2 * (c) 2003 Ronald Bultje <rbultje@ronald.bitfreak.net>
4 * ebml-read.c: read EBML data from file/stream
6 * This library is free software; you can redistribute it and/or
7 * modify it under the terms of the GNU Library General Public
8 * License as published by the Free Software Foundation; either
9 * version 2 of the License, or (at your option) any later version.
11 * This library is distributed in the hope that it will be useful,
12 * but WITHOUT ANY WARRANTY; without even the implied warranty of
13 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
14 * Library General Public License for more details.
16 * You should have received a copy of the GNU Library General Public
17 * License along with this library; if not, write to the
18 * Free Software Foundation, Inc., 59 Temple Place - Suite 330,
19 * Boston, MA 02111-1307, USA.
28 #include "ebml-read.h"
36 static void gst_ebml_read_class_init (GstEbmlReadClass *klass);
37 static void gst_ebml_read_init (GstEbmlRead *ebml);
38 static GstElementStateReturn
39 gst_ebml_read_change_state (GstElement *element);
41 static GstElementClass *parent_class = NULL;
44 gst_ebml_read_get_type (void)
46 static GType gst_ebml_read_type = 0;
48 if (!gst_ebml_read_type) {
49 static const GTypeInfo gst_ebml_read_info = {
50 sizeof (GstEbmlReadClass),
53 (GClassInitFunc) gst_ebml_read_class_init,
58 (GInstanceInitFunc) gst_ebml_read_init,
62 g_type_register_static (GST_TYPE_ELEMENT, "GstEbmlRead",
63 &gst_ebml_read_info, 0);
66 return gst_ebml_read_type;
70 gst_ebml_read_class_init (GstEbmlReadClass *klass)
72 GstElementClass *gstelement_class = (GstElementClass *) klass;
74 parent_class = g_type_class_ref (GST_TYPE_ELEMENT);
76 gstelement_class->change_state = gst_ebml_read_change_state;
80 gst_ebml_read_init (GstEbmlRead *ebml)
87 static GstElementStateReturn
88 gst_ebml_read_change_state (GstElement *element)
90 GstEbmlRead *ebml = GST_EBML_READ (element);
92 switch (GST_STATE_TRANSITION (element)) {
93 case GST_STATE_READY_TO_PAUSED:
95 return GST_STATE_FAILURE;
96 ebml->bs = gst_bytestream_new (ebml->sinkpad);
98 case GST_STATE_PAUSED_TO_READY:
99 gst_bytestream_destroy (ebml->bs);
100 while (ebml->level) {
101 GstEbmlLevel *level = ebml->level->data;
103 ebml->level = g_list_remove (ebml->level, level);
111 if (GST_ELEMENT_CLASS (parent_class)->change_state)
112 return GST_ELEMENT_CLASS (parent_class)->change_state (element);
114 return GST_STATE_SUCCESS;
118 * Return: the amount of levels in the hierarchy that the
119 * current element lies higher than the previous one.
120 * The opposite isn't done - that's auto-done using master
125 gst_ebml_read_element_level_up (GstEbmlRead *ebml)
128 guint64 pos = gst_bytestream_tell (ebml->bs);
130 while (ebml->level != NULL) {
131 GList *last = g_list_last (ebml->level);
132 GstEbmlLevel *level = last->data;
134 if (pos >= level->start + level->length) {
135 ebml->level = g_list_remove (ebml->level, level);
146 * Read: the element content data ID.
147 * Return: the number of bytes read or -1 on error.
151 gst_ebml_read_element_id (GstEbmlRead *ebml,
156 gint len_mask = 0x80, read = 1, n = 1;
159 while (gst_bytestream_peek_bytes (ebml->bs, &data, 1) != 1) {
160 GstEvent *event = NULL;
163 /* Here, we might encounter EOS */
164 gst_bytestream_get_status (ebml->bs, &remaining, &event);
166 gst_pad_event_default (ebml->sinkpad, event);
168 guint64 pos = gst_bytestream_tell (ebml->bs);
169 gst_event_unref (event);
170 GST_ELEMENT_ERROR (ebml, RESOURCE, READ, NULL,
171 ("Read error at position %llu (0x%llx)",
177 while (read <= 4 && !(total & len_mask)) {
182 guint64 pos = gst_bytestream_tell (ebml->bs);
183 GST_ELEMENT_ERROR (ebml, STREAM, DEMUX, NULL,
184 ("Invalid EBML ID size tag (0x%x) at position %llu (0x%llx)",
189 if (gst_bytestream_peek_bytes (ebml->bs, &data, read) != read) {
190 guint64 pos = gst_bytestream_tell (ebml->bs);
191 GST_ELEMENT_ERROR (ebml, RESOURCE, READ, NULL,
192 ("Read error at position %llu (0x%llx)", pos, pos));
196 total = (total << 8) | data[n++];
202 *level_up = gst_ebml_read_element_level_up (ebml);
208 * Read: element content length.
209 * Return: the number of bytes read or -1 on error.
213 gst_ebml_read_element_length (GstEbmlRead *ebml,
217 gint len_mask = 0x80, read = 1, n = 1, num_ffs = 0;
220 if (gst_bytestream_peek_bytes (ebml->bs, &data, 1) != 1) {
221 guint64 pos = gst_bytestream_tell (ebml->bs);
222 GST_ELEMENT_ERROR (ebml, RESOURCE, READ, NULL,
223 ("Read error at position %llu (0x%llx)", pos, pos));
227 while (read <= 8 && !(total & len_mask)) {
232 guint64 pos = gst_bytestream_tell (ebml->bs);
233 GST_ELEMENT_ERROR (ebml, STREAM, DEMUX, NULL,
234 ("Invalid EBML length size tag (0x%x) at position %llu (0x%llx)",
239 if ((total &= (len_mask - 1)) == len_mask - 1)
241 if (gst_bytestream_peek_bytes (ebml->bs, &data, read) != read) {
242 guint64 pos = gst_bytestream_tell (ebml->bs);
243 GST_ELEMENT_ERROR (ebml, RESOURCE, READ, NULL,
244 ("Read error at position %llu (0x%llx)", pos, pos));
250 total = (total << 8) | data[n];
255 *length = G_MAXUINT64;
263 * Read: the actual data.
264 * Return: the data, as a GstBuffer.
268 gst_ebml_read_element_data (GstEbmlRead *ebml,
271 GstBuffer *buf = NULL;
273 if (gst_bytestream_peek (ebml->bs, &buf, length) != length) {
274 guint64 pos = gst_bytestream_tell (ebml->bs);
275 GST_ELEMENT_ERROR (ebml, RESOURCE, READ, NULL,
276 ("Read error at position %llu (0x%llx)", pos, pos));
278 gst_buffer_unref (buf);
282 gst_bytestream_flush_fast (ebml->bs, length);
288 * Return: the ID of the next element.
289 * Level_up contains the amount of levels that this
290 * next element lies higher than the previous one.
294 gst_ebml_peek_id (GstEbmlRead *ebml,
301 if (gst_ebml_read_element_id (ebml, &id, level_up) < 0)
308 * Seek to a given offset.
312 gst_ebml_read_seek (GstEbmlRead *ebml,
316 GstEvent *event = NULL;
319 /* first, flush remaining buffers */
320 gst_bytestream_get_status (ebml->bs, &remaining, &event);
322 g_warning ("Unexpected event before seek");
323 gst_event_unref (event);
326 gst_bytestream_flush_fast (ebml->bs, remaining);
329 if (!gst_bytestream_seek (ebml->bs, offset, GST_SEEK_METHOD_SET)) {
330 GST_ELEMENT_ERROR (ebml, RESOURCE, SEEK, NULL,
331 ("Seek to position %llu (0x%llx) failed", offset, offset));
336 /* and now, peek a new byte. This will fail because there's a
337 * pending event. Then, take the event and return it. */
338 if (gst_bytestream_peek_bytes (ebml->bs, &data, 1)) {
339 GST_WARNING ("Unexpected data after seek - this means seek failed");
343 /* get the discont event and return */
344 gst_bytestream_get_status (ebml->bs, &remaining, &event);
346 GST_WARNING ("No discontinuity event after seek - seek failed");
348 } else if (GST_EVENT_TYPE (event) != GST_EVENT_DISCONTINUOUS) {
349 gst_pad_event_default (ebml->sinkpad, event);
358 * Skip the next element.
362 gst_ebml_read_skip (GstEbmlRead *ebml)
365 guint32 id, remaining;
369 if ((bytes = gst_ebml_read_element_id (ebml, &id, NULL)) < 0)
371 gst_bytestream_flush_fast (ebml->bs, bytes);
373 if ((bytes = gst_ebml_read_element_length (ebml, &length)) < 0)
375 gst_bytestream_flush_fast (ebml->bs, bytes);
377 /* do we have enough bytes left to skip? */
378 gst_bytestream_get_status (ebml->bs, &remaining, &event);
380 g_warning ("Unexpected event before skip");
381 gst_event_unref (event);
384 if (remaining >= length)
385 return gst_bytestream_flush (ebml->bs, length);
387 if (!(event = gst_ebml_read_seek (ebml,
388 gst_bytestream_tell (ebml->bs) + length)))
391 gst_event_unref (event);
397 * Read the next element as a GstBuffer (binary).
401 gst_ebml_read_buffer (GstEbmlRead *ebml,
408 if ((bytes = gst_ebml_read_element_id (ebml, id, NULL)) < 0)
410 gst_bytestream_flush_fast (ebml->bs, bytes);
412 if ((bytes = gst_ebml_read_element_length (ebml, &length)) < 0)
414 gst_bytestream_flush_fast (ebml->bs, bytes);
416 return ((*buf = gst_ebml_read_element_data (ebml, length)) != NULL);
420 * Read the next element as an unsigned int.
424 gst_ebml_read_uint (GstEbmlRead *ebml,
432 if (!gst_ebml_read_buffer (ebml, id, &buf))
435 data = GST_BUFFER_DATA (buf);
436 size = GST_BUFFER_SIZE (buf);
437 if (size < 1 || size > 8) {
438 GST_ELEMENT_ERROR (ebml, STREAM, DEMUX, NULL,
439 ("Invalid integer element size %d at position %llu (0x%llu)",
440 size, GST_BUFFER_OFFSET (buf), GST_BUFFER_OFFSET (buf)));
441 gst_buffer_unref (buf);
446 *num = (*num << 8) | data[GST_BUFFER_SIZE (buf) - size];
450 gst_buffer_unref (buf);
456 * Read the next element as a signed int.
460 gst_ebml_read_sint (GstEbmlRead *ebml,
466 guint size, negative = 0, n = 0;
468 if (!gst_ebml_read_buffer (ebml, id, &buf))
471 data = GST_BUFFER_DATA (buf);
472 size = GST_BUFFER_SIZE (buf);
473 if (size < 1 || size > 8) {
474 GST_ELEMENT_ERROR (ebml, STREAM, DEMUX, NULL,
475 ("Invalid integer element size %d at position %llu (0x%llx)",
476 size, GST_BUFFER_OFFSET (buf), GST_BUFFER_OFFSET (buf)));
477 gst_buffer_unref (buf);
480 if (data[0] & 0x80) {
486 *num = (*num << 8) | data[n++];
491 *num = *num - (1LL << ((8 * size) - 1));
494 gst_buffer_unref (buf);
500 * Read the next element as a float.
504 gst_ebml_read_float (GstEbmlRead *ebml,
512 if (!gst_ebml_read_buffer (ebml, id, &buf))
515 data = GST_BUFFER_DATA (buf);
516 size = GST_BUFFER_SIZE (buf);
518 if (size != 4 && size != 8 && size != 10) {
519 GST_ELEMENT_ERROR (ebml, STREAM, DEMUX, NULL,
520 ("Invalid float element size %d at position %llu (0x%llx)",
521 size, GST_BUFFER_OFFSET (buf), GST_BUFFER_OFFSET (buf)));
522 gst_buffer_unref (buf);
527 GST_ELEMENT_ERROR (ebml, CORE, NOT_IMPLEMENTED, NULL,
528 ("FIXME! 10-byte floats unimplemented"));
529 gst_buffer_unref (buf);
536 #if (G_BYTE_ORDER == G_BIG_ENDIAN)
537 f = * (gfloat *) data;
540 ((guint8 *) &f)[size - 1] = data[4 - size];
549 #if (G_BYTE_ORDER == G_BIG_ENDIAN)
550 d = * (gdouble *) data;
553 ((guint8 *) &d)[size - 1] = data[8 - size];
561 gst_buffer_unref (buf);
567 * Read the next element as an ASCII string.
571 gst_ebml_read_ascii (GstEbmlRead *ebml,
577 if (!gst_ebml_read_buffer (ebml, id, &buf))
580 *str = g_malloc (GST_BUFFER_SIZE (buf) + 1);
581 memcpy (*str, GST_BUFFER_DATA (buf), GST_BUFFER_SIZE (buf));
582 (*str)[GST_BUFFER_SIZE (buf)] = '\0';
584 gst_buffer_unref (buf);
590 * Read the next element as a UTF-8 string.
594 gst_ebml_read_utf8 (GstEbmlRead *ebml,
598 return gst_ebml_read_ascii (ebml, id, str);
602 * Read the next element as a date (nanoseconds since 1/1/2000).
606 gst_ebml_read_date (GstEbmlRead *ebml,
610 return gst_ebml_read_sint (ebml, id, date);
614 * Read the next element, but only the header. The contents
615 * are supposed to be sub-elements which can be read separately.
619 gst_ebml_read_master (GstEbmlRead *ebml,
626 if ((bytes = gst_ebml_read_element_id (ebml, id, NULL)) < 0)
628 gst_bytestream_flush_fast (ebml->bs, bytes);
630 if ((bytes = gst_ebml_read_element_length (ebml, &length)) < 0)
632 gst_bytestream_flush_fast (ebml->bs, bytes);
635 level = g_new (GstEbmlLevel, 1);
636 level->start = gst_bytestream_tell (ebml->bs);
637 level->length = length;
638 ebml->level = g_list_append (ebml->level, level);
644 * Read the next element as binary data.
648 gst_ebml_read_binary (GstEbmlRead *ebml,
655 if (!gst_ebml_read_buffer (ebml, id, &buf))
658 *length = GST_BUFFER_SIZE (buf);
659 *binary = g_memdup (GST_BUFFER_DATA (buf), GST_BUFFER_SIZE (buf));
661 gst_buffer_unref (buf);
667 * Read an EBML header.
671 gst_ebml_read_header (GstEbmlRead *ebml,
675 /* this function is the first to be called */
685 if (!(id = gst_ebml_peek_id (ebml, &level_up)))
687 if (level_up != 0 || id != GST_EBML_ID_HEADER) {
688 GST_ELEMENT_ERROR (ebml, STREAM, WRONG_TYPE, NULL, NULL);
691 if (!gst_ebml_read_master (ebml, &id))
693 g_assert (id == GST_EBML_ID_HEADER);
696 if (!(id = gst_ebml_peek_id (ebml, &level_up)))
704 /* is our read version uptodate? */
705 case GST_EBML_ID_EBMLREADVERSION: {
708 if (!gst_ebml_read_uint (ebml, &id, &num))
710 g_assert (id == GST_EBML_ID_EBMLREADVERSION);
711 if (num != GST_EBML_VERSION)
716 /* we only handle 8 byte lengths at max */
717 case GST_EBML_ID_EBMLMAXSIZELENGTH: {
720 if (!gst_ebml_read_uint (ebml, &id, &num))
722 g_assert (id == GST_EBML_ID_EBMLMAXSIZELENGTH);
723 if (num != sizeof (guint64))
728 /* we handle 4 byte IDs at max */
729 case GST_EBML_ID_EBMLMAXIDLENGTH: {
732 if (!gst_ebml_read_uint (ebml, &id, &num))
734 g_assert (id == GST_EBML_ID_EBMLMAXIDLENGTH);
735 if (num != sizeof (guint32))
740 case GST_EBML_ID_DOCTYPE: {
743 if (!gst_ebml_read_ascii (ebml, &id, &text))
745 g_assert (id == GST_EBML_ID_DOCTYPE);
755 case GST_EBML_ID_DOCTYPEREADVERSION: {
758 if (!gst_ebml_read_uint (ebml, &id, &num))
760 g_assert (id == GST_EBML_ID_DOCTYPEREADVERSION);
767 GST_WARNING ("Unknown data type 0x%x in EBML header (ignored)", id);
770 /* we ignore these two, as they don't tell us anything we care about */
771 case GST_EBML_ID_VOID:
772 case GST_EBML_ID_EBMLVERSION:
773 case GST_EBML_ID_DOCTYPEVERSION:
774 if (!gst_ebml_read_skip (ebml))