Implemented parsing of HEVC codec ids

media/base/video_codecs.cc

Index: media/base/video_codecs.cc
diff --git a/media/base/video_codecs.cc b/media/base/video_codecs.cc
index a2221a35d3dead73d0a14019d165c30116fa7ed3..3af27deffb2c3ac2ea438676dcd1b254e280263f 100644
--- a/media/base/video_codecs.cc
+++ b/media/base/video_codecs.cc
@@ -6,6 +6,7 @@
 
 #include "base/logging.h"
 #include "base/strings/string_number_conversions.h"
+#include "base/strings/string_split.h"
 #include "base/strings/string_util.h"
 
 namespace media {
@@ -62,6 +63,12 @@ std::string GetProfileName(VideoCodecProfile profile) {
       return "h264 stereo high";
     case H264PROFILE_MULTIVIEWHIGH:
       return "h264 multiview high";
+    case HEVCPROFILE_MAIN:
+      return "hevc main";
+    case HEVCPROFILE_MAIN10:
+      return "hevc main 10";
+    case HEVCPROFILE_MAIN_STILL_PICTURE:
+      return "hevc main still-picture";
     case VP8PROFILE_ANY:
       return "vp8";
     case VP9PROFILE_PROFILE0:
@@ -169,4 +176,115 @@ bool ParseAVCCodecId(const std::string& codec_id,
   return true;
 }
 
+#if BUILDFLAG(ENABLE_HEVC_DEMUXING)
+// The specification for HEVC codec id strings can be found in ISO IEC 14496-15
+// dated 2012 or newer in the Annex E.3
+bool ParseHEVCCodecId(const std::string& codec_id,
+                      VideoCodecProfile* profile,
+                      uint8_t* level_idc) {
+  if (!base::StartsWith(codec_id, "hev1.", base::CompareCase::SENSITIVE) &&
+      !base::StartsWith(codec_id, "hvc1.", base::CompareCase::SENSITIVE)) {
+    return false;
+  }
+
+  // HEVC codec id consists of:
+  const int kMaxHevcCodecIdLength =
+      5 +  // 'hev1.' or 'hvc1.' prefix (5 chars)
+      4 +  // profile, e.g. '.A12' (max 4 chars)
+      9 +  // profile_compatibility, dot + 32-bit hex number (max 9 chars)
+      5 +  // tier and level, e.g. '.H120' (max 5 chars)
+      18;  // up to 6 constraint bytes, bytes are dot-separated and hex-encoded.
+
+  if (codec_id.size() > kMaxHevcCodecIdLength) {
+    DVLOG(4) << __FUNCTION__ << ": Codec id is too long (" << codec_id << ")";
+    return false;
+  }
+
+  std::vector<std::string> elem = base::SplitString(
+      codec_id, ".", base::TRIM_WHITESPACE, base::SPLIT_WANT_ALL);

[ddorwin, 2016/04/20 22:28:15]

What happens with consecutive periods? If it creates an empty element, is that
handled correctly?

Let's add tests.

[servolk, 2016/04/20 23:14:46]

Yes, base::SPLIT_WANT_ALL means we get empty elements if there's nothing in
between two dots.

Done.

+  DCHECK(elem[0] == "hev1" || elem[0] == "hvc1");
+
+  if (elem.size() < 4) {
+    DVLOG(4) << __FUNCTION__ << ": invalid HEVC codec id " << codec_id;
+    return false;
+  }
+
+  uint8_t general_profile_space = 0;
+  if (elem[1][0] == 'A' || elem[1][0] == 'B' || elem[1][0] == 'C') {
+    general_profile_space = 1 + (elem[1][0] - 'A');
+    elem[1].erase(0, 1);
+  }
+  DCHECK(general_profile_space >= 0 && general_profile_space <= 3);
+
+  unsigned general_profile_idc = 0;
+  if (!base::StringToUint(elem[1], &general_profile_idc) ||
+      general_profile_idc > 0x1f) {
+    DVLOG(4) << __FUNCTION__ << ": invalid general_profile_idc=" << elem[1];
+    return false;
+  }
+
+  uint32_t general_profile_compatibility_flags = 0;
+  if (!base::HexStringToUInt(elem[2], &general_profile_compatibility_flags)) {
+    DVLOG(4) << __FUNCTION__
+             << ": invalid general_profile_compatibility_flags=" << elem[2];
+    return false;
+  }
+
+  if (profile) {
+    // TODO(servolk): Handle format range extension profiles as explained in
+    // HEVC standard (ISO/IEC ISO/IEC 23008-2) section A.3.5
+    if (general_profile_idc == 3 || (general_profile_compatibility_flags & 4)) {
+      *profile = HEVCPROFILE_MAIN_STILL_PICTURE;
+    }
+    if (general_profile_idc == 2 || (general_profile_compatibility_flags & 2)) {
+      *profile = HEVCPROFILE_MAIN10;
+    }
+    if (general_profile_idc == 1 || (general_profile_compatibility_flags & 1)) {
+      *profile = HEVCPROFILE_MAIN;
+    }
+  }
+
+  uint8_t general_tier_flag;
+  if (elem[3][0] == 'L' || elem[3][0] == 'H') {
+    general_tier_flag = (elem[3][0] == 'L') ? 0 : 1;
+    elem[3].erase(0, 1);
+  } else {
+    DVLOG(4) << __FUNCTION__ << ": invalid general_tier_flag=" << elem[3];
+    return false;
+  }
+  DCHECK(general_tier_flag == 0 || general_tier_flag == 1);
+
+  unsigned general_level_idc = 0;
+  if (!base::StringToUint(elem[3], &general_level_idc) ||
+      general_level_idc > 0xff) {
+    DVLOG(4) << __FUNCTION__ << ": invalid general_level_idc=" << elem[3];
+    return false;
+  }
+
+  if (level_idc)
+    *level_idc = static_cast<uint8_t>(general_level_idc);
+
+  uint8_t constraint_flags[6];
+  memset(constraint_flags, 0, sizeof(constraint_flags));
+
+  // Remove everything up to the first constraint_flags byte

[ddorwin, 2016/04/20 22:28:15]

This seems a little strange when we could just change the starting position.
Also, there is unnecessary modification of the data. Could we instead use an
iterator in the for loop?

[servolk, 2016/04/20 23:14:46]

Done.

+  elem.erase(elem.begin(), elem.begin() + 4);
+  if (elem.size() > 6) {

[ddorwin, 2016/04/20 22:28:15]

In particular, it's not immediately clear that this is really allowing up to 10
elements (nine periods). This is minor, though it does require a little more
thought.

[servolk, 2016/04/20 23:14:46]

Since we no longer remove the first elements, the number 10 is more explicit
now.

+    DVLOG(4) << __FUNCTION__ << ": unexpected number of trailing bytes in HEVC "
+             << "codec id " << codec_id;
+    return false;
+  }
+  for (size_t i = 0; i < elem.size(); ++i) {
+    unsigned constr_byte = 0;
+    if (!base::HexStringToUInt(elem[i], &constr_byte) || constr_byte > 0xFF) {
+      DVLOG(4) << __FUNCTION__ << ": invalid constraint byte=" << elem[i];
+      return false;
+    }
+    constraint_flags[i] = constr_byte;
+  }
+
+  return true;
+}
+#endif
+
 }  // namespace media