" or "/foo", we will mark all resources named "foo"

* as potentially used. However, if the string is "bar/foo" or " foo " these strings are ignored.

* This means we can potentially miss resources usages where the resource name is completed computed

* (e.g. by concatenating individual characters or taking substrings of strings that do not look

* like resource names), but that seems extremely unlikely to be a real-world scenario.

* Analyzing dex files is also supported. It follows the same rules as analyzing class files.

* For now, for reasons detailed in the code, this only applies to file-based resources like

* layouts, menus and drawables, not value-based resources like strings and dimensions.

*/

我翻译一下这段：

这个类负责搜索一颗gradle构建树（在合并资源，编译，压缩已经完成之后，最终的apk组装之前），这颗构建树确定了那些资源是未使用的，并将其删除。

它通过检查做到这一点：

合并的清单，用于查找根资源引用（例如启动图标）
合并的R类（查找分配给资源的实际整数常量）
混淆映射文件（找到原始名称到混淆后短名称的一个映射）
合并的资源（用于查找哪些资源引用其他资源，例如包括其他other drawables的drawable state lists，或包括其他布局的布局，或引用其他drawables的styles，或包括布局文件的菜单项等）
缩小的输出类（在代码中查找实际可访问的资源引用）

所有的这些，构建了一个引用图，并基于根引用（例如来自清单与剩余代码），它计算出了在app中哪些资源是实际可达的，然后将任何无法访问的内容标记为删除。

如果文件R.type.name被引用（非最终资源引用，例如在库中），或者在代码中引用相应的int值，我们通过ASM查看输出的代码来检查这一点。有一个复杂的问题是代码可以通过Resources#getIdentifier(String,String,String)，通过传递资源的名称去查查找资源，为了处理这种情况，我们使用ClassVisitor来查看是否有对特定Resources#getIdentifier方法的任何调用。如果没有，很好，使用分析是完全准确的。如果我们找到一个，我们检查所有在应用程序中的任何位置找到的字符串常量，并查看是否有看起来有关。例如，如果我们找到字符串“string / foo”或“my.pkg：string / foo”，我们将标记名为foo的字符串资源（如果有）作为可能使用的字符串资源。

 <span style="color: #0000ff;">public</span> <span style="color: #0000ff;">void</span> analyze() <span style="color: #0000ff;">throws</span><span style="color: #000000;"> IOException, ParserConfigurationException, SAXException {
        gatherResourceValues(mResourceClassDir);
        recordMapping(mProguardMapping);

        </span><span style="color: #0000ff;">for</span><span style="color: #000000;"> (File jarOrDir : mClasses) {
            recordClassUsages(jarOrDir);
        }

        recordManifestUsages(mMergedManifest);
        recordResources(mMergedResourceDir);
        keepPossiblyReferencedResources();
        dumpReferences();
        mModel.processToolsAttributes();
        mUnused </span>=<span style="color: #000000;"> mModel.findUnused();
    }</span>

public void analyze() throws IOException, ParserConfigurationException, SAXException {

gatherResourceValues(mResourceClassDir);

recordMapping(mProguardMapping);

for (File jarOrDir : mClasses) {

recordClassUsages(jarOrDir);

}

recordManifestUsages(mMergedManifest);

recordResources(mMergedResourceDir);

keepPossiblyReferencedResources();

dumpReferences();

mModel.processToolsAttributes();

mUnused = mModel.findUnused();

}

一：

<span style="color: #0000ff;">private</span> <span style="color: #0000ff;">void</span> gatherResourceValues(File file) <span style="color: #0000ff;">throws</span><span style="color: #000000;"> IOException {
        </span><span style="color: #0000ff;">if</span><span style="color: #000000;"> (file.isDirectory()) {
            File[] children </span>=<span style="color: #000000;"> file.listFiles();
            </span><span style="color: #0000ff;">if</span> (children != <span style="color: #0000ff;">null</span><span style="color: #000000;">) {
                </span><span style="color: #0000ff;">for</span><span style="color: #000000;"> (File child : children) {
                    gatherResourceValues(child);
                }
            }
        } </span><span style="color: #0000ff;">else</span> <span style="color: #0000ff;">if</span> (file.isFile() &&<span style="color: #000000;"> file.getName().equals(SdkConstants.FN_RESOURCE_CLASS)) {
            parseResourceClass(file);
        }
    }</span>

private void gatherResourceValues(File file) throws IOException {

if (file.isDirectory()) {

File[] children = file.listFiles();

if (children != null) {

for (File child : children) {

gatherResourceValues(child);

}

} else if (file.isFile() && file.getName().equals(SdkConstants.FN_RESOURCE_CLASS)) {

parseResourceClass(file);

}

}

SdkConstants.FN_RESOURCE_CLASS = R.java

可以看出它找到的是R.java文件，然后调用parseResourceClass(file);方法，将R.java中的资源都标记为可达，加入表中。

recordMapping(mProguardMapping)

解析混淆文件，将现在的把映射存于表中

recordManifestUsages(mMergedManifest);

分析Manifest文件，将找到的引用标记为可达，例如：启动图标

AndResGuard的使用

AndResGuard是何物？

AndResGuard是一个帮助你缩小APK大小的工具，他的原理类似Java Proguard，但是只针对资源。他会将原本冗长的资源路径变短，例如将res/drawable/wechat变为r/d/a。一般用于给apk瘦身。

具体请看：

https://github.com/shwenzhang/AndResGuard/blob/master/README.zh-cn.md

我们解压缩apk后可以看到这个文件

resources.arsc

这个文件是编译后的二进制资源文件，里面是id-name-value的一个map，即id与资源名称的一个映射，如：

AndResGuard。它将资源的名称进行了混淆，所以可以用它对resources.arsc进行优化，只是具体优化效果与编码方式、id数量、平均减少命名长度有关。

例如：

表一：

表二：

表二是经过优化后的，我们一眼就可以知道表2肯定比表1存储的字符要小，所以整个文件的大小肯定也要小一些，因此就达到了瘦身的效果。

总的来说它可以将res/drawable/activity_advanced_setting_for_test变为r/d/a，达到混淆与减少体积的效果，也增加了反编译后阅读代码的难度，安全性高。

那么怎么使用呢？

第一种方式：AS集成

1、在项目的build.gradle中添加依赖如下：

<span style="color: #000000;">buildscript {
    repositories {
        jcenter()
    }

    dependencies {
        classpath </span>'com.android.tools.build:gradle:2.2.3'<span style="color: #000000;">
        classpath </span>'com.tencent.mm:AndResGuard-gradle-plugin:1.1.16'<span style="color: #000000;">
    }
}</span>

buildscript {

repositories {

jcenter()

}

dependencies {

classpath 'com.android.tools.build:gradle:2.2.3'

classpath 'com.tencent.mm:AndResGuard-gradle-plugin:1.1.16'

}

}

2、在app下建立gradle文件and_res_guard.gradle，里面的内容如下：

apply plugin: 'AndResGuard'<span style="color: #000000;">

andResGuard {<br />    <span style="color: #008000;">//mappingFile = file("./resource_mapping.txt")</span>
    mappingFile </span>= <span style="color: #0000ff;">null<br /></span><span style="color: #000000;"><span style="color: #008000;">    //使用7Zip压缩</span>
    use7zip </span>= <span style="color: #0000ff;">true</span><span style="color: #000000;">
    useSign </span>= <span style="color: #0000ff;">true<br />    <br /></span><span style="color: #000000;"><span style="color: #008000;">    //打开这个开关的话，会keep住所有资源的原始路径，值混淆名称</span>
    keepRoot </span>= <span style="color: #0000ff;">false</span><span style="color: #000000;">
    compressFilePattern </span>=<span style="color: #000000;"> [
            </span>"*.png"<span style="color: #000000;">,
            </span>"*.jpg"<span style="color: #000000;">,
            </span>"*.jpeg"<span style="color: #000000;">,
            </span>"*.gif"<span style="color: #000000;">,
            </span>"resources.arsc"<span style="color: #000000;">
    ]
    whiteList </span>=<span style="color: #000000;"> [
            </span><span style="color: #008000;">//</span><span style="color: #008000;"> your icon</span>
            "R.drawable.icon"<span style="color: #000000;">,
            </span><span style="color: #008000;">//</span><span style="color: #008000;"> for fabric</span>
            "R.string.com.crashlytics.*"<span style="color: #000000;">,
            </span><span style="color: #008000;">//</span><span style="color: #008000;"> for umeng update</span>
            "R.string.tb_*"<span style="color: #000000;">,
            </span>"R.layout.tb_*"<span style="color: #000000;">,
            </span>"R.drawable.tb_*"<span style="color: #000000;">,
            </span>"R.drawable.u1*"<span style="color: #000000;">,
            </span>"R.drawable.u2*"<span style="color: #000000;">,
            </span>"R.color.tb_*"<span style="color: #000000;">,
            </span><span style="color: #008000;">//</span><span style="color: #008000;"> umeng share for sina</span>
            "R.drawable.sina*"<span style="color: #000000;">,
            </span><span style="color: #008000;">//</span><span style="color: #008000;"> for google-services.json</span>
            "R.string.google_app_id"<span style="color: #000000;">,
            </span>"R.string.gcm_defaultSenderId"<span style="color: #000000;">,
            </span>"R.string.default_web_client_id"<span style="color: #000000;">,
            </span>"R.string.ga_trackingId"<span style="color: #000000;">,
            </span>"R.string.firebase_database_url"<span style="color: #000000;">,
            </span>"R.string.google_api_key"<span style="color: #000000;">,
            </span>"R.string.google_crash_reporting_api_key"<span style="color: #000000;">,

            </span><span style="color: #008000;">//</span><span style="color: #008000;">友盟</span>
            "R.string.umeng*"<span style="color: #000000;">,
            </span>"R.string.UM*"<span style="color: #000000;">,
            </span>"R.layout.umeng*"<span style="color: #000000;">,
            </span>"R.drawable.umeng*"<span style="color: #000000;">,
            </span>"R.id.umeng*"<span style="color: #000000;">,
            </span>"R.anim.umeng*"<span style="color: #000000;">,
            </span>"R.color.umeng*"<span style="color: #000000;">,
            </span>"R.style.*UM*"<span style="color: #000000;">,
            </span>"R.style.umeng*"<span style="color: #000000;">,
    ]

    sevenzip {
        artifact </span>= 'com.tencent.mm:SevenZip:1.2.10'
        <span style="color: #008000;">//</span><span style="color: #008000;">path = "/usr/local/bin/7za"</span>
<span style="color: #000000;">    }
}</span>

apply plugin: 'AndResGuard'

andResGuard { //mappingFile = file("./resource_mapping.txt")

mappingFile = null //使用7Zip压缩

use7zip = true

useSign = true //打开这个开关的话，会keep住所有资源的原始路径，值混淆名称

keepRoot = false

compressFilePattern = [

"*.png",

"*.jpg",

"*.jpeg",

"*.gif",

"resources.arsc"

]

whiteList = [

// your icon

"R.drawable.icon",

// for fabric

"R.string.com.crashlytics.*",

// for umeng update

"R.string.tb_*",

"R.layout.tb_*",

"R.drawable.tb_*",

"R.drawable.u1*",

"R.drawable.u2*",

"R.color.tb_*",

// umeng share for sina

"R.drawable.sina*",

// for google-services.json

"R.string.google_app_id",

"R.string.gcm_defaultSenderId",

"R.string.default_web_client_id",

"R.string.ga_trackingId",

"R.string.firebase_database_url",

"R.string.google_api_key",

"R.string.google_crash_reporting_api_key",

//友盟

"R.string.umeng*",

"R.string.UM*",

"R.layout.umeng*",

"R.drawable.umeng*",

"R.id.umeng*",

"R.anim.umeng*",

"R.color.umeng*",

"R.style.*UM*",

"R.style.umeng*",

]

sevenzip {

artifact = 'com.tencent.mm:SevenZip:1.2.10'

//path = "/usr/local/bin/7za"

}

}

其中whiteList（白名单）中指定不需要进行混淆的资源路径规则，所有使用getIdentifier访问的资源都需要加入白名单，因为使用这种方式访问的资源，在代码中写死了id，比如：

int mipmapId = getResources().getIdentifier(“ic_launcher”, “mipmap”, getPackageName());，如果不加入白名单的话会报找不到资源的错误。

还有一些第三方SDK，因为有些SDK的代码中也用这种方式引用到对应的资源文件，如果对其进行混淆，会导致找不到对应资源文件，出现crash，所以不能对其资源文件进行混淆。像友盟这种喜欢用反射获取资源的SDK就是一个坑（友盟的SDK就是坑王）！对于app启动图标这样的icon可以不做混淆，推荐将其放入白名单里。

可以在white_list.md查看更多sdk的白名单配置，也欢迎大家PR自己的白名单

compressFilePattern和compressFilePattern中的通配符支持? + *

<span style="font-family: 楷体; font-size: 15px;"><code>?	Zero or one character
*	Zero or more of character
+	One or more of character<br /></code></span>

<code>? Zero or one character

* Zero or more of character

+ One or more of character </code>

白名单机制只作用于资源的specsName，不会keep住资源的路径。如果想keep住资源原有的物理路径，可以使用mappingFile。例如我想keep住icon所有folder，可以在mappingFile指向的文件添加：

<span style="color: #000000;">res path mapping:
    res</span>/mipmap-hdpi-v4 -> res/mipmap-hdpi-<span style="color: #000000;">v4
    res</span>/mipmap-mdpi-v4 -> res/mipmap-mdpi-<span style="color: #000000;">v4
    res</span>/mipmap-xhdpi-v4 -> res/mipmap-xhdpi-<span style="color: #000000;">v4
    res</span>/mipmap-xxhdpi-v4 -> res/mipmap-xxhdpi-<span style="color: #000000;">v4
    res</span>/mipmap-xxxhdpi-v4 -> res/mipmap-xxxhdpi-v4

res path mapping:

res/mipmap-hdpi-v4 -> res/mipmap-hdpi-v4

res/mipmap-mdpi-v4 -> res/mipmap-mdpi-v4

res/mipmap-xhdpi-v4 -> res/mipmap-xhdpi-v4

res/mipmap-xxhdpi-v4 -> res/mipmap-xxhdpi-v4

res/mipmap-xxxhdpi-v4 -> res/mipmap-xxxhdpi-v4

注意！

如果不是对APK size有极致的需求，请不要把resource.asrc添加进compressFilePattern. (#84 #233)
对于发布于Google Play的APP，建议不要使用7Zip压缩，因为这个会导致Google Play的优化Patch算法失效. (#233)

3、在app的build.gradle中引入：

apply from: 'and_res_guard.gradle'

1	apply from: 'and_res_guard.gradle'

编译一下便可以在gradle的task任务列表里看到混淆任务

使用：

双击对应的task可以编译debug包，release包。混淆后的apk生成在build/output/apk/AndResGuard_*目录中，默认会生成4种apk，我们选择签名、压缩、对齐后的apk即可，后缀名是*_signed_7zip_aligned.apk

第二种方式：指令集成

1、下载AndResGuard过程：

https://github.com/shwenzhang/AndResGuard

2、解压，可以看到这个文件，进入这个文件：

文件夹里面有这么一些东西：

文件作用：

jar包：用来执行命令的时候使用

build_apk.bat :windows执行脚本

set jdkpath=D:\Program Files\Java\jdk1.7<span style="color: #000000;">.0_79\bin\java.exe
set storepath</span>=<span style="color: #000000;">release.keystore
set storepass</span>=<span style="color: #000000;">testres
set keypass</span>=<span style="color: #000000;">testres
set alias</span>=<span style="color: #000000;">testres
set zipalign</span>=D:\soft\dev\android\sdk\build-tools\23.0.2<span style="color: #000000;">\zipalign.exe
</span>"%jdkpath%" -jar AndResGuard-cli-1.2.15.jar input.apk -config config.xml -out outapk -signature "%storepath%" "%storepass%" "%keypass%" "%alias%" -zipalign "%zipalign%"<span style="color: #000000;">
pause</span>

set jdkpath=D:\Program Files\Java\jdk1.7.0_79\bin\java.exe

set storepath=release.keystore

set storepass=testres

set keypass=testres

set alias=testres

set zipalign=D:\soft\dev\android\sdk\build-tools\23.0.2\zipalign.exe

"%jdkpath%" -jar AndResGuard-cli-1.2.15.jar input.apk -config config.xml -out outapk -signature "%storepath%" "%storepass%" "%keypass%" "%alias%" -zipalign "%zipalign%"

pause

build_apk.sh:linux执行脚本，由于我使用的是windows系统，所以就。