Android教程網
  1. 首頁
  2. Android 技術
  3. Android 手機
  4. Android 系統教程
  5. Android 游戲
 Android教程網 >> Android技術 >> 關於Android編程 >> 解析android截屏問題

解析android截屏問題

編輯:關於Android編程

我是基於android2.3.3系統之上的,想必大家應該知道在android源碼下面有個文件叫做screencap吧,位於frameworks\base\services\surfaceflinger\tests\screencap\screencap.cpp,你直接在linux下編譯(保存在 /system/bin/test-screencap),然後push到手機上再通過電腦去敲命令test-screencap /mnt/sdcard/scapxx.png就可以實現截屏。
復制代碼 代碼如下:
/*
  * Copyright (C) 2010 The Android Open Source Project
  *
  * 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.
  */

 #include <utils/Log.h>

 #include <binder/IPCThreadState.h>
 #include <binder/ProcessState.h>
 #include <binder/IServiceManager.h>

 #include <binder/IMemory.h>
 #include <surfaceflinger/ISurfaceComposer.h>

 #include <SkImageEncoder.h>
 #include <SkBitmap.h>

 using namespace android;

 int main(int argc, char** argv)
 {
     if (argc != 2) {
         printf("usage: %s path\n", argv[0]);
         exit(0);
     }

     const String16 name("SurfaceFlinger");
     sp<ISurfaceComposer> composer;
     getService(name, &composer);

     sp<IMemoryHeap> heap;
     uint32_t w, h;
     PixelFormat f;
     status_t err = composer->captureScreen(0, &heap, &w, &h, &f, 0, 0);
     if (err != NO_ERROR) {
         fprintf(stderr, "screen capture failed: %s\n", strerror(-err));
         exit(0);
     }

     printf("screen capture success: w=%u, h=%u, pixels=%p\n",
             w, h, heap->getBase());

     printf("saving file as PNG in %s ...\n", argv[1]);

     SkBitmap b;
     b.setConfig(SkBitmap::kARGB_8888_Config, w, h);
     b.setPixels(heap->getBase());
     SkImageEncoder::EncodeFile(argv[1], b,
             SkImageEncoder::kPNG_Type, SkImageEncoder::kDefaultQuality);

     return 0;
 }

其實這個程序真正用到的就是一個叫做capturescreen的函數,而capturescreen會調用captureScreenImplLocked這個函數
下面是代碼:
復制代碼 代碼如下:
status_t SurfaceFlinger::captureScreenImplLocked(DisplayID dpy,
         sp<IMemoryHeap>* heap,
         uint32_t* w, uint32_t* h, PixelFormat* f,
         uint32_t sw, uint32_t sh)
 {
    LOGI("captureScreenImplLocked");
     status_t result = PERMISSION_DENIED;

     // only one display supported for now
     if (UNLIKELY(uint32_t(dpy) >= DISPLAY_COUNT))
         return BAD_VALUE;

     if (!GLExtensions::getInstance().haveFramebufferObject())
         return INVALID_OPERATION;

     // get screen geometry
     const DisplayHardware& hw(graphicPlane(dpy).displayHardware());
     const uint32_t hw_w = hw.getWidth();
     const uint32_t hw_h = hw.getHeight();

     if ((sw > hw_w) || (sh > hw_h))
         return BAD_VALUE;

     sw = (!sw) ? hw_w : sw;
     sh = (!sh) ? hw_h : sh;
     const size_t size = sw * sh * 4;

     // make sure to clear all GL error flags
     while ( glGetError() != GL_NO_ERROR ) ;

     // create a FBO
     GLuint name, tname;
     glGenRenderbuffersOES(1, &tname);
     glBindRenderbufferOES(GL_RENDERBUFFER_OES, tname);
     glRenderbufferStorageOES(GL_RENDERBUFFER_OES, GL_RGBA8_OES, sw, sh);
     glGenFramebuffersOES(1, &name);
     glBindFramebufferOES(GL_FRAMEBUFFER_OES, name);
     glFramebufferRenderbufferOES(GL_FRAMEBUFFER_OES,
             GL_COLOR_ATTACHMENT0_OES, GL_RENDERBUFFER_OES, tname);

     GLenum status = glCheckFramebufferStatusOES(GL_FRAMEBUFFER_OES);
     if (status == GL_FRAMEBUFFER_COMPLETE_OES) {

         // invert everything, b/c glReadPixel() below will invert the FB
         glViewport(0, 0, sw, sh);
         glScissor(0, 0, sw, sh);
         glMatrixMode(GL_PROJECTION);
         glPushMatrix();
         glLoadIdentity();
         glOrthof(0, hw_w, 0, hw_h, 0, 1);
         glMatrixMode(GL_MODELVIEW);

         // redraw the screen entirely...
         glClearColor(0,0,0,1);
         glClear(GL_COLOR_BUFFER_BIT);

         const Vector< sp<LayerBase> >& layers(mVisibleLayersSortedByZ);
         const size_t count = layers.size();
         for (size_t i=0 ; i<count ; ++i) {
             const sp<LayerBase>& layer(layers[i]);
             layer->drawForSreenShot();
         }

         // XXX: this is needed on tegra
         glScissor(0, 0, sw, sh);

         // check for errors and return screen capture
         if (glGetError() != GL_NO_ERROR) {
             // error while rendering
             result = INVALID_OPERATION;
         } else {
             // allocate shared memory large enough to hold the
             // screen capture
             sp<MemoryHeapBase> base(
                     new MemoryHeapBase(size, 0, "screen-capture") );
             void* const ptr = base->getBase();
             if (ptr) {
                 // capture the screen with glReadPixels()
                 glReadPixels(0, 0, sw, sh, GL_RGBA, GL_UNSIGNED_BYTE, ptr);
                 if (glGetError() == GL_NO_ERROR) {
                     *heap = base;
                     *w = sw;
                     *h = sh;
                     *f = PIXEL_FORMAT_RGBA_8888;
                     result = NO_ERROR;
                 }
             } else {
                 result = NO_MEMORY;
             }
         }
         glEnable(GL_SCISSOR_TEST);
         glViewport(0, 0, hw_w, hw_h);
         glMatrixMode(GL_PROJECTION);
         glPopMatrix();
         glMatrixMode(GL_MODELVIEW);

  
     } else {
         result = BAD_VALUE;
     }

     // release FBO resources
     glBindFramebufferOES(GL_FRAMEBUFFER_OES, 0);
     glDeleteRenderbuffersOES(1, &tname);
     glDeleteFramebuffersOES(1, &name);

     hw.compositionComplete();

     return result;
 }

  
 status_t SurfaceFlinger::captureScreen(DisplayID dpy,
         sp<IMemoryHeap>* heap,
         uint32_t* width, uint32_t* height, PixelFormat* format,
         uint32_t sw, uint32_t sh)
 {
            LOGI("into captureScreen");          
     // only one display supported for now
     if (UNLIKELY(uint32_t(dpy) >= DISPLAY_COUNT))
         return BAD_VALUE;

     if (!GLExtensions::getInstance().haveFramebufferObject())
         return INVALID_OPERATION;

     class MessageCaptureScreen : public MessageBase {
         SurfaceFlinger* flinger;
         DisplayID dpy;
         sp<IMemoryHeap>* heap;
         uint32_t* w;
         uint32_t* h;
         PixelFormat* f;
         uint32_t sw;
         uint32_t sh;
         status_t result;
     public:
         MessageCaptureScreen(SurfaceFlinger* flinger, DisplayID dpy,
                 sp<IMemoryHeap>* heap, uint32_t* w, uint32_t* h, PixelFormat* f,
                 uint32_t sw, uint32_t sh)
             : flinger(flinger), dpy(dpy),
               heap(heap), w(w), h(h), f(f), sw(sw), sh(sh), result(PERMISSION_DENIED)
         {

         }
         status_t getResult() const {
            LOGI("getResult()");          
          return result;
         }
         virtual bool handler() {

  
    LOGI("handler()");
             Mutex::Autolock _l(flinger->mStateLock);

             // if we have secure windows, never allow the screen capture
             if (flinger->mSecureFrameBuffer)
                 return true;

             result = flinger->captureScreenImplLocked(dpy,
                     heap, w, h, f, sw, sh);

             return true;
         }
     };
    LOGI("before messagecapturescreen");
     sp<MessageBase> msg = new MessageCaptureScreen(this,
             dpy, heap, width, height, format, sw, sh);
     status_t res = postMessageSync(msg);
     if (res == NO_ERROR) {
         res = static_cast<MessageCaptureScreen*>( msg.get() )->getResult();
     }
     return res;
 }

而這個函數關鍵又使用了opengl的幾個函數去獲得圖片,然而opengl又去read framebuffer(這是我的理解)。如果你去用jni調用so的方法去截屏的話,就可以把screencap這個文件稍微修改一下然後做成so文件。

主要是補充一下怎麼去存放文件與編譯吧,當然我說的方法只是我做的方法不代表是很好用的。

存放:在eclipse新建一個android工程,保存後找到這個工程(如screencap)的存放位置 然後把這個文件放到android源代碼的development文件裡面,然後在你的那個工程文件裡面新建一個文件夾,名字叫做jni(這個文件夾平行於src文件夾,screencap/jni),把上面博客提到的那個C++跟mk(screencap/jni/com_android_ScreenCap_ScreenCapNative.cpp和screencap/jni/Android.mk)文件放進去,最後在把編譯的mk文件放在screencap目錄下(screencap/Android.mk);

編譯:編譯是個很偉大的工程,需要你花大量的時間與精力。直接在終端進入工程存放的所在位置,我的是Administrator/Android.2.3.3/development,然後mm(Builds all of the modules in the current directory),如果成功,那麼你運氣比較好,在終端回提示你APK保存的位置。push進手機試一試。但是往往是不成功的。你可能會遇到一些問題,比如android.permission.ACCESS_SURFACE_FLINGER ,android.permission.READ_FRAME_BUFFER(因為capturescrren這個函數是surfaceflinger裡面的函數,然而surfaceflinger裡面的opengl截屏函數會去讀取framebuffer),相關源代碼是:
復制代碼 代碼如下:
status_t SurfaceFlinger::onTransact(
     uint32_t code, const Parcel& data, Parcel* reply, uint32_t flags)
 {

     switch (code) {
         case CREATE_CONNECTION:
         case OPEN_GLOBAL_TRANSACTION:
         case CLOSE_GLOBAL_TRANSACTION:
         case SET_ORIENTATION:
         case FREEZE_DISPLAY:
         case UNFREEZE_DISPLAY:
         case BOOT_FINISHED:
         case TURN_ELECTRON_BEAM_OFF:
         case TURN_ELECTRON_BEAM_ON:
         {
             // codes that require permission check
             IPCThreadState* ipc = IPCThreadState::self();
             const int pid = ipc->getCallingPid();
             const int uid = ipc->getCallingUid();
             if ((uid != AID_GRAPHICS) && !mAccessSurfaceFlinger.check(pid, uid)) {
                 LOGE("Permission Denial: "
                         "can't access SurfaceFlinger pid=%d, uid=%d", pid, uid);
                 return PERMISSION_DENIED;
             }
             break;
         }
         case CAPTURE_SCREEN:
         {
             // codes that require permission check
             IPCThreadState* ipc = IPCThreadState::self();
             const int pid = ipc->getCallingPid();
             const int uid = ipc->getCallingUid();
             if ((uid != AID_GRAPHICS) && !mReadFramebuffer.check(pid, uid)) {

                 LOGE("Permission Denial: "
                         "can't read framebuffer pid=%d, uid=%d", pid, uid);
                 return PERMISSION_DENIED;
             }

             break;
         }
     }

     status_t err = BnSurfaceComposer::onTransact(code, data, reply, flags);
     if (err == UNKNOWN_TRANSACTION || err == PERMISSION_DENIED) {
         CHECK_INTERFACE(ISurfaceComposer, data, reply);
         if (UNLIKELY(!mHardwareTest.checkCalling())) {
             IPCThreadState* ipc = IPCThreadState::self();
             const int pid = ipc->getCallingPid();
             const int uid = ipc->getCallingUid();

    LOGI("err");
             LOGE("Permission Denial: "
                     "can't access SurfaceFlinger pid=%d, uid=%d", pid, uid);
             return PERMISSION_DENIED;
         }
         int n;
         switch (code) {
             case 1000: // SHOW_CPU, NOT SUPPORTED ANYMORE
             case 1001: // SHOW_FPS, NOT SUPPORTED ANYMORE
                 return NO_ERROR;
             case 1002:  // SHOW_UPDATES
                 n = data.readInt32();
                 mDebugRegion = n ? n : (mDebugRegion ? 0 : 1);
                 return NO_ERROR;
             case 1003:  // SHOW_BACKGROUND
                 n = data.readInt32();
                 mDebugBackground = n ? 1 : 0;
                 return NO_ERROR;
             case 1004:{ // repaint everything
                 Mutex::Autolock _l(mStateLock);
                 const DisplayHardware& hw(graphicPlane(0).displayHardware());
                 mDirtyRegion.set(hw.bounds()); // careful that's not thread-safe
                 signalEvent();
                 return NO_ERROR;
             }
             case 1005:{ // force transaction
                 setTransactionFlags(eTransactionNeeded|eTraversalNeeded);
                 return NO_ERROR;
             }
             case 1006:{ // enable/disable GraphicLog
                 int enabled = data.readInt32();
                 GraphicLog::getInstance().setEnabled(enabled);
                 return NO_ERROR;
             }
             case 1007: // set mFreezeCount
                 mFreezeCount = data.readInt32();
                 mFreezeDisplayTime = 0;
                 return NO_ERROR;
             case 1010:  // interrogate.
                 reply->writeInt32(0);
                 reply->writeInt32(0);
                 reply->writeInt32(mDebugRegion);
                 reply->writeInt32(mDebugBackground);
                 return NO_ERROR;
             case 1013: {
                 Mutex::Autolock _l(mStateLock);
                 const DisplayHardware& hw(graphicPlane(0).displayHardware());
                 reply->writeInt32(hw.getPageFlipCount());
             }
             return NO_ERROR;
         }
     }
     return err;
 }

這個僅僅只是開始!  你會發現你即使在xml裡面添加相應的權限仍然會有這個問題出現,為什麼呢?在packageManger文件裡面發現相關代碼:
復制代碼 代碼如下:
 int checkSignaturesLP(Signature[] s1, Signature[] s2) {
        if (s1 == null) {
            return s2 == null
                    ? PackageManager.SIGNATURE_NEITHER_SIGNED
                    : PackageManager.SIGNATURE_FIRST_NOT_SIGNED;
        }
        if (s2 == null) {
            return PackageManager.SIGNATURE_SECOND_NOT_SIGNED;
        }
        HashSet<Signature> set1 = new HashSet<Signature>();
        for (Signature sig : s1) {
            set1.add(sig);
        }
        HashSet<Signature> set2 = new HashSet<Signature>();
        for (Signature sig : s2) {
            set2.add(sig);
        }
        // Make sure s2 contains all signatures in s1.
        if (set1.equals(set2)) {
            return PackageManager.SIGNATURE_MATCH;
        }
        return PackageManager.SIGNATURE_NO_MATCH;
    }

  

 // Check for shared user signatures
        if (pkgSetting.sharedUser != null && pkgSetting.sharedUser.signatures.mSignatures != null) {
            if (checkSignaturesLP(pkgSetting.sharedUser.signatures.mSignatures,
                    pkg.mSignatures) != PackageManager.SIGNATURE_MATCH) {
                Slog.e(TAG, "Package " + pkg.packageName
                        + " has no signatures that match those in shared user "
                        + pkgSetting.sharedUser.name + "; ignoring!");
                mLastScanError = PackageManager.INSTALL_FAILED_SHARED_USER_INCOMPATIBLE;
                return false;
            }
        }
        return true;

  
    private boolean verifySignaturesLP(PackageSetting pkgSetting,
            PackageParser.Package pkg) {

        // Check for shared user signatures
        if (pkgSetting.sharedUser != null && pkgSetting.sharedUser.signatures.mSignatures != null) {

        if (checkSignaturesLP(pkgSetting.sharedUser.signatures.mSignatures,
                    pkg.mSignatures) != PackageManager.SIGNATURE_MATCH) {

            Slog.e(TAG, "Package " + pkg.packageName
                        + " has no signatures that match those in shared user "
                        + pkgSetting.sharedUser.name + "; ignoring!");
                mLastScanError = PackageManager.INSTALL_FAILED_SHARED_USER_INCOMPATIBLE;
                return false;

            }
        }
        return true;
    }

你在終端輸入adb logcat | grep PackageManager 你會發現這兩個權限根本沒有賦予給你的apk,我的理解是,程序需要權限,然後apk仍然需要權限。那怎麼樣給apk賦予權限呢,兩個方法,一個是在我上面說的screencap/Android.mk裡面添加platform一行,然後在回到mm。還有一個方法就是通過sign。這兩個方法都是給apk賦予system權限,但是我試過這兩種方法,都有問題,就是在adb install的時候會顯示簽名不兼容,查看源代碼會發現uid跟gid不匹配。這些是我這段時間發現的問題,大家有問題可以交流交流。
再說說幾個簡單的應用層截屏吧,很簡單,就是幾個函數調用而已
復制代碼 代碼如下:
View view = getWindow().getDecorView();  
       Display display = this.getWindowManager().getDefaultDisplay();  
       view.layout(0, 0, display.getWidth(), display.getHeight());  
       view.setDrawingCacheEnabled(true);//允許當前窗口保存緩存信息,這樣  getDrawingCache()方法才會返回一個Bitmap  
       Bitmap bmp = Bitmap.createBitmap(view.getDrawingCache());

我對這個程序的理解就是,它僅僅只能截取當前的activity,也就是說如果你運行這個程序後它就截取你這個程序的當前屏幕的信息。我們假設你做成一個按鈕,在點擊按鈕後sleep5秒再調用這個方法(假設你的activity叫做screen)。當你點擊按鈕以後,然後你再點擊home或者返回按鈕,等到5秒後你那個程序就會截取到你當前屏幕?不是!它只會截取那個運行於後台的screen這個activity。

這些只是我的一點小小的總結,而且肯定有不對的地方,希望大家一起來解決截屏的問題!
  1. 上一頁:
  2. 下一頁:
熱門文章
閱讀排行版
Copyright © Android教程網 All Rights Reserved