Jean-Baptiste Queru wrote: > [still bcc android-platform, android-framework, android-porting] > > We're now one step closer. Change 9356 was submitted, so that the > source tree "as is" can be compiled. > > You still need to set BUILD_WITHOUT_PV=true. > > JBQ > > On Sat, Mar 21, 2009 at 10:35 PM, Jean-Baptiste Queru <j...@android.com> > wrote: > >> [still bcc android-platform, android-framework, android-porting] >> >> Latest status: >> >> -we're back to one patch: 9356. >> >> -build with "BUILD_WITHOUT_PV=true make" >> >> -quickly tested both on emulator and dream, and seems to work well >> enough to not have the phone app crash in a loop. >> >> -I expect to submit 9356 on Monday morning PDT. >> >> JBQ >> >> On Sat, Mar 21, 2009 at 11:50 AM, Jean-Baptiste Queru <j...@android.com> >> wrote: >> >>> I've been working on a "better" batch of patches (now it feels more >>> like open-heart surgery with a meat cleaver): >>> >>> -no need to delete the opencore directory or to remove it from the manifest. >>> >>> -you need to repo download changes 9355, 9356 and 9357. No need to >>> take 9300. I know it's 3 changes instead of 1, because I had to touch >>> a few more parts of the system, but those are much cleaner. >>> >>> -only tested on the emulator, so those might very well cause >>> regressions on dream. >>> >>> JBQ >>> >>> On Thu, Mar 19, 2009 at 1:07 PM, Jean-Baptiste Queru <j...@android.com> >>> wrote: >>> >>>> I've put together a hack that allows the system to compile and start >>>> all the way to the home app. I worked with the delicateness of >>>> open-heart surgery performed with a chainsaw. >>>> >>>> Steps: >>>> >>>> -remove the opencore files ( rm -rf external/opencore >>>> .repo/projects/external/opencore.git ). Remove opencore from your >>>> .repo/manifest.xml if you intend to repo sync the entire world but >>>> don't want to have to remove opencore every single time. >>>> >>>> -patch in change 9300 ( repo download platform/frameworks/base 9300/1 ) >>>> >>>> -most probably do a clean build ( rm -rf out/ ; make ) >>>> >>>> I've "tested" on a device/release/generic/userdebug build. On my >>>> machine, it compiles, launches. The media process dies (which probably >>>> means that downloads are busted too), as well as the music player. The >>>> browser starts and can access the network. >>>> >>>> JBQ >>>> >>>> On Wed, Mar 18, 2009 at 7:57 PM, Jean-Baptiste Queru <j...@android.com> >>>> wrote: >>>> >>>>> I've submitted the merge (106 projects!), and I believe that the tree >>>>> is in the state that it should be. >>>>> >>>>> Caveats: >>>>> >>>>> -THE BUILD IS BROKEN. You've been warned. There's been some drift >>>>> around OpenCORE (probably situations where new code was written in >>>>> cupcake that uses OpenCORE 1, or where APIs were removed in cupcake >>>>> that OpenCORE 2 relies on). >>>>> >>>>> The proper command to try to merge the OpenCORE code should be "git >>>>> merge remotes/korg/cupcake" (I'm typing from memory). >>>>> >>>>> -I'm not 100% sure that the server contains exactly what it should. >>>>> I've had a filesystem failure right as I was trying to verify it, and >>>>> I'm not gonna be able to verify until at least sometime tomorrow. >>>>> >>>>> JBQ >>>>> >>>>> On Wed, Mar 18, 2009 at 12:43 PM, Jean-Baptiste Queru <j...@android.com> >>>>> wrote: >>>>> >>>>>> I expect to start submitting the changes in about an hour, i.e. >>>>>> between 1:30pm and 2pm PDT. >>>>>> >>>>>> Starting right now, you may want to avoid initiating a new repo sync, >>>>>> unless you're OK ending up with a tree that might not even compile. >>>>>> >>>>>> JBQ >>>>>> >>>>>> On Tue, Mar 17, 2009 at 6:35 PM, Jean-Baptiste Queru <j...@android.com> >>>>>> wrote: >>>>>> >>>>>>> [bcc android-platform, android-framework, android-porting] >>>>>>> >>>>>>> I'm working on merging the latest cupcake code drop into master. The >>>>>>> task is quite hairy, so the following two guidelines probably apply: >>>>>>> >>>>>>> -please don't submit anything in gerrit, as you'll just get in my way. >>>>>>> -now is a good time to repo sync master, as I'm going to have to >>>>>>> submit the result of the merge in a state where it doesn't build, and >>>>>>> I have no idea how long it'll take to get it to build again after >>>>>>> that. >>>>>>> >>>>>>> Thanks, >>>>>>> JBQ >>>>>>> >>>>>>> -- >>>>>>> Jean-Baptiste M. "JBQ" Queru >>>>>>> Android Engineer, Google. >>>>>>> >>>>>>> Questions sent directly to me that have no reason for being private >>>>>>> will likely get ignored or forwarded to a public forum with no further >>>>>>> warning. >>>>>>> >>>>>>> >>>>>> >>>>>> -- >>>>>> Jean-Baptiste M. "JBQ" Queru >>>>>> Android Engineer, Google. >>>>>> >>>>>> Questions sent directly to me that have no reason for being private >>>>>> will likely get ignored or forwarded to a public forum with no further >>>>>> warning. >>>>>> >>>>>> >>>>> >>>>> -- >>>>> Jean-Baptiste M. "JBQ" Queru >>>>> Android Engineer, Google. >>>>> >>>>> Questions sent directly to me that have no reason for being private >>>>> will likely get ignored or forwarded to a public forum with no further >>>>> warning. >>>>> >>>>> >>>> >>>> -- >>>> Jean-Baptiste M. "JBQ" Queru >>>> Android Engineer, Google. >>>> >>>> Questions sent directly to me that have no reason for being private >>>> will likely get ignored or forwarded to a public forum with no further >>>> warning. >>>> >>>> >>> >>> -- >>> Jean-Baptiste M. "JBQ" Queru >>> Android Engineer, Google. >>> >>> Questions sent directly to me that have no reason for being private >>> will likely get ignored or forwarded to a public forum with no further >>> warning. >>> >>> >> >> -- >> Jean-Baptiste M. "JBQ" Queru >> Android Engineer, Google. >> >> Questions sent directly to me that have no reason for being private >> will likely get ignored or forwarded to a public forum with no further >> warning. >> >> > > > > Maybe this one can help someone to compile. I patch the koolu version of the opencore, to compile on master.
Michael --~--~---------~--~----~------------~-------~--~----~ unsubscribe: android-porting+unsubscr...@googlegroups.com website: http://groups.google.com/group/android-porting -~----------~----~----~----~------~----~------~--~---
The interface to the registerBuffers has changed. The user must allocate a BufferHeap structure and use it for the registerBuffers. Rename the setFrameCallback to setPreviewCallback Signed-off-by: Michael Trimarchi <mich...@panicking.kicks-ass.org> --- diff --git a/android/android_surface_output.cpp b/android/android_surface_output.cpp index 262a329..1c22d3b 100755 --- a/android/android_surface_output.cpp +++ b/android/android_surface_output.cpp @@ -1045,7 +1045,12 @@ OSCL_EXPORT_REF bool AndroidSurfaceOutput::initCheck() LOGE("Error creating frame buffer heap"); return false; } - status_t ret = mSurface->registerBuffers(displayWidth, displayHeight, frameWidth, frameHeight, PIXEL_FORMAT_RGB_565, mFrameHeap); + + ISurface::BufferHeap buffer( + displayWidth, displayHeight, frameWidth, frameHeight, + PIXEL_FORMAT_RGB_565, 0, 0, mFrameHeap); + + status_t ret = mSurface->registerBuffers(buffer); // create frame buffers for (int i = 0; i < kBufferCount; i++) { @@ -1075,7 +1080,11 @@ OSCL_EXPORT_REF bool AndroidSurfaceOutput::initCheck() mHeapPmem = new MemoryHeapPmem(master, 0); mHeapPmem->slap(); master.clear(); - status_t ret = mSurface->registerBuffers(displayWidth, displayHeight, frameWidth, frameHeight, PIXEL_FORMAT_YCbCr_420_SP, mHeapPmem); + ISurface::BufferHeap buffer( + displayWidth, displayHeight, frameWidth, frameHeight, + PIXEL_FORMAT_YCbCr_420_SP, 0, 0, mHeapPmem); + + status_t ret = mSurface->registerBuffers(buffer); // create frame buffers for (int i = 0; i < kBufferCount; i++) { @@ -1123,7 +1132,12 @@ OSCL_EXPORT_REF PVMFStatus AndroidSurfaceOutput::WriteFrameBuf(uint8* aData, uin master.clear(); // register frame buffers with SurfaceFlinger - status_t ret = mSurface->registerBuffers(iVideoDisplayWidth, iVideoDisplayHeight, iVideoWidth, iVideoHeight, PIXEL_FORMAT_YCbCr_420_SP, mHeapPmem); + ISurface::BufferHeap buffer( + iVideoDisplayWidth, iVideoDisplayHeight, + iVideoWidth, iVideoHeight, + PIXEL_FORMAT_YCbCr_420_SP, 0, 0, mHeapPmem); + + status_t ret = mSurface->registerBuffers(buffer); mInitialized = true; } diff --git a/android/author/android_camera_input.cpp b/android/author/android_camera_input.cpp index 7d8fae9..ba6e75b 100644 --- a/android/author/android_camera_input.cpp +++ b/android/author/android_camera_input.cpp @@ -65,7 +65,7 @@ AndroidCameraInput::~AndroidCameraInput() { LOGV("destructor"); if (mCamera != NULL) { - mCamera->setFrameCallback(NULL, this, FRAME_CALLBACK_FLAG_NOOP); + mCamera->setPreviewCallback(NULL, this, FRAME_CALLBACK_FLAG_NOOP); if ((mFlags & FLAGS_HOT_CAMERA) == 0) { LOGV("camera was cold when we started, stopping preview"); mCamera->stopPreview(); @@ -887,7 +887,7 @@ PVMFStatus AndroidCameraInput::DoStart() { LOGV("DoStart"); iState = STATE_STARTED; - mCamera->setFrameCallback(preview_callback, this, FRAME_CALLBACK_FLAG_CAMCORDER); + mCamera->setPreviewCallback(preview_callback, this, FRAME_CALLBACK_FLAG_CAMCORDER); AddDataEventToQueue(iMilliSecondsPerDataEvent); return PVMFSuccess; @@ -923,7 +923,7 @@ PVMFStatus AndroidCameraInput::DoStop(const AndroidCameraInputCmd& aCmd) { LOGV("DoStop"); iDataEventCounter = 0; - mCamera->setFrameCallback(NULL, this); + mCamera->setPreviewCallback(NULL, this); // if no buffers pending, complete the stop command if (mFrameRefCount == 0) {