Android 热修复方案Tinker(三) Dex补丁加载

93 阅读 0 评论 62 点赞

我是靠谱客的博主幽默皮卡丘，最近开发中收集的这篇文章主要介绍Android 热修复方案Tinker(三) Dex补丁加载，觉得挺不错的，现在分享给大家，希望可以做个参考。

概述

基于Tinker V1.7.5

Android 热修复方案Tinker(一) Application改造
Android 热修复方案Tinker(二) 补丁加载流程
Android 热修复方案Tinker(三) Dex补丁加载
Android 热修复方案Tinker(四) 资源补丁加载
Android 热修复方案Tinker(五) SO补丁加载
Android 热修复方案Tinker(六) Gradle插件实现
Android 热修复方案Tinker(七) 插桩实现
带注释的源码

之前有说到Tinker的修复原理是跟Qzone类似,这里就详细分析一下为什么这样做可以修复补丁.虽然其他Android版本的源码实现可能不一样,但是都是基于相同的原理.所以这里就以Android 6.0的源码为例介绍原理.具体每个系统版本的不同实现下面会详细说明.

首先从加载dex文件的入口开始看, /libcore/dalvik/src/main/java/dalvik/system/DexClassLoader.java这个类很简单,只是继承了BaseDexClassLoader在构造方法中调用了父类的构造方法.

public class DexClassLoader extends BaseDexClassLoader {
    public DexClassLoader(String dexPath, String optimizedDirectory,
            String libraryPath, ClassLoader parent) {
        super(dexPath, new File(optimizedDirectory), libraryPath, parent);
    }
}

继续进入/libcore/dalvik/src/main/java/dalvik/system/BaseDexClassLoader.java类中,BaseDexClassLoader在构造方法中创建出了一个很重要的对象pathList,至于为什么说他重要可以看下面的findclass方法.findclass方法是根据类名在运行时从dex文件中找出并将其返回回来,而真正的findclass是通过pathList对象的方法来操作的.

public class BaseDexClassLoader extends ClassLoader {
    private final DexPathList pathList;

    public BaseDexClassLoader(String dexPath, File optimizedDirectory,
            String libraryPath, ClassLoader parent) {
        super(parent);
        this.pathList = new DexPathList(this, dexPath, libraryPath, optimizedDirectory);
    }

    @Override
    protected Class<?> findClass(String name) throws ClassNotFoundException {
        List<Throwable> suppressedExceptions = new ArrayList<Throwable>();
        Class c = pathList.findClass(name, suppressedExceptions);
        if (c == null) {
            ClassNotFoundException cnfe = new ClassNotFoundException("Didn't find class "" + name + "" on path: " + pathList);
            for (Throwable t : suppressedExceptions) {
                cnfe.addSuppressed(t);
            }
            throw cnfe;
        }
        return c;
    }
}

最终在/libcore/dalvik/src/main/java/dalvik/system/DexPathList.java类中找到findclass方法,该方法是按顺序遍历dexElements,只要dexElement中的dex文件中包含有该class就加载出class然后直接return.所以利用findclass这种特性把补丁包dex插入dexElements的首位,系统在findClass的时候就优先拿到补丁包中的class,从而达到修复bug的目的.

final class DexPathList {

    private final Element[] dexElements;

    public Class findClass(String name, List<Throwable> suppressed) {
        for (Element element : dexElements) {
            DexFile dex = element.dexFile;

            if (dex != null) {
                Class clazz = dex.loadClassBinaryName(name, definingContext, suppressed);
                if (clazz != null) {
                    return clazz;
                }
            }
        }
        if (dexElementsSuppressedExceptions != null) {
            suppressed.addAll(Arrays.asList(dexElementsSuppressedExceptions));
        }
        return null;
    }
}

校验Dex文件

讲过dex修复的原理,回到Tinker的dex补丁加载流程.在loadTinkerJars之后,先确保之前checkComplete时是否筛选出物理有效的dex文件以供加载,再拿到PathClassLoader供后面使用.

if (dexList.isEmpty()) {
    Log.w(TAG, "there is no dex to load");
    return true;
}

PathClassLoader classLoader = (PathClassLoader) TinkerDexLoader.class.getClassLoader();
if (classLoader != null) {
    Log.i(TAG, "classloader: " + classLoader.toString());
} else {
    Log.e(TAG, "classloader is null");
    ShareIntentUtil.setIntentReturnCode(intentResult, ShareConstants.ERROR_LOAD_PATCH_VERSION_DEX_CLASSLOADER_NULL);
    return false;
}

在TinkerLoader.tryLoad时只是校验了dex_meta.txt文件的签名信息,并没有校验所有的dex文件的合法性.如果在ApplicationLike处配置了tinkerLoadVerifyFlag为true, 则每次加载dex补丁之前都对文件做MD5,并对比dex_meta.txt中对应的MD5信息.

for (DexDiffPatchInfo info : dexList) {
    //for dalvik, ignore art support dex
    if (isJustArtSupportDex(info)) {
        continue;
    }
    String path = dexPath + info.realName;
    File file = new File(path);

    if (tinkerLoadVerifyFlag) {
        long start = System.currentTimeMillis();
        String checkMd5 = isArtPlatForm ? info.destMd5InArt : info.destMd5InDvm;
        if (!PatchFileUtil.verifyDexFileMd5(file, checkMd5)) {
            //it is good to delete the mismatch file
            IntentUtil.setIntentReturnCode(intentResult, Constants.ERROR_LOAD_PATCH_VERSION_DEX_MD5_MISMATCH);
            intentResult.putExtra(IntentUtil.INTENT_PATCH_MISMATCH_DEX_PATH,
                    file.getAbsolutePath());
            return false;
        }
        Log.i(TAG, "verify dex file:" + file.getPath() + " md5, use time: " + (System.currentTimeMillis() - start));
    }
    legalFiles.add(file);
}

OTA在特定情况下重新load Dex

在v1.7.5的版本开始有了isSystemOTA判断,只要用户是ART环境并且做了OTA升级则在加载dex补丁的时候就会先把最近一次的补丁全部DexFile.loadDex一遍.这么做的原因是有些场景做了OTA后,oat的规则可能发生变化,在这种情况下去加载上个系统版本oat过的dex就会出现问题.

if (isSystemOTA) {
    parallelOTAResult = true;
    parallelOTAThrowable = null;
    Log.w(TAG, "systemOTA, try parallel oat dexes!!!!!");

    ParallelDexOptimizer.optimizeAll(
            legalFiles, optimizeDir,
            new ParallelDexOptimizer.ResultCallback() {
                @Override
                public void onSuccess(File dexFile, File optimizedDir) {
                    // Do nothing.
                }
                @Override
                public void onFailed(File dexFile, File optimizedDir, Throwable thr) {
                    parallelOTAResult = false;
                    parallelOTAThrowable = thr;
                }
            }
    );
    if (!parallelOTAResult) {
        Log.e(TAG, "parallel oat dexes failed");
        intentResult.putExtra(IntentUtil.INTENT_PATCH_EXCEPTION, parallelOTAThrowable);
        IntentUtil.setIntentReturnCode(intentResult, Constants.ERROR_LOAD_PATCH_VERSION_PARALLEL_DEX_OPT_EXCEPTION);
        return false;
    }
}

dex补丁的重置是在线程池中执行,并且利用CountDownLatch挂起主线程,直到线程池中的task都执行完毕再恢复主线程.在很极端的情况下可能会造成ANR.

private static boolean optimizeAllLocked(Collection<File> dexFiles, File optimizedDir, AtomicInteger successCount, ResultCallback cb) {
    final CountDownLatch lauch = new CountDownLatch(dexFiles.size());
    final ExecutorService threadPool = Executors.newCachedThreadPool();
    long startTick = System.nanoTime();
    for (File dexFile : dexFiles) {
        OptimizeWorker worker = new OptimizeWorker(dexFile, optimizedDir, successCount, lauch, cb);
        threadPool.submit(worker);
    }
    try {
        lauch.await();
        long timeCost = (System.nanoTime() - startTick) / 1000000;
        if (successCount.get() == dexFiles.size()) {
            Log.i(TAG, "All dexes are optimized successfully, cost: " + timeCost + " ms.");
            return true;
        } else {
            Log.e(TAG, "Dexes optimizing failed, some dexes are not optimized.");
            return false;
        }
    } catch (InterruptedException e) {
        Log.w(TAG, "Dex optimizing was interrupted.", e);
        return false;
    } finally {
        threadPool.shutdown();
    }
}

加载Dex

经过一系列的校验,终于到真正加载dex补丁的步骤了.Tinker加载dex补丁按照系统版本不同分成了四条分支.同样加载失败之后记录失败信息到intentResult中.

V4 Android SDK版本小于14
V14 Android SDK版本小于19
V19 Android SDK版本小于 23
V23 Android SDK版本大于等于23
- Android N 改造ClassLoader

try {
    SystemClassLoaderAdder.installDexes(application, classLoader, optimizeDir, legalFiles);
} catch (Throwable e) {
    Log.e(TAG, "install dexes failed");
    intentResult.putExtra(IntentUtil.INTENT_PATCH_EXCEPTION, e);
    IntentUtil.setIntentReturnCode(intentResult, Constants.ERROR_LOAD_PATCH_VERSION_DEX_LOAD_EXCEPTION);
    return false;
}
Log.i(TAG, "after loaded classloader: " + application.getClassLoader().toString());

V4 Android SDK版本小于14

在Android SDK4到14之间PathClassLoader.java的实现是直接继承自ClassLoader,findClass时是根据mFiles数组来遍历mDexs数组(类似于dexElements).从mDexs数组中的dex根据类名来加载Class,规则也是按照遍历的顺序加载,只要有加载出来的Class就直接return掉.

Android 2.3.6版本源码

public class PathClassLoader extends ClassLoader {

private final String path;

private final String[] mPaths;
private final File[] mFiles;
private final ZipFile[] mZips;
private final DexFile[] mDexs;

@Override
protected Class<?> findClass(String name) throws ClassNotFoundException
{
    //System.out.println("PathClassLoader " + this + ": findClass '" + name + "'");

    byte[] data = null;
    int length = mPaths.length;

    for (int i = 0; i < length; i++) {
        //System.out.println("My path is: " + mPaths[i]);

        if (mDexs[i] != null) {
            Class clazz = mDexs[i].loadClassBinaryName(name, this);
            if (clazz != null)
                return clazz;
        } else if (mZips[i] != null) {
            String fileName = name.replace('.', '/') + ".class";
            data = loadFromArchive(mZips[i], fileName);
        } else {
            File pathFile = mFiles[i];
            if (pathFile.isDirectory()) {
                String fileName =
                    mPaths[i] + "/" + name.replace('.', '/') + ".class";
                data = loadFromDirectory(fileName);
            } else {
                //System.out.println("PathClassLoader: can't find '"
                //    + mPaths[i] + "'");
            }

        }
    }

    throw new ClassNotFoundException(name + " in loader " + this);
}

在DexClassLoader.java的构造方法中可以看到path,mPaths,mFiles, mZips和mDexs五个关键属性之间是互相联系的,所以在做热修复时要同时对这五个属性同步操作,来确保数据的一致性.

this.path = path;
this.libPath = libPath;

mPaths = path.split(":");
int length = mPaths.length;

//System.out.println("PathClassLoader: " + mPaths);
mFiles = new File[length];
mZips = new ZipFile[length];
mDexs = new DexFile[length];

...

/* open all Zip and DEX files up front */
for (int i = 0; i < length; i++) {
    //System.out.println("My path is: " + mPaths[i]);
    File pathFile = new File(mPaths[i]);
    mFiles[i] = pathFile;

    if (pathFile.isFile()) {
        try {
            mZips[i] = new ZipFile(pathFile);
        }
        catch (IOException ioex) {
        }
        if (wantDex) {
            /* we need both DEX and Zip, because dex has no resources */
            try {
                mDexs[i] = new DexFile(pathFile);
            }
            catch (IOException ioex) {}
        }
    }
}

所以在Tinker中,要加载这类系统的补丁包最核心的地方就是path,mPaths,mFiles, mZips和mDexs五个属性的的操作.根据补丁文件的个数建立四个关键属性对应的数组,再通过遍历补丁文件,对四个数组和一个字符串进行填充.再利用反射将新的数组插入到原数组头部,完成补丁加载的过程.

private static void install(ClassLoader loader, List<File> additionalClassPathEntries, File optimizedDirectory)
        throws IllegalArgumentException, IllegalAccessException,
        NoSuchFieldException, IOException {
    int extraSize = additionalClassPathEntries.size();

    Field pathField = ReflectUtil.findField(loader, "path");

    StringBuilder path = new StringBuilder((String) pathField.get(loader));
    String[] extraPaths = new String[extraSize];
    File[] extraFiles = new File[extraSize];
    ZipFile[] extraZips = new ZipFile[extraSize];
    DexFile[] extraDexs = new DexFile[extraSize];

    for (ListIterator<File> iterator = additionalClassPathEntries.listIterator();
         iterator.hasNext();) {
        File additionalEntry = iterator.next();
        String entryPath = additionalEntry.getAbsolutePath();
        path.append(':').append(entryPath);
        int index = iterator.previousIndex();
        extraPaths[index] = entryPath;
        extraFiles[index] = additionalEntry;
        extraZips[index] = new ZipFile(additionalEntry);
        //edit by zhangshaowen
        String outputPathName = PatchFileUtil.optimizedPathFor(additionalEntry, optimizedDirectory);
        //for below 4.0, we must input jar or zip
        extraDexs[index] = DexFile.loadDex(entryPath, outputPathName, 0);
    }

    pathField.set(loader, path.toString());
    ReflectUtil.expandFieldArray(loader, "mPaths", extraPaths);
    ReflectUtil.expandFieldArray(loader, "mFiles", extraFiles);
    ReflectUtil.expandFieldArray(loader, "mZips", extraZips);
    try {
        ReflectUtil.expandFieldArray(loader, "mDexs", extraDexs);
    } catch (Exception e) {

    }
}

对原数组的操作是利用反射,先拿到原数组的对象original, 再根据original的类型长度以及补丁数组的长度重新创建出一个新数组combined.接下来使用arraycopy将补丁数组和原数组copy到combined中,最后将该数组赋值给filedName对应的属性.

public static void expandFieldArray(Object instance, String fieldName, Object[] extraElements)
        throws NoSuchFieldException, IllegalArgumentException, IllegalAccessException {
    Field jlrField = findField(instance, fieldName);

    Object[] original = (Object[]) jlrField.get(instance);
    Object[] combined = (Object[]) Array.newInstance(original.getClass().getComponentType(), original.length + extraElements.length);

    // NOTE: changed to copy extraElements first, for patch load first

    System.arraycopy(extraElements, 0, combined, 0, extraElements.length);
    System.arraycopy(original, 0, combined, extraElements.length, original.length);

    jlrField.set(instance, combined);
}

V14 Android SDK版本小于19

在这个Android版本的区间内不再像老版本的那样要维护四个数组,源码从中抽离出了一个类DexPathList.java,加载dex的关键数组也变成了dexElements,并且dexElements是根据makeDexElements方法生成的.对比过源码其实就可以发现dexElements其实就是老版本中mFiles, mZips和mDexs的封装,makeDexElements方法就是老版本DexClassLoader.java构造方法中对数组初始化的动作.

Android 4.2.2版本源码

final class DexPathList {
    /** list of dex/resource (class path) elements */
    private final Element[] dexElements;


    public Class findClass(String name) {
        for (Element element : dexElements) {
            DexFile dex = element.dexFile;

            if (dex != null) {
                Class clazz = dex.loadClassBinaryName(name, definingContext);
                if (clazz != null) {
                    return clazz;
                }
            }
        }

        return null;
    }

    /**
     * Makes an array of dex/resource path elements, one per element of
     * the given array.
     */
    private static Element[] makeDexElements(ArrayList<File> files,
            File optimizedDirectory) {
        ArrayList<Element> elements = new ArrayList<Element>();

        /*
         * Open all files and load the (direct or contained) dex files
         * up front.
         */
        for (File file : files) {
            File zip = null;
            DexFile dex = null;
            String name = file.getName();

            if (name.endsWith(DEX_SUFFIX)) {
                // Raw dex file (not inside a zip/jar).
                try {
                    dex = loadDexFile(file, optimizedDirectory);
                } catch (IOException ex) {
                    System.logE("Unable to load dex file: " + file, ex);
                }
            } else if (name.endsWith(APK_SUFFIX) || name.endsWith(JAR_SUFFIX)
                    || name.endsWith(ZIP_SUFFIX)) {
                zip = file;

                try {
                    dex = loadDexFile(file, optimizedDirectory);
                } catch (IOException ignored) {
                    /*
                     * IOException might get thrown "legitimately" by
                     * the DexFile constructor if the zip file turns
                     * out to be resource-only (that is, no
                     * classes.dex file in it). Safe to just ignore
                     * the exception here, and let dex == null.
                     */
                }
            } else {
                System.logW("Unknown file type for: " + file);
            }

            if ((zip != null) || (dex != null)) {
                elements.add(new Element(file, zip, dex));
            }
        }

        return elements.toArray(new Element[elements.size()]);
    }
}

系统既然自己做了封装,那么我们反射调用起来也会更方便.首先反射拿到反射得到PathClassLoader中的pathList对象,再将补丁文件通过反射调用makeDexElements得到补丁文件的Element[],再将补丁包的Element数组插入到dexElements中,方法如V4.完成补丁加载.

private static void install(ClassLoader loader, List<File> additionalClassPathEntries,
                            File optimizedDirectory)
        throws IllegalArgumentException, IllegalAccessException,
        NoSuchFieldException, InvocationTargetException, NoSuchMethodException {
    /* The patched class loader is expected to be a descendant of
     * dalvik.system.BaseDexClassLoader. We modify its
     * dalvik.system.DexPathList pathList field to append additional DEX
     * file entries.
     */

    Field pathListField = ReflectUtil.findField(loader, "pathList");
    Object dexPathList = pathListField.get(loader);
    //通过反射调用makeDexElements方法生成补丁包的dex数组,再将其插入到dexElements的头部
    ReflectUtil.expandFieldArray(dexPathList, "dexElements", makeDexElements(dexPathList,
            new ArrayList<File>(additionalClassPathEntries), optimizedDirectory));
}

/**
 * A wrapper around
 * {@code private static final dalvik.system.DexPathList#makeDexElements}.
 */
private static Object[] makeDexElements(
        Object dexPathList, ArrayList<File> files, File optimizedDirectory)
        throws IllegalAccessException, InvocationTargetException,
        NoSuchMethodException {
    Method makeDexElements =
            ReflectUtil.findMethod(dexPathList, "makeDexElements", ArrayList.class, File.class);

    //反射调用makeDexElements方法根据files得到新dexElements数组
    return (Object[]) makeDexElements.invoke(dexPathList, files, optimizedDirectory);
}

V19 Android SDK版本小于 23

在该版本系统区间中,加载补丁涉及到的修改只是增加了一个exElementsSuppressedExceptions异常数组的维护.所以在加载补丁的时候就跟V14差不多了.既然只是多了一个异常的管理,为什么Tinker源码在利用反射找makeDexElements(ArrayList,File,ArrayList),如果找不到就接着找makeDexElements(List,File,List)？为了在Android源码中找到答案我去查找了4.4, 5.0,5.1版本的DexPathList源码,发现方法的参数都是ArrayList,根本没有List.百思不得姐之后就问了一下Tinker的作者,他们说在线上发现有机子的rom中这个方法的参数就是List.

private static Object[] makeDexElements(
        Object dexPathList, ArrayList<File> files, File optimizedDirectory,
        ArrayList<IOException> suppressedExceptions)
        throws IllegalAccessException, InvocationTargetException, NoSuchMethodException {

    Method makeDexElements = null;
    try {
        makeDexElements = ReflectUtil.findMethod(dexPathList, "makeDexElements", ArrayList.class, File.class,
                ArrayList.class);
    } catch (NoSuchMethodException e) {
        Log.e(TAG, "NoSuchMethodException: makeDexElements(ArrayList,File,ArrayList) failure");
        try {
            makeDexElements = ReflectUtil.findMethod(dexPathList, "makeDexElements", List.class, File.class, List.class);
        } catch (NoSuchMethodException e1) {
            Log.e(TAG, "NoSuchMethodException: makeDexElements(List,File,List) failure");
            throw e1;
        }
    }

    return (Object[]) makeDexElements.invoke(dexPathList, files, optimizedDirectory, suppressedExceptions);
}

V23 Android SDK版本大于等于23

因为V23中包含有Android 7.0的系统版本,由于Android N混合编译与对热补丁影响解析,这会造成要修复的class被缓存在App image中,App image中的class会插入PathClassLoader中,而PathClassLoader 加载补丁的时候不会替换缓存的class,最终会导致在全量更新的情况下有可能部分类是从base.apk中加载,部分类是从patch.dex中加载,抛出IllegalAccessError.Tinker的解决方案是在运行时改写PathClassLoader来加载类,让App image中的缓存失效.

所以要解决N里面混编的问题,核心着手点就是要替换PathClassLoader使他在加载dex的时候不加载做过优化的dex文件,重新加载原始的dex文件.这个点要从哪里切入呢? 在Android 7.0的源码中定位到了在makePathElements方法中调用的loadDexFile方法.从代码上来看是要在调用的时候有传递有效的optimizedDirectory参数,就会去opt过的路径下加载dex文件.所以我们在调用的时候不传optimizedDirectory参数就可以达到重新加载原始dex文件从而去除混编优化的目的.

private static DexFile loadDexFile(File file, File optimizedDirectory, ClassLoader loader,
                                   Element[] elements)
        throws IOException {
    if (optimizedDirectory == null) {
        return new DexFile(file, loader, elements);
    } else {
        String optimizedPath = optimizedPathFor(file, optimizedDirectory);
        return DexFile.loadDex(file.getPath(), optimizedPath, 0, loader, elements);
    }
}

知道了解决方案和切入点,接下来分析一下Tinker做法.在加载补丁之前利用反射替换原PathClassLoader以及与它相关的所有引用.首先根据原PathClassLoader的parent 构建出AndroidNClassLoader;再反射拿到original的pathList;接着反射拿到pathList对象的definingContext属性,因为该属性是original的引用,需要拿到之后替换成新loader的引用;继续反射拿到androidNClassLoader的pathList对象,并且替换成original的;再反射拿到original的pathList的dexElements,并且遍历出dexElements中真实的dex文件名之后存储起来;接下来反射拿到original的pathList的makePathElements方法并调用注意方法第二个参数optDir要设置为null,重新生成dexElements数组,并替换原来的数组.最终完成AndroidNClassLoader的创建,以及子类引用的替换.

private static AndroidNClassLoader createAndroidNClassLoader(PathClassLoader original) throws Exception {
    //let all element ""
    AndroidNClassLoader androidNClassLoader = new AndroidNClassLoader("",  original);
    Field originPathList = ShareReflectUtil.findField(original, "pathList");
    Object originPathListObject = originPathList.get(original);
    //should reflect definingContext also
    Field originClassloader = ShareReflectUtil.findField(originPathListObject, "definingContext");
    originClassloader.set(originPathListObject, androidNClassLoader);
    //copy pathList
    Field pathListField = ShareReflectUtil.findField(androidNClassLoader, "pathList");
    //just use PathClassloader's pathList
    pathListField.set(androidNClassLoader, originPathListObject);

    //we must recreate dexFile due to dexCache
    List<File> additionalClassPathEntries = new ArrayList<>();
    Field dexElement = ShareReflectUtil.findField(originPathListObject, "dexElements");
    Object[] originDexElements = (Object[]) dexElement.get(originPathListObject);
    for (Object element : originDexElements) {
        DexFile dexFile = (DexFile) ShareReflectUtil.findField(element, "dexFile").get(element);
        additionalClassPathEntries.add(new File(dexFile.getName()));
        //protect for java.lang.AssertionError: Failed to close dex file in finalizer.
        oldDexFiles.add(dexFile);
    }
    Method makePathElements = ShareReflectUtil.findMethod(originPathListObject, "makePathElements", List.class, File.class,
        List.class);
    ArrayList<IOException> suppressedExceptions = new ArrayList<>();
    Object[] newDexElements = (Object[]) makePathElements.invoke(originPathListObject, additionalClassPathEntries, null, suppressedExceptions);
    dexElement.set(originPathListObject, newDexElements);
    return androidNClassLoader;
}

做完新AndroidNClassLoader的创建之后就是替换真正的ClassLoader的引用了.在全局Context中持有的LoadedApk的对象mPackageInfo的属性中,有一个ClassLoader类的对象mClassLoader.层层反射将mClassLoader的引用替换为上面创建出来的AndroidNClassLoader对象.同时将Thread中持有的ClassLoader也同步替换为AndroidNClassLoader.至此PathClassLoader的修改和替换都已经完成了,接下来就可以正常得加载补丁dex了.

String defBase = "mBase";
String defPackageInfo = "mPackageInfo";
String defClassLoader = "mClassLoader";

Context baseContext = (Context) ReflectUtil.findField(application, defBase).get(application);
Object basePackageInfo = ReflectUtil.findField(baseContext, defPackageInfo).get(baseContext);
Field classLoaderField = ReflectUtil.findField(basePackageInfo, defClassLoader);
Thread.currentThread().setContextClassLoader(reflectClassLoader);
classLoaderField.set(basePackageInfo, reflectClassLoader);

在Android系统在该版本区间之内时,DexPathList类中的findclass方法跟V19相比是没有变化的.但是生成dexElements数组用的方法名发生了变化.所以在这个版本中反射生成补丁包的Element[]就需要兼容这些变化.

Android 6.0.0版本源码, 相比老版本makeDexElements(ArrayList,File,ArrayList)方法变成了makePathElements(List,File,List).

 /**
 * Makes an array of dex/resource path elements, one per element of
 * the given array.
 */
private static Element[] makePathElements(List<File> files, File optimizedDirectory,
                                          List<IOException> suppressedExceptions) {
    ···
}

Android 7.0.0版本源码,该方法名又发生了变化.根据职能做了一些区分和重载.

 /**
 * Makes an array of dex/resource path elements, one per element of
 * the given array.
 */
private static Element[] makeDexElements(List<File> files, File optimizedDirectory,
                                         List<IOException> suppressedExceptions,
                                         ClassLoader loader) {
    return makeElements(files, optimizedDirectory, suppressedExceptions, false, loader);
}

/**
 * Makes an array of directory/zip path elements, one per element of the given array.
 */
private static Element[] makePathElements(List<File> files,
                                          List<IOException> suppressedExceptions,
                                          ClassLoader loader) {
    return makeElements(files, null, suppressedExceptions, true, loader);
}

private static Element[] makePathElements(List<File> files, File optimizedDirectory,
                                          List<IOException> suppressedExceptions) {
    return makeElements(files, optimizedDirectory, suppressedExceptions, false, null);
}

private static Element[] makeElements(List<File> files, File optimizedDirectory,
                                      List<IOException> suppressedExceptions,
                                      boolean ignoreDexFiles,
                                      ClassLoader loader) {
    ···
}

结合上面两个系统版本的分析,在反射findMethod时做多种情况的处理,同时也避免V19中描述的rom问题.

private static Object[] makePathElements(
        Object dexPathList, ArrayList<File> files, File optimizedDirectory,
        ArrayList<IOException> suppressedExceptions)
        throws IllegalAccessException, InvocationTargetException, NoSuchMethodException {

    Method makePathElements;
    try {
        makePathElements = ReflectUtil.findMethod(dexPathList, "makePathElements", List.class, File.class,
                List.class);
    } catch (NoSuchMethodException e) {
        Log.e(TAG, "NoSuchMethodException: makePathElements(List,File,List) failure");
        try {
            makePathElements = ReflectUtil.findMethod(dexPathList, "makePathElements", ArrayList.class, File.class, ArrayList.class);
        } catch (NoSuchMethodException e1) {
            Log.e(TAG, "NoSuchMethodException: makeDexElements(ArrayList,File,ArrayList) failure");
            try {
                Log.e(TAG, "NoSuchMethodException: try use v19 instead");
                return V19.makeDexElements(dexPathList, files, optimizedDirectory, suppressedExceptions);
            } catch (NoSuchMethodException e2) {
                Log.e(TAG, "NoSuchMethodException: makeDexElements(List,File,List) failure");
                throw e2;
            }
        }
    }

    return (Object[]) makePathElements.invoke(dexPathList, files, optimizedDirectory, suppressedExceptions);
}

卸载 Dex补丁

校验和加载Dex补丁的流程都已经分析完了.接下来就是验证补丁是否加载成功,这里是通过反射拿到TinkerTestDexLoad.isPatch静态属性来判断.在没有补丁加载的情况下都是返回false的,在补丁中修改isPatch属性为true.所以只要反射拿到isPatch的属性为true就说明补丁已经成功加载进来了.如果返回false则卸载Dex补丁恢复到加载之前的数据状态.

private static boolean checkDexInstall(ClassLoader classLoader) throws ClassNotFoundException, NoSuchFieldException, IllegalAccessException {
    Class<?> clazz = Class.forName(CHECK_DEX_CLASS, true, classLoader);
    Field filed = ShareReflectUtil.findField(clazz, CHECK_DEX_FIELD);
    boolean isPatch = (boolean) filed.get(null);
    Log.w(TAG, "checkDexInstall result:" + isPatch);
    return isPatch;
}

补丁的卸载就不需要跟加载一样要区分那么多版本,因为Android 4.0开始就都是在操作DexPathList类中的dexElements数组,Android 4.0之前是在V4补丁加载中提到过的那四个数组.卸载就根据补丁的数量把数组头部的数据都移除掉.其中reduceFieldArray的作用跟加载补丁V4中提到的expandFieldArray完全相反.

public static void uninstallPatchDex(ClassLoader classLoader) throws Throwable {
    if (sPatchDexCount <= 0) {
        return;
    }
    if (Build.VERSION.SDK_INT >= 14) {
        Field pathListField = ShareReflectUtil.findField(classLoader, "pathList");
        Object dexPathList = pathListField.get(classLoader);
        ShareReflectUtil.reduceFieldArray(dexPathList, "dexElements", sPatchDexCount);
    } else {
        ShareReflectUtil.reduceFieldArray(classLoader, "mPaths", sPatchDexCount);
        ShareReflectUtil.reduceFieldArray(classLoader, "mFiles", sPatchDexCount);
        ShareReflectUtil.reduceFieldArray(classLoader, "mZips", sPatchDexCount);
        try {
            ShareReflectUtil.reduceFieldArray(classLoader, "mDexs", sPatchDexCount);
        } catch (Exception e) {
        }
    }
}