mochitest test_buffered.html reports gizmo.mp4: Should be buffered in one range - got 0, expected 1 With the media.fragmented-mp4.exposed true and the Apple PlatformDecoderModule.

Comment on attachment 8475640 [details] [diff] [review] Fix final buffered range calculation Review of attachment 8475640 [details] [diff] [review]: ----------------------------------------------------------------- ::: media/libstagefright/binding/Index.cpp @@ +97,5 @@ > mMoofParser->RebuildFragmentedIndex(aByteRanges); > } > > +Microseconds > +Index::GetEndCompositionIfBufffered(const nsTArray<MediaByteRange>& aByteRanges) Autocomplete fail. Consider naming GetEndCompositionIfBufffffffered. ::: media/libstagefright/binding/mp4_demuxer.cpp @@ +228,4 @@ > nsTArray<Interval<Microseconds>> ranges; > mPrivate->mIndexes[i]->ConvertByteRangesToTimeRanges(aByteRanges, &ranges); > + if (lastComposition) { > + if (endCompositions[i]) { if (lastComposition && endCompositions[i]) {

I got a crash on Windows running mochitests in test_unseekable with the patch here: 12644 INFO Assertion failure: !mRemaining, at c:\Users\Bob\src\mozilla\purple\media\libstagefright\binding\include\mp4_demuxer/ByteReader.h:35 12645 INFO draining. 12646 INFO mp4_demuxer::BoxReader::~BoxReader+0x0000001A [xul +0x00000000001BDFDA] 12647 INFO mp4_demuxer::Tfhd::Tfhd+0x00000243 [xul +0x00000000001BDA13] (c:\users\bob\src\mozilla\purple\media\libstagefright\binding\moofparser.cpp, line 252) 12648 INFO mp4_demuxer::Moof::ParseTraf+0x00000085 [xul +0x00000000001BD695] (c:\users\bob\src\mozilla\purple\media\libstagefright\binding\moofparser.cpp, line 103) 12649 INFO mp4_demuxer::Moof::Moof+0x000000AA [xul +0x00000000001BD50A] (c:\users\bob\src\mozilla\purple\media\libstagefright\binding\moofparser.cpp, line 93) 12650 INFO mp4_demuxer::MoofParser::RebuildFragmentedIndex+0x000000CB [xul +0x00000000001BCCFB] (c:\users\bob\src\mozilla\purple\media\libstagefright\binding\mo ofparser.cpp, line 24) 12651 INFO mp4_demuxer::Index::UpdateMoofIndex+0x0000002E [xul +0x00000000001BC86E] (c:\users\bob\src\mozilla\purple\media\libstagefright\binding\index.cpp, lin e 98) 12652 INFO mp4_demuxer::MP4Demuxer::UpdateIndex+0x00000066 [xul +0x00000000001BD436] (c:\users\bob\src\mozilla\purple\media\libstagefright\binding\mp4_demuxer.c pp, line 206) 12653 INFO mozilla::MP4Reader::UpdateIndex+0x000000AA [xul +0x0000000002B1D80A] (c:\users\bob\src\mozilla\purple\content\media\fmp4\mp4reader.cpp, line 799) 12654 INFO nsRunnableMethodImpl<void (__thiscall mozilla::MP4Reader::*)(void),void,1>::Run+0x0000001A [xul +0x0000000002B1E8AA] (c:\users\bob\src\mozilla\purple \objdir\dist\include\nsthreadutils.h, line 393) 12655 INFO mozilla::MediaTaskQueue::Runner::Run+0x0000017E [xul +0x0000000002A20DFE] (c:\users\bob\src\mozilla\purple\content\media\mediataskqueue.cpp, line 201 ) 12656 INFO nsThreadPool::Run+0x000002B2 [xul +0x00000000002D3D72] (c:\users\bob\src\mozilla\purple\xpcom\threads\nsthreadpool.cpp, line 222) 12657 INFO nsThread::ProcessNextEvent+0x000003F2 [xul +0x00000000002D0C82] (c:\users\bob\src\mozilla\purple\xpcom\threads\nsthread.cpp, line 770) 12658 INFO NS_ProcessNextEvent+0x00000058 [xul +0x0000000000321A28] (c:\users\bob\src\mozilla\purple\xpcom\glue\nsthreadutils.cpp, line 265) 12659 INFO mozilla::ipc::MessagePumpForNonMainThreads::Run+0x000002B5 [xul +0x00000000007FFF95] (c:\users\bob\src\mozilla\purple\ipc\glue\messagepump.cpp, line 355) 12660 INFO MessageLoop::RunInternal+0x0000004E [xul +0x000000000077E5CE] (c:\users\bob\src\mozilla\purple\ipc\chromium\src\base\message_loop.cc, line 230) 12661 INFO MessageLoop::RunHandler+0x00000082 [xul +0x000000000077E4F2] (c:\users\bob\src\mozilla\purple\ipc\chromium\src\base\message_loop.cc, line 223) 12662 INFO MessageLoop::Run+0x0000001D [xul +0x000000000077E3FD] (c:\users\bob\src\mozilla\purple\ipc\chromium\src\base\message_loop.cc, line 197) 12663 INFO nsThread::ThreadFunc+0x00000118 [xul +0x00000000002CF368] (c:\users\bob\src\mozilla\purple\xpcom\threads\nsthread.cpp, line 354) 12664 INFO _PR_NativeRunThread+0x000000DB [nss3 +0x00000000002CAD3B] (c:\users\bob\src\mozilla\purple\nsprpub\pr\src\threads\combined\pruthr.c, line 397) 12665 INFO pr_root+0x00000019 [nss3 +0x00000000002B2969] (c:\users\bob\src\mozilla\purple\nsprpub\pr\src\md\windows\w95thred.c, line 90) 12666 INFO endthreadex+0x0000003A [MSVCR100 +0x000000000005C556] 12667 INFO endthreadex+0x000000E4 [MSVCR100 +0x000000000005C600] 12668 INFO BaseThreadInitThunk+0x0000000E [KERNEL32 +0x000000000001919F] 12669 INFO RtlInitializeExceptionChain+0x00000084 [ntdll +0x000000000004A22B] 12670 INFO RtlInitializeExceptionChain+0x0000005A [ntdll +0x000000000004A201] TEST-INFO | Main app process: exit status 80000003 69:13.36 TEST-UNEXPECTED-FAIL | /tests/content/media/test/test_unseekable.html | application terminated with exit code 2147483651 69:13.36 runtests.py | Application ran for: 1:09:01.621000 69:13.36 zombiecheck | Reading PID log: c:\users\bob\appdata\local\temp\tmplxz8c7pidlog PROCESS-CRASH | /tests/content/media/test/test_unseekable.html | application crashed [unknown top frame] Crash dump filename: c:\users\bob\appdata\local\temp\tmpcw4mzt.mozrunner\minidumps\8887c7fd-85d1-4fae-9c29-d31830b5dd50.dmp MINIDUMP_STACKWALK not set, can't process dump. 69:13.36 Stopping web server System JS : ERROR file:///c:/Users/Bob/src/mozilla/purple/objdir/dist/bin/components/nsB 69:13.43 Stopping web socket server 69:13.43 Stopping ssltunnel 69:13.43 69:13.44 WARNING | leakcheck | missing output line for total leaks! 69:13.44 runtests.py | Running tests: end. 12671 INFO SUITE-END | took 4141s

Incorporate review comments from Edwin, carry forward r+. I don't hit the crash on Mac, and since this helps us pass try there, I think we should go ahead and land this and follow up with a separate bug for the windows issue.

(In reply to Chris Pearce (:cpearce) from comment #5) > I got a crash on Windows running mochitests in test_unseekable with the > patch here: This means that the tfhd has some bytes at the end that are unaccounted for. There should be no reason for a platform difference here. Does it fail consistently?

Please use needinfo? to get my attention. I get too much bugspam to keep track. The test also fails, because the test expects the file to be unseekable, but with your demuxer we can seek in it. :D 152 INFO TEST-UNEXPECTED-FAIL | /tests/content/media/test/test_unseekable.html | Should have no seekable ranges in unseekable media: big-buck-bunny-unseekable.mp4 - got 1, expected 0