This menas we can have GMPVideoDecoder's AVCC -> AnnexB conversion done by the H264Converter, and
simplify the code in WidevineVideoDecoder.
MozReview-Commit-ID: 3HT5VXth6LL
--HG--
extra : rebase_source : b840489edafa5dc981ba44f722d92083a40e34cd
The work I did in bug 1306314 seems to have either regressed or never worked
properly for multiple CDM same origin processes. I'd guess the decryptor IPDL
protocol actor ID must not be as unique as I thought. So if we just use a
counter managed by the GMPDecrytorChild, we'll get a per CDM unique ID, which
is sufficient.
MozReview-Commit-ID: KSh72ptX5fn
--HG--
extra : rebase_source : 9dd558aa9b2e9154e70fc328009b79e1daa884b2
This makes it easier to reuse in the ChromiumCDM code.
Also add an ExtractBuffer() method, which allows us to Move() the contained nsTArray
out without needing to copy the data.
MozReview-Commit-ID: 9suJSfXTVYy
--HG--
extra : rebase_source : 6eec99eb5329f3b8c3bb14d22459fee3bd95caf5
This makes it easier to reuse in the ChromiumCDM code.
Also add an ExtractBuffer() method, which allows us to Move() the contained nsTArray
out without needing to copy the data.
MozReview-Commit-ID: 9suJSfXTVYy
--HG--
extra : rebase_source : 89540b254249833cf8bb09792bb33cc402977d5a
This means we can reuse LogToConsole inside the new CDM decoder backend.
This change also makes GMPUtils.cpp build in non-unified build mode.
MozReview-Commit-ID: AFkdHIos4X2
--HG--
extra : rebase_source : d31e794ce94fa724a90b1cfa842a86d119a4e2d1
extra : source : 6cad0b06a556795f6d6de123bb5a153ff06062f5
This prevents the Log macro from colliding with the Log function on
IPC ParamTraits definitions.
MozReview-Commit-ID: Hd2v6ilbmGc
--HG--
extra : rebase_source : d26d495878706fe5a2009dd33d226cc71193be13
Sometimes the build breaks because this file uses nsPrintfCString but
doesn't include its header.
MozReview-Commit-ID: CcawXkMucdA
--HG--
extra : rebase_source : 3b36138053c1ffa557fd59af37cf1cfa4166493a
The job id is just a counter, so rather than have other users of DecryptJob
reimplement their own counter, we can push the id/counter code into DecryptJob
itself.
MozReview-Commit-ID: 3RB8ctplWkK
--HG--
extra : rebase_source : f6cd7fddb2bf208419cf314cd7b01c508d68b49e
extra : source : 9b18193d0c1ccdedec25f7395fe124a86660b9d5
This means it can be used in other CDMProxies, specifically the
upcoming Chromium CDM code in bug 1315850.
MozReview-Commit-ID: G26xclqhtSw
--HG--
extra : rebase_source : 041c2cb41ba444e0dea8de3ddcc6a119d480f4f7
extra : source : c7f66edac83a6662d99f59a48f70c539f8ccecc8
The job id is just a counter, so rather than have other users of DecryptJob
reimplement their own counter, we can push the id/counter code into DecryptJob
itself.
MozReview-Commit-ID: 3RB8ctplWkK
--HG--
extra : rebase_source : 7ee0e48ab7638d1f713ee6dd852ab4b5f91119e8
extra : source : 9b18193d0c1ccdedec25f7395fe124a86660b9d5
This means it can be used in other CDMProxies, specifically the
upcoming Chromium CDM code in bug 1315850.
MozReview-Commit-ID: G26xclqhtSw
--HG--
extra : rebase_source : c4c99063f7423b8cbbd49865d2d45eca32c254ce
extra : source : c7f66edac83a6662d99f59a48f70c539f8ccecc8
This means we don't need to include GMPService.h in GMPCrashHelperHolder.h,
in order to use the GMPService inside GMPCrashHelperHolder.
This which prevents an inclusion cycle between GMPService.h and GMPCrashHelperHolder.h
in another patch I'm working on.
MozReview-Commit-ID: AbcXvv4UMyl
--HG--
extra : rebase_source : bdae6e1fbbbe8ce4100f51d2339f69c23f12859f
Fixed coding style of files encountered in P1 and P2.
MozReview-Commit-ID: LApVu9K2oto
--HG--
extra : rebase_source : e3bb296baaec9df2011ff312fec2eda19dd125e6
This works, at least on Windows, if the NSPR_LOG_FILE is set at a file
in the OS temp dir. This means we can turn on CDM logging in release
builds, in the sandboxed child process, without needing to recompile
to #define on logging.
This will make debugging issues with the CDM easier.
MozReview-Commit-ID: 6cAxMy4lv3T
--HG--
extra : rebase_source : eb75bba8e0dc38d1a0137cef28b7589ded43351a
Note: Only the Adobe GMP used enum storage, so not that it's unused we may
as well remove this.
MozReview-Commit-ID: JtmQ69eJzaI
--HG--
extra : rebase_source : 29929e680dc1692b957b34ce274c4944743768e8
GMP gtests fail on ASAN builds now since the GMPLoader requires a sandbox
starter, and ASAN doesn't run with a GMP sandbox. So only enforce that we
need a sandbox starter if we've built with sandboxing enabled.
MozReview-Commit-ID: GptxIZ7TFIy
--HG--
extra : rebase_source : 6265f91a9c80555b63f71ac5da116450d4728df1
The GMPLoader code was in plugin-container so that it was covered by
Adobe's voucher of plugin-container, but that's no longer necessary.
MozReview-Commit-ID: 3VRBAohRI9I
--HG--
extra : rebase_source : 58a30855ade14af4c4b1420edabd3abb398f232e