matroska: add new parsing code

Add a new EBML parser implementation that should allow significant
improvements to the Matroska demuxer. The new parsing code is not
actually used yet by the demuxer. The only changes to existing code in
this commit are to generate the MATROSKA_ID_* / EBML_ID_* macro
definitions from the new implementation and to rename some of them
(the new implementation uses names matching the official Matroska spec).

The main parser implementation is added in ebml.c. There are two new
generated files, ebml_defs.c and ebml_types.h, that contain
definitions of EBML elements. Those are generated by the new script
TOOLS/matroska.py. There's a new Makefile target "generated_ebml" that
run the script to refresh the content of the generated files.

1 files changed, 397 insertions, 0 deletions

diff --git a/TOOLS/matroska.py b/TOOLS/matroska.py
new file mode 100755
index 0000000000..8368f35185
--- /dev/null
+++ b/TOOLS/matroska.py
@@ -0,0 +1,397 @@
+#!/usr/bin/python
+"""
+Generate C definitions for parsing Matroska files.
+Can also be used to directly parse Matroska files and display their contents.
+"""
+
+#
+# This file is part of MPlayer.
+#
+# MPlayer is free software; you can redistribute it and/or modify
+# it under the terms of the GNU General Public License as published by
+# the Free Software Foundation; either version 2 of the License, or
+# (at your option) any later version.
+#
+# MPlayer is distributed in the hope that it will be useful,
+# but WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+# GNU General Public License for more details.
+#
+# You should have received a copy of the GNU General Public License along
+# with MPlayer; if not, write to the Free Software Foundation, Inc.,
+# 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
+#
+
+
+elements_ebml = (
+    'EBML, 1a45dfa3, sub', (
+        'EBMLVersion, 4286, uint',
+        'EBMLReadVersion, 42f7, uint',
+        'EBMLMaxIDLength, 42f2, uint',
+        'EBMLMaxSizeLength, 42f3, uint',
+        'DocType, 4282, str',
+        'DocTypeVersion, 4287, uint',
+        'DocTypeReadVersion, 4285, uint',
+    ),
+
+    'CRC32, bf, binary',
+    'Void, ec, binary',
+)
+
+elements_matroska = (
+    'Segment, 18538067, sub', (
+
+        'SeekHead*, 114d9b74, sub', (
+            'Seek*, 4dbb, sub', (
+                'SeekID, 53ab, ebml_id',
+                'SeekPosition, 53ac, uint',
+            ),
+        ),
+
+        'Info*, 1549a966, sub', (
+            'SegmentUID, 73a4, binary',
+            'PrevUID, 3cb923, binary',
+            'NextUID, 3eb923, binary',
+            'TimecodeScale, 2ad7b1, uint',
+            'DateUTC, 4461, sint',
+            'Title, 7ba9, str',
+            'MuxingApp, 4d80, str',
+            'WritingApp, 5741, str',
+            'Duration, 4489, float',
+        ),
+
+        'Cluster*, 1f43b675, sub', (
+            'Timecode, e7, uint',
+            'BlockGroup*, a0, sub', (
+                'Block, a1, binary',
+                'BlockDuration, 9b, uint',
+                'ReferenceBlock*, fb, sint',
+            ),
+            'SimpleBlock*, a3, binary',
+        ),
+
+        'Tracks*, 1654ae6b, sub', (
+            'TrackEntry*, ae, sub', (
+                'TrackNumber, d7, uint',
+                'TrackUID, 73c5, uint',
+                'TrackType, 83, uint',
+                'FlagEnabled, b9, uint',
+                'FlagDefault, 88, uint',
+                'FlagForced, 55aa, uint',
+                'FlagLacing, 9c, uint',
+                'MinCache, 6de7, uint',
+                'DefaultDuration, 23e383, uint',
+                'TrackTimecodeScale, 23314f, float',
+                'MaxBlockAdditionID, 55ee, uint',
+                'Name, 536e, str',
+                'Language, 22b59c, str',
+                'CodecID, 86, str',
+                'CodecPrivate, 63a2, binary',
+                'CodecDecodeAll, aa, uint',
+                'Video, e0, sub', (
+                    'FlagInterlaced, 9a, uint',
+                    'PixelWidth, b0, uint',
+                    'PixelHeight, ba, uint',
+                    'DisplayWidth, 54b0, uint',
+                    'DisplayHeight, 54ba, uint',
+                    'FrameRate, 2383e3, float',
+                ),
+                'Audio, e1, sub', (
+                    'SamplingFrequency, b5, float',
+                    'Channels, 9f, uint',
+                    'BitDepth, 6264, uint',
+                ),
+                'ContentEncodings, 6d80, sub', (
+                    'ContentEncoding*, 6240, sub', (
+                        'ContentEncodingOrder, 5031, uint',
+                        'ContentEncodingScope, 5032, uint',
+                        'ContentEncodingType, 5033, uint',
+                        'ContentCompression, 5034, sub', (
+                            'ContentCompAlgo, 4254, uint',
+                            'ContentCompSettings, 4255, binary',
+                        ),
+                    ),
+                ),
+            ),
+        ),
+
+        'Cues, 1c53bb6b, sub', (
+            'CuePoint*, bb, sub', (
+                'CueTime, b3, uint',
+                'CueTrackPositions*, b7, sub', (
+                    'CueTrack, f7, uint',
+                    'CueClusterPosition, f1, uint',
+                ),
+            ),
+        ),
+
+        'Attachments, 1941a469, sub', (
+            'AttachedFile*, 61a7, sub', (
+                'FileName, 466e, str',
+                'FileMimeType, 4660, str',
+                'FileData, 465c, binary',
+                'FileUID, 46ae, uint',
+            ),
+        ),
+
+        'Chapters, 1043a770, sub', (
+            'EditionEntry*, 45b9, sub', (
+                'EditionUID, 45bc, uint',
+                'EditionFlagHidden, 45bd, uint',
+                'EditionFlagDefault, 45db, uint',
+                'EditionFlagOrdered, 45dd, uint',
+                'ChapterAtom*, b6, sub', (
+                    'ChapterUID, 73c4, uint',
+                    'ChapterTimeStart, 91, uint',
+                    'ChapterTimeEnd, 92, uint',
+                    'ChapterFlagHidden, 98, uint',
+                    'ChapterFlagEnabled, 4598, uint',
+                    'ChapterSegmentUID, 6e67, binary',
+                    'ChapterSegmentEditionUID, 6ebc, uint',
+                    'ChapterDisplay*, 80, sub', (
+                        'ChapString, 85, str',
+                        'ChapLanguage*, 437c, str',
+                    ),
+                ),
+            ),
+        ),
+        'Tags*, 1254c367, sub', (
+            'Tag*, 7373, sub', (
+                'Targets, 63c0, sub', (
+                    'TargetTypeValue, 68ca, uint',
+                 ),
+            ),
+        ),
+    ),
+)
+
+
+import sys
+from math import ldexp
+
+def byte2num(s):
+    return int(s.encode('hex'), 16)
+
+def camelcase_to_words(name):
+    parts = []
+    start = 0
+    for i in range(1, len(name)):
+        if name[i].isupper() and (name[i-1].islower() or
+                                  name[i+1:i+2].islower()):
+            parts.append(name[start:i])
+            start = i
+    parts.append(name[start:])
+    return '_'.join(parts).lower()
+
+class MatroskaElement(object):
+
+    def __init__(self, name, elid, valtype, namespace):
+        self.name = name
+        self.definename = '%s_ID_%s' % (namespace, name.upper())
+        self.fieldname = camelcase_to_words(name)
+        self.structname = 'ebml_' + self.fieldname
+        self.elid = elid
+        self.valtype = valtype
+        if valtype == 'sub':
+            self.ebmltype = 'EBML_TYPE_SUBELEMENTS'
+            self.valname = 'struct %s' % self.structname
+        else:
+            self.ebmltype = 'EBML_TYPE_' + valtype.upper()
+            try:
+                self.valname = {'uint': 'uint64_t', 'str': 'struct bstr',
+                                'binary': 'struct bstr', 'ebml_id': 'uint32_t',
+                                'float': 'double', 'sint': 'int64_t',
+                                }[valtype]
+            except KeyError:
+                raise SyntaxError('Unrecognized value type ' + valtype)
+        self.subelements = ()
+
+    def add_subelements(self, subelements):
+        self.subelements = subelements
+        self.subids = set(x[0].elid for x in subelements)
+
+elementd = {}
+elementlist = []
+def parse_elems(l, namespace):
+    subelements = []
+    for el in l:
+        if isinstance(el, str):
+            name, hexid, eltype = [x.strip() for x in el.split(',')]
+            multiple = name.endswith('*')
+            name = name.strip('*')
+            new = MatroskaElement(name, hexid, eltype, namespace)
+            elementd[hexid] = new
+            elementlist.append(new)
+            subelements.append((new, multiple))
+        else:
+            new.add_subelements(parse_elems(el, namespace))
+    return subelements
+
+parse_elems(elements_ebml, 'EBML')
+parse_elems(elements_matroska, 'MATROSKA')
+
+def generate_C_header():
+    print('// Generated by TOOLS/matroska.py, do not edit manually')
+    print
+
+    for el in elementlist:
+        print('#define %-40s 0x%s' % (el.definename, el.elid))
+
+    print
+
+    for el in reversed(elementlist):
+        if not el.subelements:
+            continue
+        print
+        print('struct %s {' % el.structname)
+        l = max(len(subel.valname) for subel, multiple in el.subelements)+1
+        for subel, multiple in el.subelements:
+            print('    %-*s %s%s;' % (l, subel.valname, (' ', '*')[multiple],
+                                    subel.fieldname))
+        print
+        for subel, multiple in el.subelements:
+            print('    int  n_%s;' % (subel.fieldname))
+        print('};')
+
+    for el in elementlist:
+        if not el.subelements:
+            continue
+        print('extern const struct ebml_elem_desc %s_desc;' % el.structname)
+
+    print
+    print('#define MAX_EBML_SUBELEMENTS %d' % max(len(el.subelements)
+                                                  for el in elementlist))
+
+
+
+def generate_C_definitions():
+    print('// Generated by TOOLS/matroska.py, do not edit manually')
+    print
+    for el in reversed(elementlist):
+        print
+        if el.subelements:
+            print('#define N %s' % el.fieldname)
+            print('E_S("%s", %d)' % (el.name, len(el.subelements)))
+            for subel, multiple in el.subelements:
+                print('F(%s, %s, %d)' % (subel.definename, subel.fieldname,
+                                         multiple))
+            print('}};')
+            print('#undef N')
+        else:
+            print('E("%s", %s, %s)' % (el.name, el.fieldname, el.ebmltype))
+
+def read(s, length):
+    t = s.read(length)
+    if len(t) != length:
+        raise IOError
+    return t
+
+def read_id(s):
+    t = read(s, 1)
+    i = 0
+    mask = 128
+    if ord(t) == 0:
+        raise SyntaxError
+    while not ord(t) & mask:
+        i += 1
+        mask >>= 1
+    t += read(s, i)
+    return t
+
+def read_vint(s):
+    t = read(s, 1)
+    i = 0
+    mask = 128
+    if ord(t) == 0:
+        raise SyntaxError
+    while not ord(t) & mask:
+        i += 1
+        mask >>= 1
+    t = chr(ord(t) & (mask - 1))
+    t += read(s, i)
+    return i+1, byte2num(t)
+
+def read_str(s, length):
+    return read(s, length)
+
+def read_uint(s, length):
+    t = read(s, length)
+    return byte2num(t)
+
+def read_sint(s, length):
+    i = read_uint(s, length)
+    mask = 1 << (length * 8 - 1)
+    if i & mask:
+        i -= 2 * mask
+    return i
+
+def read_float(s, length):
+    t = read(s, length)
+    i = byte2num(t)
+    if length == 4:
+        f = ldexp((i & 0x7fffff) + (1 << 23), (i >> 23 & 0xff) - 150)
+        if i & (1 << 31):
+            f = -f
+        return f
+    raise SyntaxError
+
+def parse_one(s, depth, parent, maxlen):
+    elid = read_id(s).encode('hex')
+    elem = elementd.get(elid)
+    if parent is not None and elid not in parent.subids and elid not in ('ec', 'bf'):
+        print('Unexpected:', elid)
+        if 1:
+            raise NotImplementedError
+    size, length = read_vint(s)
+    this_length = len(elid) / 2 + size + length
+    if elem is not None:
+        if elem.valtype != 'skip':
+            print depth, elid, elem.name, 'size:', length, 'value:',
+        if elem.valtype == 'sub':
+            print('subelements:')
+            while length > 0:
+                length -= parse_one(s, depth + 1, elem, length)
+            if length < 0:
+                raise SyntaxError
+        elif elem.valtype == 'str':
+            print 'string', repr(read_str(s, length))
+        elif elem.valtype in ('binary', 'ebml_id'):
+            t = read_str(s, length)
+            dec = ''
+            if elem.valtype == 'ebml_id':
+                idelem = elementd.get(t.encode('hex'))
+                if idelem is None:
+                    dec = '(UNKNOWN)'
+                else:
+                    dec = '(%s)' % idelem.name
+            if len(t) < 20:
+                t = t.encode('hex')
+            else:
+                t = '<skipped %d bytes>' % len(t)
+            print 'binary', t, dec
+        elif elem.valtype == 'uint':
+            print 'uint', read_uint(s, length)
+        elif elem.valtype == 'sint':
+            print 'sint', read_sint(s, length)
+        elif elem.valtype == 'float':
+            print 'float', read_float(s, length)
+        elif elem.valtype == 'skip':
+            read(s, length)
+        else:
+            raise NotImplementedError
+    else:
+        print(depth, 'Unknown element:', elid, 'size:', length)
+        read(s, length)
+    return this_length
+
+def parse_toplevel(s):
+    parse_one(s, 0, None, 1 << 63)
+
+if sys.argv[1] == '--generate-header':
+    generate_C_header()
+elif sys.argv[1] == '--generate-definitions':
+    generate_C_definitions()
+else:
+    s = open(sys.argv[1])
+    while 1:
+        parse_toplevel(s)