腾讯性能检测库Matrix源码阅读

作者: 普通的程序员 | 来源:发表于2019-04-15 16:27 被阅读44次

0.前面随便写点什么

1.初始化

  Matrix.Builder builder = new Matrix.Builder(application);
 // build matrix
  builder.patchListener(new TestPluginListener(this)); 
// add general pluginListener
  DynamicConfigImplDemo dynamicConfig = new DynamicConfigImplDemo();
 // dynamic config
  
  // init plugin 
  IOCanaryPlugin ioCanaryPlugin = new IOCanaryPlugin(new IOConfig.Builder()
                    .dynamicConfig(dynamicConfig)
                    .build());
  //add to matrix               
  builder.plugin(ioCanaryPlugin); 
  //init matrix
  Matrix.init(builder.build());
  // start plugin 
  ioCanaryPlugin.start();

建造者模式，内部类Builder
patchListener就是监听每一个patch的生命周期
DynamicConfigImplDemo 实际是IDynamicConfig的实现类，用于动态设置参数，里面也有个matrix-config.xml是参数配置文件。

Plugin后续再说。

Matrix内部维护了一个HashSet

 public Builder plugin(Plugin plugin) {
            if (plugins == null) {
                plugins = new HashSet<>();
            }
            String tag = plugin.getTag();
            for (Plugin exist : plugins) {
                if (tag.equals(exist.getTag())) {
                    throw new RuntimeException(String.format("plugin with tag %s is already exist", tag));
                }
            }
            plugins.add(plugin);
            return this;
        }

先用tag做判断，再者HashSet也不允许有重复的值存在，也就是Matrix里同一类型的plugin只能有一个。

接下来就是

public Matrix build() {
            if (pluginListener == null) {
                pluginListener = new DefaultPluginListener(application);
            }
            return new Matrix(application, pluginListener, plugins);
        }

private Matrix(Application app, PluginListener listener, HashSet<Plugin> plugins) {
        this.application = app;
        this.pluginListener = listener;
        this.plugins = plugins;
        for (Plugin plugin : plugins) {
            plugin.init(application, pluginListener);
            pluginListener.onInit(plugin);
        }
    }

public static Matrix init(Matrix matrix) {
        if (matrix == null) {
            throw new RuntimeException("Matrix init, Matrix should not be null.");
        }
        synchronized (Matrix.class) {
            if (sInstance == null) {
                sInstance = matrix;
            } else {
                MatrixLog.e(TAG, "Matrix instance is already set. this invoking will be ignored");
            }
        }
        return sInstance;
    }

至此，完成了Matrix装饰类的初始化。

其余的每一个Plugin插件初始化斗差不多，就用demo里的IOPlugin
每一个Plugin像glide那样，传入一个动态参数。
在Matrix的构造方法里，plugin依次调用init方法，plugin的init方法初始化PluginCore核心类，最终plugin的start方法也是core核心类的start方法。

2.内存监控MemoryCanaryPlugin

初始化多了一个HashSet维护的activity集合

MemoryCanaryPlugin类
public void init(Application app, PluginListener listener) {
        super.init(app, listener);
        mCore = new MemoryCanaryCore(this);
        HashSet<String> specialActivities = mMemoryConfig.getSpecialActivities();
        if (specialActivities != null) {
            for (String activity : specialActivities) {
                mCore.addSpecial(activity);
            }
        }
    }

MemoryConfig类
public HashSet<String> getSpecialActivities() {
        String strActivities = mDynamicConfig.get(IDynamicConfig.ExptEnum.clicfg_matrix_memory_special_activities.name(), DEFAULT_SPECIAL_ACTIVITIES);
        if (strActivities == null || "".equals(strActivities)) {
            return null;
        }

        String[] vecActivities = strActivities.split(";");
        HashSet<String> result = new HashSet<>();
        for (String activity : vecActivities) {
            if ("".equals(activity)) {
                continue;
            }

            result.add(activity);
        }
        return result;
    }

重点关注MemoryCanaryCore类
start方法首先检测机器条件是否允许，获取内存信息，注册生命周期回调callback----添加到set里进行管理，还有组件回调---在合适的时候dumpMem信息

public void start() {
        mLevel = DeviceUtil.getLevel(mContext);
        if (!isApiLevelOk() || mLevel == DeviceUtil.LEVEL.LOW || mLevel == DeviceUtil.LEVEL.BAD || mLevel == DeviceUtil.LEVEL.UN_KNOW) {
            mIsOpen = false;
            return;
        }

        mIsOpen = true;

        //all in kb
        mTotalMemory = DeviceUtil.getTotalMemory(mContext) / 1024;
        mLowMemoryThreshold = DeviceUtil.getLowMemoryThresold(mContext) / 1024;
        mMemoryClass = DeviceUtil.getMemoryClass(mContext);
        if (mLowMemoryThreshold >= mTotalMemory || mLowMemoryThreshold == 0) {
            mIsOpen = false;
        }

        ((Application) mContext).registerActivityLifecycleCallbacks(mActivityLifecycleCallback);
        mContext.registerComponentCallbacks(mComponentCallback);
    }

private final Application.ActivityLifecycleCallbacks mActivityLifecycleCallback = new Application.ActivityLifecycleCallbacks() {
        @Override
        public void onActivityCreated(Activity activity, Bundle bundle) {
            if (!mActivityMap.containsKey(activity.getClass().hashCode())) {
                mActivityMap.put(activity.getClass().hashCode(), activity.getClass().getSimpleName());
                if (mSpecialActivityName.contains(activity.getClass().getSimpleName())) {
                    mSpecialActivity.add(activity.getClass().hashCode());
                }
            }

            MatrixLog.d(TAG, "activity create:" + activity.getClass().getSimpleName());
        }
@Override
        public void onActivityStarted(final Activity activity) {
            onShow(activity);
        }
.....//关键代码
    };

private final ComponentCallbacks2 mComponentCallback = new ComponentCallbacks2() {
        @Override
        public void onTrimMemory(final int i) {
            switch (i) {
                case TRIM_MEMORY_RUNNING_CRITICAL:
                    break;
                case TRIM_MEMORY_COMPLETE:
                    break;
                default:
                    return;
            }

            mHandler.post(new Runnable() {
                @Override
                public void run() {
                    detectAppMemoryInfo(mShowingActivity, true, i);
                }
            });
        }
.....//关键代码
     @Override
        public void onLowMemory() {
            mHandler.post(new Runnable() {
                @Override
                public void run() {
                    detectAppMemoryInfo(mShowingActivity, true, LOW_MEMORY_TRIM);
                }
            });
        }
    };

detectAppMemoryInfo这个我就不多说了
在activity生命周期的activityStart里面有个onShow方法
问题：
没有太懂，为什么对activity做了个IFragmentActivity的判断？
这个IFragmentActivity是一个Interface，也没有看见哪个activity实现了这个接口。

detectAppMemoryInfo之后，最终是调用detectAppMemoryInfoImpl
并且会触发onDetectIssue回调。
最终由pluginListener.onReportIssue(issue);把内存信息发送出去。

MemoryCanaryPlugin没有做内存泄露的检测，可能是没必要重复造轮子吧。

3.性能监控TracePlugin

这个监控应该是四个角度

    TracePlugin类的四个成员变量
    private FPSTracer mFPSTracer;//帧率
    private EvilMethodTracer mEvilMethodTracer;//应该是耗时方法
    private FrameTracer mFrameTracer;//帧数
    private StartUpTracer mStartUpTracer;//启动速度

初始化

public void init(Application app, PluginListener listener) {
        super.init(app, listener);
        MatrixLog.i(TAG, "trace plugin init, trace config: %s", mTraceConfig.toString());
        if (Build.VERSION.SDK_INT < Build.VERSION_CODES.JELLY_BEAN) {
            MatrixLog.e(TAG, "[FrameBeat] API is low Build.VERSION_CODES.JELLY_BEAN(16), TracePlugin is not supported");
            unSupportPlugin();
            return;
        }

        ApplicationLifeObserver.init(app);

        mFrameTracer = new FrameTracer(this);//帧数检测是默认开启的
        if (mTraceConfig.isMethodTraceEnable()) {
            mStartUpTracer = new StartUpTracer(this, mTraceConfig);
        }
        if (mTraceConfig.isFPSEnable()) {
            mFPSTracer = new FPSTracer(this, mTraceConfig);
        }

        if (mTraceConfig.isMethodTraceEnable()) {
            mEvilMethodTracer = new EvilMethodTracer(this, mTraceConfig);
        }
    }

ApplicationLifeObserver实现了ActivityLifecycleCallbacks接口，内部是监控activity是否在前台。
start后，重点关注FrameBeat.getInstance().onCreate();
然后先看mFPSTracer.onCreate();

public void start() {
        super.start();
        if (!isSupported()) {
            return;
        }

        new Handler(Looper.getMainLooper()).post(new Runnable() {
            @Override
            public void run() {
                FrameBeat.getInstance().onCreate();
            }
        });

        if (null != mFPSTracer) {
            mFPSTracer.onCreate();
        }
      //....省略
    }

实际上呢，这个FrameBeat就是借助了Choreographer.FrameCallback
监控相邻两次 Vsync 事件通知的时间差
(360的开源方案https://github.com/Qihoo360/ArgusAPM)

微信自己的公众号也说了
https://mp.weixin.qq.com/s/W4-1tfepKg2XMYvVn62B-Q

确实在FrameBeat类，在activity的Resume时间注册这个回调，在pause反注册，doFrame里完成事件的记录。

public void doFrame(long frameTimeNanos) {
        if (isPause) {
            return;
        }
        if (frameTimeNanos < mLastFrameNanos || mLastFrameNanos <= 0) {
            mLastFrameNanos = frameTimeNanos;
            if (null != mChoreographer) {
                mChoreographer.postFrameCallback(this);
            }
            return;
        }

        if (null != mFrameListeners) {

            for (IFrameBeatListener listener : mFrameListeners) {
                listener.doFrame(mLastFrameNanos, frameTimeNanos);
            }

            if (null != mChoreographer) {
                mChoreographer.postFrameCallback(this);
            }

            mLastFrameNanos = frameTimeNanos;
        }

    }

由IFrameBeatListener完成回调动作，BaseTracer实现了IFrameBeatListener接口，但是内部是一个空实现，最终由子类FrameTracer来实现真正的统计方法。

public void doFrame(final long lastFrameNanos, final long frameNanos) {
        if (!isDrawing) {
            return;
        }
        isDrawing = false;
        final int droppedCount = (int) ((frameNanos - lastFrameNanos) / REFRESH_RATE_MS);
        for (final IDoFrameListener listener : mDoFrameListenerList) {
            listener.doFrameSync(lastFrameNanos, frameNanos, getScene(), droppedCount);
            if (null != listener.getHandler()) {
                listener.getHandler().post(new AsyncDoFrameTask(listener,
                        lastFrameNanos, frameNanos, getScene(), droppedCount));
            }
        }
    }

continue...

腾讯性能检测库Matrix源码阅读

目录

0.前面随便写点什么

1.初始化

2.内存监控MemoryCanaryPlugin

3.性能监控TracePlugin

相关文章

网友评论

延伸阅读

深度阅读

栏目导航

热点阅读

优化