[FFmpeg-devel] [PATCH v15] avformat/dashdec: add dash demuxer base version
wm4
nfxjfg at googlemail.com
Mon Aug 28 14:48:03 EEST 2017
On Sun, 27 Aug 2017 22:19:31 +0800
Steven Liu <lq at chinaffmpeg.org> wrote:
> ffmpeg need a dash demuxer for demux the dash formats base on
> https://github.com/samsamsam-iptvplayer/exteplayer3/blob/master/tmp/ffmpeg/patches/3.2.2/000001_add_dash_demux.patch
>
> TODO:
> 1. support multi bitrate dash
>
> v2 fixed:
> 1. from autodetect to disabled
> 2. from camelCase code style to ffmpeg code style
> 3. from RepType to AVMediaType
> 4. fix variable typo
> 5. change time value from uint32_t to uint64_t
> 6. removed be used once API
> 7. change 'time(NULL)`, except it is not 2038-safe.' to av_gettime and av_timegm
> 8. merge complex free operation to free_fragment
> 9. use API from snprintf to av_asprintf
>
> v3 fixed:
> 1. fix typo from --enabled-xml2 to --enable-xml2
>
> v4 fixed:
> 1. from --enable-xml2 to --enable-libxml2
> 2. move system includes to top
> 3. remove nouse includes
> 4. rename enum name
> 5. add a trailing comma for the last entry enum
> 6. fix comment typo
> 7. add const to DASHContext class front
> 8. check sscanf if return arguments and give warning message when error
> 9. check validity before free seg->url and seg
> 10. check if the val is null, before use atoll
>
> v5 fixed:
> 1. fix typo from mainifest to manifest
>
> v6 fixed:
> 1. from realloc to av_realloc
> 2. from free to av_free
>
> v7 fixed:
> 1. remove the -lxml2 from configure when require_pkg_config
>
> v8 fixed:
> 1. fix replace filename template by av_asprintf secure problem
>
> v9 modified:
> 1. make manifest parser clearly
>
> v10 fixed:
> 1. fix function API name code style
> 2. remove redundant strreplace call
> 3. remove redundant memory operation and check return value from get_content_url()
> 4. add space between ) and {
> 5. remove no need to log the value for print
>
> v11 fixed:
> 1. from atoll to strtoll
>
> v12 fixed:
> 1. remove strreplace and instead by av_strreplace
>
> v13 fixed:
> 1. fix bug: cannot play:
> http://dash.edgesuite.net/akamai/bbb_30fps/bbb_30fps.mpd
>
> v14 fixed:
> 1. fix bug: TLS connection was non-properly terminated
> 2. fix bug: No trailing CRLF found in HTTP header
>
> v15 fixed:
> 1. play youtube link: ffmpeg -i $(youtube-dl -J "https://www.youtube.com/watch?v=XmL19DOP_Ls" | jq -r ".requested_formats[0].manifest_url")
> 2. code refine for timeline living stream
>
> Reviewed-by: Clément Bœsch <u at pkh.me>
> Reviewed-by: Michael Niedermayer <michael at niedermayer.cc>
> Reviewed-by: Carl Eugen Hoyos <cehoyos at ag.or.at>
> Reviewed-by: Rodger Combs <rodger.combs at gmail.com>
> Reviewed-by: Moritz Barsnick <barsnick at gmx.net>
> Reviewed-by: Nicolas George <george at nsup.org>
> Reviewed-by: Ricardo Constantino <wiiaboo at gmail.com>
> Reviewed-by: wm4 <nfxjfg at googlemail.com>
> Tested-by: Andy Furniss <adf.lists at gmail.com>
> Reported-by: Andy Furniss <adf.lists at gmail.com>
> Signed-off-by: Steven Liu <lq at chinaffmpeg.org>
> Signed-off-by: samsamsam <samsamsam at o2.pl>
I only have some superficial comments, since I don't think I have a
full overview over the DASH protocol and this code to tell whether the
more intricate details are handled correctly. As such it's not a full
review.
But I'd like to object to the code duplication. Also, stuffing this
into a single source file isn't ideal.
Also, I don't remember having given a full review on this before.
> ---
> configure | 4 +
> libavformat/Makefile | 1 +
> libavformat/allformats.c | 2 +-
> libavformat/dashdec.c | 1981 ++++++++++++++++++++++++++++++++++++++++++++++
> 4 files changed, 1987 insertions(+), 1 deletion(-)
> create mode 100644 libavformat/dashdec.c
>
> diff --git a/configure b/configure
> index 05f6dcc99a..7a7d61fa13 100755
> --- a/configure
> +++ b/configure
> @@ -272,6 +272,7 @@ External library support:
> --enable-libxcb-shape enable X11 grabbing shape rendering [autodetect]
> --enable-libxvid enable Xvid encoding via xvidcore,
> native MPEG-4/Xvid encoder exists [no]
> + --enable-libxml2 enable XML parsing using the C library libxml2 [no]
> --enable-libzimg enable z.lib, needed for zscale filter [no]
> --enable-libzmq enable message passing via libzmq [no]
> --enable-libzvbi enable teletext support via libzvbi [no]
> @@ -1576,6 +1577,7 @@ EXTERNAL_LIBRARY_LIST="
> libvpx
> libwavpack
> libwebp
> + libxml2
> libzimg
> libzmq
> libzvbi
> @@ -2937,6 +2939,7 @@ avi_muxer_select="riffenc"
> caf_demuxer_select="iso_media riffdec"
> caf_muxer_select="iso_media"
> dash_muxer_select="mp4_muxer"
> +dash_demuxer_deps="libxml2"
> dirac_demuxer_select="dirac_parser"
> dts_demuxer_select="dca_parser"
> dtshd_demuxer_select="dca_parser"
> @@ -5996,6 +5999,7 @@ enabled openssl && { use_pkg_config openssl openssl/ssl.h OPENSSL_init
> check_lib openssl openssl/ssl.h SSL_library_init -lssl -lcrypto -lws2_32 -lgdi32 ||
> die "ERROR: openssl not found"; }
> enabled qtkit_indev && { check_header_objcc QTKit/QTKit.h || disable qtkit_indev; }
> +enabled libxml2 && require_pkg_config libxml-2.0 libxml2/libxml/xmlversion.h xmlCheckVersion
>
> if enabled gcrypt; then
> GCRYPT_CONFIG="${cross_prefix}libgcrypt-config"
> diff --git a/libavformat/Makefile b/libavformat/Makefile
> index f2b465cfa2..3d478749d0 100644
> --- a/libavformat/Makefile
> +++ b/libavformat/Makefile
> @@ -133,6 +133,7 @@ OBJS-$(CONFIG_CRC_MUXER) += crcenc.o
> OBJS-$(CONFIG_DATA_DEMUXER) += rawdec.o
> OBJS-$(CONFIG_DATA_MUXER) += rawenc.o
> OBJS-$(CONFIG_DASH_MUXER) += dashenc.o
> +OBJS-$(CONFIG_DASH_DEMUXER) += dashdec.o
> OBJS-$(CONFIG_DAUD_DEMUXER) += dauddec.o
> OBJS-$(CONFIG_DAUD_MUXER) += daudenc.o
> OBJS-$(CONFIG_DCSTR_DEMUXER) += dcstr.o
> diff --git a/libavformat/allformats.c b/libavformat/allformats.c
> index cd8200ea1c..aeb9b710fe 100644
> --- a/libavformat/allformats.c
> +++ b/libavformat/allformats.c
> @@ -96,7 +96,7 @@ static void register_all(void)
> REGISTER_DEMUXER (CINE, cine);
> REGISTER_DEMUXER (CONCAT, concat);
> REGISTER_MUXER (CRC, crc);
> - REGISTER_MUXER (DASH, dash);
> + REGISTER_MUXDEMUX(DASH, dash);
> REGISTER_MUXDEMUX(DATA, data);
> REGISTER_MUXDEMUX(DAUD, daud);
> REGISTER_DEMUXER (DCSTR, dcstr);
> diff --git a/libavformat/dashdec.c b/libavformat/dashdec.c
> new file mode 100644
> index 0000000000..4718ce24ab
> --- /dev/null
> +++ b/libavformat/dashdec.c
> @@ -0,0 +1,1981 @@
> +/*
> + * Dynamic Adaptive Streaming over HTTP demux
> + * Copyright (c) 2017 samsamsam at o2.pl based on HLS demux
> + * Copyright (c) 2017 Steven Liu
> + *
> + * This file is part of FFmpeg.
> + *
> + * FFmpeg is free software; you can redistribute it and/or
> + * modify it under the terms of the GNU Lesser General Public
> + * License as published by the Free Software Foundation; either
> + * version 2.1 of the License, or (at your option) any later version.
> + *
> + * FFmpeg 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
> + * Lesser General Public License for more details.
> + *
> + * You should have received a copy of the GNU Lesser General Public
> + * License along with FFmpeg; if not, write to the Free Software
> + * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
> + */
> +#include <libxml/parser.h>
> +#include "libavutil/intreadwrite.h"
> +#include "libavutil/opt.h"
> +#include "libavutil/time.h"
> +#include "libavutil/parseutils.h"
> +#include "internal.h"
> +#include "avio_internal.h"
> +
> +#define INITIAL_BUFFER_SIZE 32768
> +
> +struct fragment {
> + int64_t url_offset;
> + int64_t size;
> + char *url;
> +};
> +
> +/*
> + * reference to : ISO_IEC_23009-1-DASH-2012
> + * Section: 5.3.9.6.2
> + * Table: Table 17 — Semantics of SegmentTimeline element
> + * */
> +struct timeline {
> + /* t: Element or Attribute Name
> + * specifies the MPD start time, in @timescale units,
> + * the first Segment in the series starts relative to the beginning of the Period.
> + * The value of this attribute must be equal to or greater than the sum of the previous S
> + * element earliest presentation time and the sum of the contiguous Segment durations.
> + * If the value of the attribute is greater than what is expressed by the previous S element,
> + * it expresses discontinuities in the timeline.
> + * If not present then the value shall be assumed to be zero for the first S element
> + * and for the subsequent S elements, the value shall be assumed to be the sum of
> + * the previous S element's earliest presentation time and contiguous duration
> + * (i.e. previous S at t + @d * (@r + 1)).
> + * */
> + int64_t t;
> + /* r: Element or Attribute Name
> + * specifies the repeat count of the number of following contiguous Segments with
> + * the same duration expressed by the value of @d. This value is zero-based
> + * (e.g. a value of three means four Segments in the contiguous series).
> + * */
> + int64_t r;
> + /* d: Element or Attribute Name
> + * specifies the Segment duration, in units of the value of the @timescale.
> + * */
> + int64_t d;
Maybe these fields should have better names...
> +};
> +
> +enum DASHTmplUrlType {
> + TMP_URL_TYPE_UNSPECIFIED,
> + TMP_URL_TYPE_NUMBER,
> + TMP_URL_TYPE_TIME,
> +};
> +
> +/*
> + * Each playlist has its own demuxer. If it is currently active,
> + * it has an opened AVIOContext too, and potentially an AVPacket
> + * containing the next packet from this stream.
> + */
> +struct representation {
> + char *url_template;
> + char *url_template_pattern;
> + char *url_template_format;
> + enum DASHTmplUrlType tmp_url_type;
> + AVIOContext pb;
> + AVIOContext *input;
> + AVFormatContext *parent;
> + AVFormatContext *ctx;
> + AVPacket pkt;
> + int rep_idx;
> + int rep_count;
> + int stream_index;
> +
> + enum AVMediaType type;
> + int64_t target_duration;
> +
> + int n_fragments;
> + struct fragment **fragments; /* VOD list of fragment for profile */
> +
> + int n_timelines;
> + struct timeline **timelines;
> +
> + int64_t first_seq_no;
> + int64_t last_seq_no;
> + int64_t start_number; /* used in case when we have dynamic list of segment to know which segments are new one*/
> +
> + int64_t fragment_duration;
> + int64_t fragment_timescale;
> +
> + int64_t presentation_timeoffset;
> +
> + int64_t cur_seq_no;
> + int64_t cur_seg_offset;
> + int64_t cur_seg_size;
> + struct fragment *cur_seg;
> +
> + /* Currently active Media Initialization Section */
> + struct fragment *init_section;
> + uint8_t *init_sec_buf;
> + uint32_t init_sec_buf_size;
> + uint32_t init_sec_data_len;
> + uint32_t init_sec_buf_read_offset;
> + int fix_multiple_stsd_order;
> + int64_t cur_timestamp;
> + int is_restart_needed;
> +};
> +
> +typedef struct DASHContext {
> + const AVClass *class;
> + char *base_url;
> + struct representation *cur_video;
> + struct representation *cur_audio;
> +
> + /* MediaPresentationDescription Attribute */
> + uint64_t media_presentation_duration;
> + uint64_t suggested_presentation_delay;
> + uint64_t availability_start_time;
> + uint64_t publish_time;
> + uint64_t minimum_update_period;
> + uint64_t time_shift_buffer_depth;
> + uint64_t min_buffer_time;
> +
> + /* Period Attribute */
> + uint64_t period_duration;
> + uint64_t period_start;
> +
> + int is_live;
> + AVIOInterruptCB *interrupt_callback;
> + char *user_agent; ///< holds HTTP user agent set as an AVOption to the HTTP protocol context
> + char *cookies; ///< holds HTTP cookie values set in either the initial response or as an AVOption to the HTTP protocol context
> + char *headers; ///< holds HTTP headers set as an AVOption to the HTTP protocol context
> + AVDictionary *avio_opts;
> +} DASHContext;
> +
> +static uint64_t get_current_time_in_sec(void)
> +{
> + return av_gettime() / 1000000;
> +}
> +
> +static uint64_t get_utc_date_time_insec(AVFormatContext *s, const char *datetime)
> +{
> + struct tm timeinfo;
> + int year = 0;
> + int month = 0;
> + int day = 0;
> + int hour = 0;
> + int minute = 0;
> + int ret = 0;
> + float second = 0.0;
> +
> + /* ISO-8601 date parser */
> + if (!datetime)
> + return 0;
> +
> + ret = sscanf(datetime, "%d-%d-%dT%d:%d:%fZ", &year, &month, &day, &hour, &minute, &second);
> + /* year, month, day, hour, minute, second 6 arguments */
> + if (ret != 6) {
> + av_log(s, AV_LOG_WARNING, "get_utc_date_time_insec get a wrong time format\n");
> + }
> + timeinfo.tm_year = year - 1900;
> + timeinfo.tm_mon = month - 1;
> + timeinfo.tm_mday = day;
> + timeinfo.tm_hour = hour;
> + timeinfo.tm_min = minute;
> + timeinfo.tm_sec = (int)second;
> +
> + return av_timegm(&timeinfo);
> +}
> +
> +static uint32_t get_duration_insec(AVFormatContext *s, const char *duration)
> +{
> + /* ISO-8601 duration parser */
> + uint32_t days = 0;
> + uint32_t hours = 0;
> + uint32_t mins = 0;
> + uint32_t secs = 0;
> + uint32_t size = 0;
> + float value = 0;
> + uint8_t type = 0;
> + const char *ptr = duration;
> +
> + while (*ptr) {
> + if (*ptr == 'P' || *ptr == 'T') {
> + ptr++;
> + continue;
> + }
> +
> + if (sscanf(ptr, "%f%c%n", &value, &type, &size) != 2) {
%c takes a char*, %n takes a int*.
> + av_log(s, AV_LOG_WARNING, "get_duration_insec get a wrong time format\n");
> + return 0; /* parser error */
> + }
> + switch (type) {
> + case 'D':
> + days = (uint32_t)value;
> + break;
> + case 'H':
> + hours = (uint32_t)value;
> + break;
> + case 'M':
> + mins = (uint32_t)value;
> + break;
> + case 'S':
> + secs = (uint32_t)value;
> + break;
> + default:
> + // handle invalid type
> + break;
> + }
> + ptr += size;
> + }
> + return ((days * 24 + hours) * 60 + mins) * 60 + secs;
> +}
> +
> +static int64_t get_segment_start_time_based_on_timeline(struct representation *pls, int64_t cur_seq_no)
> +{
> + int64_t start_time = 0;
> + int64_t i = 0;
> + int64_t j = 0;
> + int64_t num = 0;
> +
> + if (pls->n_timelines) {
> + for (i = 0; i < pls->n_timelines; i++) {
> + if (pls->timelines[i]->t > 0) {
> + start_time = pls->timelines[i]->t;
> + }
> + if (num == cur_seq_no)
> + goto finish;
> +
> + start_time += pls->timelines[i]->d;
> + for (j = 0; j < pls->timelines[i]->r; j++) {
> + num++;
> + if (num == cur_seq_no)
> + goto finish;
> + start_time += pls->timelines[i]->d;
> + }
> + num++;
> + }
> + }
> +finish:
> + return start_time;
> +}
> +
> +static int64_t calc_next_seg_no_from_timelines(struct representation *pls, int64_t cur_time)
> +{
> + int64_t i = 0;
> + int64_t j = 0;
> + int64_t num = 0;
> + int64_t start_time = 0;
> +
> + for (i = 0; i < pls->n_timelines; i++) {
> + if (pls->timelines[i]->t > 0) {
> + start_time = pls->timelines[i]->t;
> + }
> + if (start_time > cur_time)
> + goto finish;
> +
> + start_time += pls->timelines[i]->d;
> + for (j = 0; j < pls->timelines[i]->r; j++) {
> + num++;
> + if (start_time > cur_time)
> + goto finish;
> + start_time += pls->timelines[i]->d;
> + }
> + num++;
> + }
> +
> + return -1;
> +
> +finish:
> + return num;
> +}
> +
> +static void free_fragment(struct fragment **seg)
> +{
> + if (!(*seg)) {
> + return;
> + }
> + av_freep(&(*seg)->url);
> + av_freep(seg);
> +}
> +
> +static void free_fragment_list(struct representation *pls)
> +{
> + int i;
> +
> + for (i = 0; i < pls->n_fragments; i++) {
> + free_fragment(&pls->fragments[i]);
> + }
> + av_freep(&pls->fragments);
> + pls->n_fragments = 0;
> +}
> +
> +static void free_timelines_list(struct representation *pls)
> +{
> + int i;
> +
> + for (i = 0; i < pls->n_timelines; i++) {
> + av_freep(&pls->timelines[i]);
> + }
> + av_freep(&pls->timelines);
> + pls->n_timelines = 0;
> +}
> +
> +static void free_representation(struct representation *pls)
> +{
> + free_fragment_list(pls);
> + free_timelines_list(pls);
> + free_fragment(&pls->cur_seg);
> + free_fragment(&pls->init_section);
> + av_freep(&pls->init_sec_buf);
> + av_freep(&pls->pb.buffer);
> + if (pls->input)
> + ff_format_io_close(pls->parent, &pls->input);
> + if (pls->ctx) {
> + pls->ctx->pb = NULL;
> + avformat_close_input(&pls->ctx);
> + }
> +
> + av_free(pls->url_template_pattern);
> + av_free(pls->url_template_format);
> + av_free(pls->url_template);
> + av_free(pls);
> +}
> +
> +static void update_options(char **dest, const char *name, void *src)
> +{
> + av_freep(dest);
> + av_opt_get(src, name, AV_OPT_SEARCH_CHILDREN, (uint8_t**)dest);
> + if (*dest)
> + av_freep(dest);
> +}
> +
> +static int open_url(AVFormatContext *s, AVIOContext **pb, const char *url,
> + AVDictionary *opts, AVDictionary *opts2, int *is_http)
> +{
> + DASHContext *c = s->priv_data;
> + AVDictionary *tmp = NULL;
> + const char *proto_name = NULL;
> + int ret;
> + void *p = NULL;
> +
> + av_dict_copy(&tmp, opts, 0);
> + av_dict_copy(&tmp, opts2, 0);
> +
> + if (av_strstart(url, "crypto", NULL)) {
> + if (url[6] == '+' || url[6] == ':')
> + proto_name = avio_find_protocol_name(url + 7);
> + }
> +
> + if (!proto_name)
> + proto_name = avio_find_protocol_name(url);
> +
> + if (!proto_name)
> + return AVERROR_INVALIDDATA;
> +
> + // only http(s) & file are allowed
> + if (!av_strstart(proto_name, "http", NULL) && !av_strstart(proto_name, "file", NULL)) {
> + return AVERROR_INVALIDDATA;
> + }
> + if (!strncmp(proto_name, url, strlen(proto_name)) && url[strlen(proto_name)] == ':') {
> + ;
> + } else if (av_strstart(url, "crypto", NULL) && !strncmp(proto_name, url + 7, strlen(proto_name)) && url[7 + strlen(proto_name)] == ':') {
That's awfully convoluted, and I'd probably take a while to figure out
what this even does.
In fact, this is just copy&pasted from hls.c.
> + ;
> + } else if (strcmp(proto_name, "file") || !strncmp(url, "file,", 5)) {
> + return AVERROR_INVALIDDATA;
> + }
(What does it mean if proto_name starts with "file,"?)
> + ret = s->io_open(s, pb, url, AVIO_FLAG_READ, &tmp);
I think demuxers are supposed to call some wrapper?
> + if (ret >= 0) {
> + // update cookies on http response with setcookies.
> + p = (s->flags & AVFMT_FLAG_CUSTOM_IO) ? NULL : s->pb;
> + update_options(&c->cookies, "cookies", p);
> + av_dict_set(&opts, "cookies", c->cookies, 0);
> + }
Some questionable code copy&pasted from HLS, I guess. I'm not sure if I
ever found out why it even checks for AVFMT_FLAG_CUSTOM_IO.
> + av_dict_free(&tmp);
> +
> + if (is_http)
> + *is_http = av_strstart(proto_name, "http", NULL);
> +
> + return ret;
> +
> +}
> +
> +static char *get_content_url(xmlNodePtr *baseurl_nodes,
> + int n_baseurl_nodes,
> + xmlChar *rep_id_val,
> + xmlChar *rep_bandwidth_val,
> + xmlChar *val)
> +{
> + int i;
> + xmlChar *text;
> + char *url = NULL;
> + char *tmp_str = av_mallocz(MAX_URL_SIZE);
I wonder if a static buffer would make the code below simpler, since
this is a fixed size buffer anyway.
> + char *tmp_str_2 = NULL;
> +
> + if (!tmp_str) {
> + return NULL;
> + }
> + for (i = 0; i < n_baseurl_nodes; ++i) {
> + if (baseurl_nodes[i] &&
> + baseurl_nodes[i]->children &&
> + baseurl_nodes[i]->children->type == XML_TEXT_NODE) {
> + text = xmlNodeGetContent(baseurl_nodes[i]->children);
> + if (text) {
> + tmp_str_2 = av_mallocz(MAX_URL_SIZE);
> + if (!tmp_str_2) {
> + av_free(tmp_str);
> + return NULL;
> + }
> + ff_make_absolute_url(tmp_str_2, MAX_URL_SIZE, tmp_str, text);
> + av_free(tmp_str);
> + tmp_str = tmp_str_2;
> + xmlFree(text);
> + }
> + }
> + }
> + if (val)
> + av_strlcat(tmp_str, (const char*)val, MAX_URL_SIZE);
> +
> + if (rep_id_val) {
> + url = av_strireplace(tmp_str, "$RepresentationID$", (const char*)rep_id_val);
> + av_free(tmp_str);
> + tmp_str = url;
> + }
> + if (rep_bandwidth_val && tmp_str)
> + url = av_strireplace(tmp_str, "$Bandwidth$", (const char*)rep_bandwidth_val);
> + if (tmp_str != url)
> + av_free(tmp_str);
> + return url;
> +}
> +
> +static xmlChar *get_val_from_nodes_tab(xmlNodePtr *nodes, const int n_nodes, const xmlChar *attrname)
> +{
> + int i;
> + xmlChar *val;
> +
> + for (i = 0; i < n_nodes; ++i) {
> + if (nodes[i]) {
> + val = xmlGetProp(nodes[i], attrname);
> + if (val)
> + return val;
> + }
> + }
> +
> + return NULL;
> +}
> +
> +static xmlNodePtr find_child_node_by_name(xmlNodePtr rootnode, const xmlChar *nodename)
> +{
> + xmlNodePtr node = rootnode;
> + if (!node) {
> + return NULL;
> + }
> +
> + node = xmlFirstElementChild(node);
> + while (node) {
> + if (!xmlStrcmp(node->name, nodename)) {
> + return node;
> + }
> + node = xmlNextElementSibling(node);
> + }
> + return NULL;
> +}
> +
> +static int get_repl_pattern_and_format(const char *i_url, const char *i_marker, char **o_pattern, char **o_format)
> +{
> + int ret = -1;
> + int marker_len = 0;
> + int format_len = 0;
> + char *prefix = NULL;
> + char *start = NULL;
> + char *end = NULL;
> +
> +
> + if (av_stristr(i_url, i_marker)) {
> + *o_pattern = av_strdup(i_marker);
> + *o_format = av_strdup("%"PRId64);
> + ret = 0;
> + } else {
> + prefix = av_strdup(i_marker);
> + marker_len = strlen(prefix)-1;
> + prefix[marker_len] = '\0';
> + start = av_stristr(i_url, prefix);
> + if (!start)
> + goto finish;
> +
> + end = strchr(start + 1, '$');
> + if (!end)
> + goto finish;
> +
> + if (start[marker_len] != '%')
> + goto finish;
> +
> + if (end[-1] != 'd')
> + goto finish;
> +
> + format_len = end - start - marker_len - 1 + strlen(PRId64);
> + *o_format = av_mallocz(format_len+1);
> + av_strlcpy(*o_format, start + marker_len, end - start - marker_len -1);
> + av_strlcat(*o_format, PRId64, strlen(*o_format) + strlen(PRId64));
*o_format is later passed as format string to snprintf(). This is
probably not safe, although I have a hard time understand just what
this code outputs.
Also, the entire function is missing a lot of allocation failure error
checks.
> + *o_pattern = av_mallocz(end - start + 2);
> + if (*o_pattern) {
> + ret = AVERROR(EINVAL);
> + goto finish;
> + }
> + av_strlcpy(*o_pattern, start, end - start + 1);
> + ret = 0;
> +
> +finish:
> + av_free(prefix);
> + }
> +
> + return ret;
> +}
> +
> +static enum AVMediaType get_content_type(xmlNodePtr node)
> +{
> + enum AVMediaType type = AVMEDIA_TYPE_UNKNOWN;
> + int i = 0;
> + const char *attr;
> + xmlChar *val = NULL;
> +
> + if (node) {
> + while (type == AVMEDIA_TYPE_UNKNOWN && i < 2) {
> + attr = (i) ? "mimeType" : "contentType";
> + val = xmlGetProp(node, attr);
> + if (val) {
> + if (av_stristr((const char *)val, "video")) {
> + type = AVMEDIA_TYPE_VIDEO;
> + } else if (av_stristr((const char *)val, "audio")) {
> + type = AVMEDIA_TYPE_AUDIO;
> + }
> + xmlFree(val);
> + }
> + i++;
> + }
Making this a loop looks pretty convoluted.
> + }
> + return type;
> +}
> +
> +static int parse_manifest_segmenturlnode(AVFormatContext *s, struct representation *rep,
> + xmlNodePtr fragmenturl_node,
> + xmlNodePtr *baseurl_nodes,
> + xmlChar *rep_id_val,
> + xmlChar *rep_bandwidth_val)
> +{
> + xmlChar *initialization_val = NULL;
> + xmlChar *media_val = NULL;
> +
> + if (!xmlStrcmp(fragmenturl_node->name, (const xmlChar *)"Initialization")) {
> + initialization_val = xmlGetProp(fragmenturl_node, "sourceURL");
> + if (initialization_val) {
> + rep->init_section = av_mallocz(sizeof(struct fragment));
> + if (!rep->init_section) {
> + xmlFree(initialization_val);
> + return AVERROR(ENOMEM);
> + }
> + rep->init_section->url = get_content_url(baseurl_nodes, 4,
> + rep_id_val,
> + rep_bandwidth_val,
> + initialization_val);
> + if (!rep->init_section->url) {
> + av_free(rep->init_section);
> + xmlFree(initialization_val);
> + return AVERROR(ENOMEM);
> + }
> + rep->init_section->size = -1;
> + xmlFree(initialization_val);
> + }
> + } else if (!xmlStrcmp(fragmenturl_node->name, (const xmlChar *)"SegmentURL")) {
> + media_val = xmlGetProp(fragmenturl_node, "media");
> + if (media_val) {
> + struct fragment *seg = av_mallocz(sizeof(struct fragment));
> + if (!seg) {
> + xmlFree(media_val);
> + return AVERROR(ENOMEM);
> + }
> + seg->url = get_content_url(baseurl_nodes, 4,
> + rep_id_val,
> + rep_bandwidth_val,
> + media_val);
> + if (!seg->url) {
> + av_free(seg);
> + xmlFree(media_val);
> + return AVERROR(ENOMEM);
> + }
> + seg->size = -1;
> + dynarray_add(&rep->fragments, &rep->n_fragments, seg);
> + xmlFree(media_val);
> + }
> + }
> +
> + return 0;
> +}
> +
> +static int parse_manifest_segmenttimeline(AVFormatContext *s, struct representation *rep,
> + xmlNodePtr fragment_timeline_node)
> +{
> + xmlAttrPtr attr = NULL;
> + xmlChar *val = NULL;
> +
> + if (!xmlStrcmp(fragment_timeline_node->name, (const xmlChar *)"S")) {
> + struct timeline *tml = av_mallocz(sizeof(struct timeline));
> + if (!tml) {
> + return AVERROR(ENOMEM);
> + }
> + attr = fragment_timeline_node->properties;
> + while (attr) {
> + val = xmlGetProp(fragment_timeline_node, attr->name);
> +
> + if (!val) {
> + av_log(s, AV_LOG_WARNING, "parse_manifest_segmenttimeline attr->name = %s val is NULL\n", attr->name);
> + continue;
> + }
> +
> + if (!xmlStrcmp(attr->name, (const xmlChar *)"t")) {
> + tml->t = (int64_t)strtoll(val, NULL, 10);
> + } else if (!xmlStrcmp(attr->name, (const xmlChar *)"r")) {
> + tml->r =(int64_t) strtoll(val, NULL, 10);
> + } else if (!xmlStrcmp(attr->name, (const xmlChar *)"d")) {
> + tml->d = (int64_t)strtoll(val, NULL, 10);
> +// rep->fragment_duration = (int64_t) strtoll(val, NULL, 10);
Stray comment?
> + }
> + attr = attr->next;
> + xmlFree(val);
> + }
> + dynarray_add(&rep->timelines, &rep->n_timelines, tml);
> + }
> +
> + return 0;
> +}
> +
> +static int parse_manifest_representation(AVFormatContext *s, const char *url,
> + xmlNodePtr node,
> + xmlNodePtr adaptionset_node,
> + xmlNodePtr mpd_baseurl_node,
> + xmlNodePtr period_baseurl_node,
> + xmlNodePtr fragment_template_node,
> + xmlNodePtr content_component_node,
> + xmlNodePtr adaptionset_baseurl_node)
> +{
> + int32_t ret = 0;
> + int32_t audio_rep_idx = 0;
> + int32_t video_rep_idx = 0;
> + char *temp_string = NULL;
> + DASHContext *c = s->priv_data;
> + struct representation *rep = NULL;
> + struct fragment *seg = NULL;
> + xmlNodePtr representation_segmenttemplate_node = NULL;
> + xmlNodePtr representation_baseurl_node = NULL;
> + xmlNodePtr representation_segmentlist_node = NULL;
> + xmlNodePtr fragment_timeline_node = NULL;
> + xmlNodePtr fragment_templates_tab[2];
> + xmlChar *duration_val = NULL;
> + xmlChar *presentation_timeoffset_val = NULL;
> + xmlChar *startnumber_val = NULL;
> + xmlChar *timescale_val = NULL;
> + xmlChar *initialization_val = NULL;
> + xmlChar *media_val = NULL;
> + xmlNodePtr baseurl_nodes[4];
> + xmlNodePtr representation_node = node;
> + xmlChar *rep_id_val = xmlGetProp(representation_node, "id");
> + xmlChar *rep_bandwidth_val = xmlGetProp(representation_node, "bandwidth");
> + enum AVMediaType type = AVMEDIA_TYPE_UNKNOWN;
> +
> + // try get information from representation
> + if (type == AVMEDIA_TYPE_UNKNOWN)
> + type = get_content_type(representation_node);
> + // try get information from contentComponen
> + if (type == AVMEDIA_TYPE_UNKNOWN)
> + type = get_content_type(content_component_node);
> + // try get information from adaption set
> + if (type == AVMEDIA_TYPE_UNKNOWN)
> + type = get_content_type(adaptionset_node);
> + if (type == AVMEDIA_TYPE_UNKNOWN) {
> + av_log(s, AV_LOG_VERBOSE, "Parsing '%s' - skipp not supported representation type\n", url);
> + } else if ((type == AVMEDIA_TYPE_VIDEO && !c->cur_video) || (type == AVMEDIA_TYPE_AUDIO && !c->cur_audio)) {
> + // convert selected representation to our internal struct
> + rep = av_mallocz(sizeof(struct representation));
> + if (!rep) {
> + ret = AVERROR(ENOMEM);
> + goto end;
> + }
> + representation_segmenttemplate_node = find_child_node_by_name(representation_node, "SegmentTemplate");
> + representation_baseurl_node = find_child_node_by_name(representation_node, "BaseURL");
> + representation_segmentlist_node = find_child_node_by_name(representation_node, "SegmentList");
> +
> + baseurl_nodes[0] = mpd_baseurl_node;
> + baseurl_nodes[1] = period_baseurl_node;
> + baseurl_nodes[2] = adaptionset_baseurl_node;
> + baseurl_nodes[3] = representation_baseurl_node;
> +
> + if (representation_segmenttemplate_node || fragment_template_node) {
> + fragment_timeline_node = NULL;
> + fragment_templates_tab[0] = representation_segmenttemplate_node;
> + fragment_templates_tab[1] = fragment_template_node;
> +
> + presentation_timeoffset_val = get_val_from_nodes_tab(fragment_templates_tab, 2, "presentationTimeOffset");
> + duration_val = get_val_from_nodes_tab(fragment_templates_tab, 2, "duration");
> + startnumber_val = get_val_from_nodes_tab(fragment_templates_tab, 2, "startNumber");
> + timescale_val = get_val_from_nodes_tab(fragment_templates_tab, 2, "timescale");
> + initialization_val = get_val_from_nodes_tab(fragment_templates_tab, 2, "initialization");
> + media_val = get_val_from_nodes_tab(fragment_templates_tab, 2, "media");
> +
> + if (initialization_val) {
> + rep->init_section = av_mallocz(sizeof(struct fragment));
> + if (!rep->init_section) {
> + av_free(rep);
> + ret = AVERROR(ENOMEM);
> + goto end;
> + }
> + rep->init_section->url = get_content_url(baseurl_nodes, 4, rep_id_val, rep_bandwidth_val, initialization_val);
> + if (!rep->init_section->url) {
> + av_free(rep->init_section);
> + av_free(rep);
> + ret = AVERROR(ENOMEM);
> + goto end;
> + }
> + rep->init_section->size = -1;
> + xmlFree(initialization_val);
> + }
> +
> + if (media_val) {
> + rep->url_template = get_content_url(baseurl_nodes, 4, rep_id_val, rep_bandwidth_val, media_val);
> + temp_string = rep->url_template;
> + if (temp_string) {
> + if (av_stristr(temp_string, "$Number")) {
> + get_repl_pattern_and_format(temp_string, "$Number$", &(rep->url_template_pattern), &(rep->url_template_format));
> + rep->tmp_url_type = TMP_URL_TYPE_NUMBER; /* Number-Based. */
> + } else if (av_stristr(temp_string, "$Time")) {
> + get_repl_pattern_and_format(temp_string, "$Time$", &(rep->url_template_pattern), &(rep->url_template_format));
> + rep->tmp_url_type = TMP_URL_TYPE_TIME; /* Time-Based. */
> + } else {
> + temp_string = NULL;
> + }
> + }
> + xmlFree(media_val);
> + }
> +
> + if (presentation_timeoffset_val) {
> + rep->presentation_timeoffset = (int64_t) strtoll(presentation_timeoffset_val, NULL, 10);
> + xmlFree(presentation_timeoffset_val);
> + }
> + if (duration_val) {
> + rep->fragment_duration = (int64_t) strtoll(duration_val, NULL, 10);
> + xmlFree(duration_val);
> + }
> + if (timescale_val) {
> + rep->fragment_timescale = (int64_t) strtoll(timescale_val, NULL, 10);
> + xmlFree(timescale_val);
> + }
> + if (startnumber_val) {
> + rep->first_seq_no = (int64_t) strtoll(startnumber_val, NULL, 10);
> + xmlFree(startnumber_val);
> + }
> +
> + fragment_timeline_node = find_child_node_by_name(representation_segmenttemplate_node, "SegmentTimeline");
> +
> + if (!fragment_timeline_node)
> + fragment_timeline_node = find_child_node_by_name(fragment_template_node, "SegmentTimeline");
> + if (fragment_timeline_node) {
> + fragment_timeline_node = xmlFirstElementChild(fragment_timeline_node);
> + while (fragment_timeline_node) {
> + ret = parse_manifest_segmenttimeline(s, rep, fragment_timeline_node);
> + if (ret < 0) {
> + return ret;
> + }
> + fragment_timeline_node = xmlNextElementSibling(fragment_timeline_node);
> + }
> + }
> + } else if (representation_baseurl_node && !representation_segmentlist_node) {
> + seg = av_mallocz(sizeof(struct fragment));
> + if (!seg) {
> + ret = AVERROR(ENOMEM);
> + goto end;
> + }
> + seg->url = get_content_url(baseurl_nodes, 4, rep_id_val, rep_bandwidth_val, NULL);
> + if (!seg->url) {
> + av_free(seg);
> + ret = AVERROR(ENOMEM);
> + goto end;
> + }
> + seg->size = -1;
> + dynarray_add(&rep->fragments, &rep->n_fragments, seg);
> + } else if (representation_segmentlist_node) {
> + // TODO: https://www.brendanlong.com/the-structure-of-an-mpeg-dash-mpd.html
> + // http://www-itec.uni-klu.ac.at/dash/ddash/mpdGenerator.php?fragmentlength=15&type=full
> + xmlNodePtr fragmenturl_node = NULL;
> + duration_val = xmlGetProp(representation_segmentlist_node, "duration");
> + timescale_val = xmlGetProp(representation_segmentlist_node, "timescale");
> + if (duration_val) {
> + rep->fragment_duration = (int64_t) strtoll(duration_val, NULL, 10);
Doesn't check for errors. Not sure if that's a problem. Same with the
call below.
> + xmlFree(duration_val);
> + }
> + if (timescale_val) {
> + rep->fragment_timescale = (int64_t) strtoll(timescale_val, NULL, 10);
> + xmlFree(timescale_val);
> + }
> + fragmenturl_node = xmlFirstElementChild(representation_segmentlist_node);
> + while (fragmenturl_node) {
> + ret = parse_manifest_segmenturlnode(s, rep, fragmenturl_node,
> + baseurl_nodes,
> + rep_id_val,
> + rep_bandwidth_val);
> + if (ret < 0) {
> + return ret;
> + }
> + fragmenturl_node = xmlNextElementSibling(fragmenturl_node);
> + }
> +
> + fragment_timeline_node = find_child_node_by_name(representation_segmenttemplate_node, "SegmentTimeline");
> +
> + if (!fragment_timeline_node)
> + fragment_timeline_node = find_child_node_by_name(fragment_template_node, "SegmentTimeline");
> + if (fragment_timeline_node) {
> + fragment_timeline_node = xmlFirstElementChild(fragment_timeline_node);
> + while (fragment_timeline_node) {
> + ret = parse_manifest_segmenttimeline(s, rep, fragment_timeline_node);
> + if (ret < 0) {
> + return ret;
> + }
> + fragment_timeline_node = xmlNextElementSibling(fragment_timeline_node);
> + }
> + }
> + } else {
> + free_representation(rep);
> + rep = NULL;
> + av_log(s, AV_LOG_ERROR, "Unknown format of Representation node id[%s] \n", (const char *)rep_id_val);
> + }
> +
> + if (rep) {
> + if (rep->fragment_duration > 0 && !rep->fragment_timescale)
> + rep->fragment_timescale = 1;
> + if (type == AVMEDIA_TYPE_VIDEO) {
> + rep->rep_idx = video_rep_idx;
> + c->cur_video = rep;
> + } else {
> + rep->rep_idx = audio_rep_idx;
> + c->cur_audio = rep;
> + }
> + }
> + }
> +
> + video_rep_idx += type == AVMEDIA_TYPE_VIDEO;
> + audio_rep_idx += type == AVMEDIA_TYPE_AUDIO;
> +
> +end:
> + if (rep_id_val)
> + xmlFree(rep_id_val);
> + if (rep_bandwidth_val)
> + xmlFree(rep_bandwidth_val);
> +
> + return ret;
> +}
> +
> +static int parse_manifest_adaptationset(AVFormatContext *s, const char *url,
> + xmlNodePtr adaptionset_node,
> + xmlNodePtr mpd_baseurl_node,
> + xmlNodePtr period_baseurl_node)
> +{
> + int ret = 0;
> + xmlNodePtr fragment_template_node = NULL;
> + xmlNodePtr content_component_node = NULL;
> + xmlNodePtr adaptionset_baseurl_node = NULL;
> + xmlNodePtr node = NULL;
> +
> + node = xmlFirstElementChild(adaptionset_node);
> + while (node) {
> + if (!xmlStrcmp(node->name, (const xmlChar *)"SegmentTemplate")) {
Why xmlStrcmp, and not strcmp?
> + fragment_template_node = node;
> + } else if (!xmlStrcmp(node->name, (const xmlChar *)"ContentComponent")) {
> + content_component_node = node;
> + } else if (!xmlStrcmp(node->name, (const xmlChar *)"BaseURL")) {
> + adaptionset_baseurl_node = node;
> + } else if (!xmlStrcmp(node->name, (const xmlChar *)"Representation")) {
> + ret = parse_manifest_representation(s, url, node,
> + adaptionset_node,
> + mpd_baseurl_node,
> + period_baseurl_node,
> + fragment_template_node,
> + content_component_node,
> + adaptionset_baseurl_node);
> + if (ret < 0) {
> + return ret;
> + }
> + }
> + node = xmlNextElementSibling(node);
> + }
> + return 0;
> +}
> +
> +
> +static int parse_manifest(AVFormatContext *s, const char *url, AVIOContext *in)
> +{
> + DASHContext *c = s->priv_data;
> + int ret = 0;
> + int close_in = 0;
> + uint8_t *new_url = NULL;
> + int64_t filesize = 0;
> + char *buffer = NULL;
> + AVDictionary *opts = NULL;
> + xmlDoc *doc = NULL;
> + xmlNodePtr root_element = NULL;
> + xmlNodePtr node = NULL;
> + xmlNodePtr period_node = NULL;
> + xmlNodePtr mpd_baseurl_node = NULL;
> + xmlNodePtr period_baseurl_node = NULL;
> + xmlNodePtr adaptionset_node = NULL;
> + xmlAttrPtr attr = NULL;
> + xmlChar *val = NULL;
> + uint32_t perdiod_duration_sec = 0;
> + uint32_t perdiod_start_sec = 0;
> + int32_t audio_rep_idx = 0;
> + int32_t video_rep_idx = 0;
> +
> + if (!in) {
> + close_in = 1;
> + /* This is XML manifest there is no need to set range header */
> + av_dict_set(&opts, "seekable", "0", 0);
> + // broker prior HTTP options that should be consistent across requests
> + av_dict_set(&opts, "user-agent", c->user_agent, 0);
> + av_dict_set(&opts, "cookies", c->cookies, 0);
> + av_dict_set(&opts, "headers", c->headers, 0);
> +
> + ret = avio_open2(&in, url, AVIO_FLAG_READ, c->interrupt_callback, &opts);
> + av_dict_free(&opts);
> + if (ret < 0)
> + return ret;
> + }
> +
> + if (av_opt_get(in, "location", AV_OPT_SEARCH_CHILDREN, &new_url) >= 0) {
> + c->base_url = av_strdup(new_url);
> + } else {
> + c->base_url = av_strdup(url);
> + }
> +
> + filesize = avio_size(in);
> + if (filesize <= 0) {
> + filesize = 8 * 1024;
> + }
> +
> + buffer = av_mallocz(filesize);
> + if (!buffer) {
> + return AVERROR(ENOMEM);
> + }
> +
> + filesize = avio_read(in, buffer, filesize);
> + if (filesize > 0) {
> + LIBXML_TEST_VERSION
> +
> + doc = xmlReadMemory(buffer, filesize, c->base_url, NULL, 0);
> + root_element = xmlDocGetRootElement(doc);
> + node = root_element;
> +
> + if (!node) {
> + ret = AVERROR_INVALIDDATA;
> + av_log(s, AV_LOG_ERROR, "Unable to parse '%s' - missing root node\n", url);
> + goto cleanup;
> + }
> +
> + if (node->type != XML_ELEMENT_NODE ||
> + xmlStrcmp(node->name, (const xmlChar *)"MPD")) {
> + ret = AVERROR_INVALIDDATA;
> + av_log(s, AV_LOG_ERROR, "Unable to parse '%s' - wrong root node name[%s] type[%d]\n", url, node->name, (int)node->type);
> + goto cleanup;
> + }
> +
> + val = xmlGetProp(node, "type");
> + if (!val) {
> + av_log(s, AV_LOG_ERROR, "Unable to parse '%s' - missing type attrib\n", url);
> + ret = AVERROR_INVALIDDATA;
> + goto cleanup;
> + }
> + if (!xmlStrcmp(val, (const xmlChar *)"dynamic"))
> + c->is_live = 1;
> + xmlFree(val);
> +
> + attr = node->properties;
> + while (attr) {
> + val = xmlGetProp(node, attr->name);
> +
> + if (!xmlStrcmp(attr->name, (const xmlChar *)"availabilityStartTime")) {
> + c->availability_start_time = get_utc_date_time_insec(s, (const char *)val);
> + } else if (!xmlStrcmp(attr->name, (const xmlChar *)"publishTime")) {
> + c->publish_time = get_utc_date_time_insec(s, (const char *)val);
> + } else if (!xmlStrcmp(attr->name, (const xmlChar *)"minimumUpdatePeriod")) {
> + c->minimum_update_period = get_duration_insec(s, (const char *)val);
> + } else if (!xmlStrcmp(attr->name, (const xmlChar *)"timeShiftBufferDepth")) {
> + c->time_shift_buffer_depth = get_duration_insec(s, (const char *)val);
> + } else if (!xmlStrcmp(attr->name, (const xmlChar *)"minBufferTime")) {
> + c->min_buffer_time = get_duration_insec(s, (const char *)val);
> + } else if (!xmlStrcmp(attr->name, (const xmlChar *)"suggestedPresentationDelay")) {
> + c->suggested_presentation_delay = get_duration_insec(s, (const char *)val);
> + } else if (!xmlStrcmp(attr->name, (const xmlChar *)"mediaPresentationDuration")) {
> + c->media_presentation_duration = get_duration_insec(s, (const char *)val);
> + }
> + attr = attr->next;
> + xmlFree(val);
> + }
> +
> + mpd_baseurl_node = find_child_node_by_name(node, "BaseURL");
> +
> + // at now we can handle only one period, with the longest duration
> + node = xmlFirstElementChild(node);
> + while (node) {
> + if (!xmlStrcmp(node->name, (const xmlChar *)"Period")) {
> + perdiod_duration_sec = 0;
> + perdiod_start_sec = 0;
> + attr = node->properties;
> + while (attr) {
> + val = xmlGetProp(node, attr->name);
> + if (!xmlStrcmp(attr->name, (const xmlChar *)"duration")) {
> + perdiod_duration_sec = get_duration_insec(s, (const char *)val);
> + } else if (!xmlStrcmp(attr->name, (const xmlChar *)"start")) {
> + perdiod_start_sec = get_duration_insec(s, (const char *)val);
> + }
> + attr = attr->next;
> + xmlFree(val);
> + }
> + if ((perdiod_duration_sec) >= (c->period_duration)) {
> + period_node = node;
> + c->period_duration = perdiod_duration_sec;
> + c->period_start = perdiod_start_sec;
> + if (c->period_start > 0)
> + c->media_presentation_duration = c->period_duration;
> + }
> + }
> + node = xmlNextElementSibling(node);
> + }
> + if (!period_node) {
> + av_log(s, AV_LOG_ERROR, "Unable to parse '%s' - missing Period node\n", url);
> + ret = AVERROR_INVALIDDATA;
> + goto cleanup;
> + }
> +
> + adaptionset_node = xmlFirstElementChild(period_node);
> + while (adaptionset_node) {
> + if (!xmlStrcmp(adaptionset_node->name, (const xmlChar *)"BaseURL")) {
> + period_baseurl_node = adaptionset_node;
> + } else if (!xmlStrcmp(adaptionset_node->name, (const xmlChar *)"AdaptationSet")) {
> + parse_manifest_adaptationset(s, url, adaptionset_node, mpd_baseurl_node, period_baseurl_node);
> + }
> + adaptionset_node = xmlNextElementSibling(adaptionset_node);
> + }
> + if (c->cur_video) {
> + c->cur_video->rep_count = video_rep_idx;
> + c->cur_video->fix_multiple_stsd_order = 1;
> + av_log(s, AV_LOG_VERBOSE, "rep_idx[%d]\n", (int)c->cur_video->rep_idx);
> + av_log(s, AV_LOG_VERBOSE, "rep_count[%d]\n", (int)video_rep_idx);
> + }
> + if (c->cur_audio) {
> + c->cur_audio->rep_count = audio_rep_idx;
> + }
> +cleanup:
> + /*free the document */
> + xmlFreeDoc(doc);
> + xmlCleanupParser();
> + } else {
> + av_log(s, AV_LOG_ERROR, "Unable to read to offset '%s'\n", url);
> + ret = AVERROR_INVALIDDATA;
Having error handling so far from where it happens doesn't make it very
readable.
> + }
> +
> + av_free(new_url);
> + av_free(buffer);
> + if (close_in) {
> + avio_close(in);
> + }
> + return ret;
> +}
> +
> +static int64_t calc_cur_seg_no(AVFormatContext *s, struct representation *pls)
> +{
> + DASHContext *c = s->priv_data;
> + int64_t num = 0;
> + int64_t start_time_offset = 0;
> +
> + if (c->is_live) {
> + if (pls->n_fragments) {
> + num = pls->first_seq_no;
> + } else if (pls->n_timelines) {
> + start_time_offset = get_segment_start_time_based_on_timeline(pls, 0xFFFFFFFF) - pls->timelines[pls->first_seq_no]->t; // total duration of playlist
> + if (start_time_offset < 60 * pls->fragment_timescale)
> + start_time_offset = 0;
> + else
> + start_time_offset = start_time_offset - 60 * pls->fragment_timescale;
> +
> + num = calc_next_seg_no_from_timelines(pls, pls->timelines[pls->first_seq_no]->t + start_time_offset);
> + if (num == -1)
> + num = pls->first_seq_no;
> + } else {
> + if (pls->presentation_timeoffset) {
> + num = pls->presentation_timeoffset * pls->fragment_timescale / pls->fragment_duration;
> + } else if (c->publish_time > 0) {
> + num = pls->first_seq_no + (((c->publish_time - c->availability_start_time) - c->suggested_presentation_delay) * pls->fragment_timescale) / pls->fragment_duration;
> + } else {
> + num = pls->first_seq_no + (((get_current_time_in_sec() - c->availability_start_time) - c->suggested_presentation_delay) * pls->fragment_timescale) / pls->fragment_duration;
> + }
> + }
> + } else {
> + num = pls->first_seq_no;
> + }
> + return num;
> +}
> +
> +static int64_t calc_min_seg_no(AVFormatContext *s, struct representation *pls)
> +{
> + DASHContext *c = s->priv_data;
> + int64_t num = 0;
> +
> + if (c->is_live && pls->fragment_duration) {
> + num = pls->first_seq_no + (((get_current_time_in_sec() - c->availability_start_time) - c->time_shift_buffer_depth) * pls->fragment_timescale) / pls->fragment_duration;
> + } else {
> + num = pls->first_seq_no;
> + }
> + return num;
> +}
> +
> +static int64_t calc_max_seg_no(struct representation *pls)
> +{
> + DASHContext *c = pls->parent->priv_data;
> + int64_t num = 0;
> +
> + if (pls->n_fragments) {
> + num = pls->first_seq_no + pls->n_fragments - 1;
> + } else if (pls->n_timelines) {
> + int i = 0;
> + num = pls->first_seq_no + pls->n_timelines - 1;
> + for (i = 0; i < pls->n_timelines; i++) {
> + num += pls->timelines[i]->r;
> + }
> + } else if (c->is_live) {
> + num = pls->first_seq_no + (((get_current_time_in_sec() - c->availability_start_time)) * pls->fragment_timescale) / pls->fragment_duration;
> + } else {
> + num = pls->first_seq_no + (c->media_presentation_duration * pls->fragment_timescale) / pls->fragment_duration;
> + }
> +
> + return num;
> +}
> +
> +static void move_timelines(struct representation *rep_src, struct representation *rep_dest)
> +{
> + if (rep_dest && rep_src ) {
> + free_timelines_list(rep_dest);
> + rep_dest->timelines = rep_src->timelines;
> + rep_dest->n_timelines = rep_src->n_timelines;
> + rep_dest->first_seq_no = rep_src->first_seq_no;
> + rep_dest->last_seq_no = calc_max_seg_no(rep_dest);
> + rep_src->timelines = NULL;
> + rep_src->n_timelines = 0;
> + rep_dest->cur_seq_no = rep_src->cur_seq_no;
> + }
> +}
> +
> +static void move_segments(struct representation *rep_src, struct representation *rep_dest)
> +{
> + if (rep_dest && rep_src ) {
> + free_fragment_list(rep_dest);
> + if (rep_src->start_number > (rep_dest->start_number + rep_dest->n_fragments))
> + rep_dest->cur_seq_no = 0;
> + else
> + rep_dest->cur_seq_no += rep_src->start_number - rep_dest->start_number;
> + rep_dest->fragments = rep_src->fragments;
> + rep_dest->n_fragments = rep_src->n_fragments;
> + rep_dest->parent = rep_src->parent;
> + rep_dest->last_seq_no = calc_max_seg_no(rep_dest);
> + rep_src->fragments = NULL;
> + rep_src->n_fragments = 0;
> + }
> +}
> +
> +
> +static int refresh_manifest(AVFormatContext *s)
> +{
> +
> + int ret = 0;
> + DASHContext *c = s->priv_data;
> +
> + // save current context
> + struct representation *cur_video = c->cur_video;
> + struct representation *cur_audio = c->cur_audio;
> + char *base_url = c->base_url;
> +
> + c->base_url = NULL;
> + c->cur_video = NULL;
> + c->cur_audio = NULL;
> + ret = parse_manifest(s, s->filename, NULL);
> + if (ret)
> + goto finish;
> +
> + if (cur_video && cur_video->timelines || cur_audio && cur_audio->timelines) {
> + // calc current time
> + int64_t currentVideoTime = 0;
> + int64_t currentAudioTime = 0;
> + if (cur_video && cur_video->timelines)
> + currentVideoTime = get_segment_start_time_based_on_timeline(cur_video, cur_video->cur_seq_no) / cur_video->fragment_timescale;
> + if (cur_audio && cur_audio->timelines)
> + currentAudioTime = get_segment_start_time_based_on_timeline(cur_audio, cur_audio->cur_seq_no) / cur_audio->fragment_timescale;
> + // update segments
> + if (cur_video && cur_video->timelines) {
> + c->cur_video->cur_seq_no = calc_next_seg_no_from_timelines(c->cur_video, currentVideoTime * cur_video->fragment_timescale - 1);
> + if (c->cur_video->cur_seq_no >= 0) {
> + move_timelines(c->cur_video, cur_video);
> + }
> + }
> + if (cur_audio && cur_audio->timelines) {
> + c->cur_audio->cur_seq_no = calc_next_seg_no_from_timelines(c->cur_audio, currentAudioTime * cur_audio->fragment_timescale - 1);
> + if (c->cur_audio->cur_seq_no >= 0) {
> + move_timelines(c->cur_audio, cur_audio);
> + }
> + }
> + }
> + if (cur_video && cur_video->fragments) {
> + move_segments(c->cur_video, cur_video);
> + }
> + if (cur_audio && cur_audio->fragments) {
> + move_segments(c->cur_audio, cur_audio);
> + }
> +
> +finish:
> + // restore context
> + if (c->base_url)
> + av_free(base_url);
> + else
> + c->base_url = base_url;
> + if (c->cur_audio)
> + free_representation(c->cur_audio);
> + if (c->cur_video)
> + free_representation(c->cur_video);
> + c->cur_audio = cur_audio;
> + c->cur_video = cur_video;
> + return ret;
> +}
> +
> +static struct fragment *get_current_fragment(struct representation *pls)
> +{
> + int64_t min_seq_no = 0;
> + int64_t max_seq_no = 0;
> + struct fragment *seg = NULL;
> + struct fragment *seg_ptr = NULL;
> + DASHContext *c = pls->parent->priv_data;
> +
> + while (( !ff_check_interrupt(c->interrupt_callback)&& pls->n_fragments > 0)) {
> + if (pls->cur_seq_no < pls->n_fragments) {
> + seg_ptr = pls->fragments[pls->cur_seq_no];
> + seg = av_mallocz(sizeof(struct fragment));
> + if (!seg) {
> + return NULL;
> + }
> + seg->url = av_strdup(seg_ptr->url);
> + if (!seg->url) {
> + av_free(seg);
> + return NULL;
> + }
> + seg->size = seg_ptr->size;
> + seg->url_offset = seg_ptr->url_offset;
> + return seg;
> + } else if (c->is_live) {
> + av_usleep(1000*1000);
Why does it sleep an arbitrary amount of time?
> + refresh_manifest(pls->parent);
> + } else {
> + break;
> + }
> + }
> + if (c->is_live) {
> + while (!ff_check_interrupt(c->interrupt_callback)) {
> + min_seq_no = calc_min_seg_no(pls->parent, pls);
> + max_seq_no = calc_max_seg_no(pls);
> +
> + if (pls->cur_seq_no <= min_seq_no) {
> + av_log(pls->parent, AV_LOG_VERBOSE, "old fragment: cur[%"PRId64"] min[%"PRId64"] max[%"PRId64"], playlist %d\n", (int64_t)pls->cur_seq_no, min_seq_no, max_seq_no, (int)pls->rep_idx);
> + if (c->is_live && (pls->timelines || pls->fragments)) {
> + refresh_manifest(pls->parent);
> + }
> + pls->cur_seq_no = calc_cur_seg_no(pls->parent, pls);
> + } else if (pls->cur_seq_no > max_seq_no) {
> + av_log(pls->parent, AV_LOG_VERBOSE, "new fragment: min[%"PRId64"] max[%"PRId64"], playlist %d\n", min_seq_no, max_seq_no, (int)pls->rep_idx);
> + av_usleep(1000*1000);
Same. I also wonder if it wouldn't be better to avoid the duplicated
retry loop, if possible.
> + if (c->is_live && (pls->timelines || pls->fragments)) {
> + refresh_manifest(pls->parent);
> + }
> + continue;
> + }
> + break;
> + }
> + seg = av_mallocz(sizeof(struct fragment));
> + if (!seg) {
> + return NULL;
> + }
> + } else if (pls->cur_seq_no <= pls->last_seq_no) {
> + seg = av_mallocz(sizeof(struct fragment));
> + if (!seg) {
> + return NULL;
> + }
> + }
> + if (seg) {
> + if (pls->tmp_url_type != TMP_URL_TYPE_UNSPECIFIED) {
> + int64_t val = pls->tmp_url_type == TMP_URL_TYPE_NUMBER ? pls->cur_seq_no : get_segment_start_time_based_on_timeline(pls, pls->cur_seq_no);
> + int size = snprintf(NULL, 0, pls->url_template_format, val); // calc needed buffer size
Others have pointed it out: it's not ok to use input data as format
printf string. Also, I don't think passing NULL as buffer is allowed
(even if the buffer size is 0).
> +
> + if (size > 0) {
> + char *tmp_val = av_mallocz(size + 1);
> + snprintf(tmp_val, size+1, pls->url_template_format, val);
> + seg->url = av_strireplace(pls->url_template, pls->url_template_pattern, tmp_val);
> + av_free(tmp_val);
> + }
> + }
> +
> + if (!seg->url) {
> + av_log(pls->parent, AV_LOG_ERROR, "Unable to resolve template url '%s'\n", pls->url_template);
> + seg->url = av_strdup(pls->url_template);
> + if (!seg->url) {
> + return NULL;
> + }
> + }
> +
> + seg->size = -1;
> + }
> +
> + return seg;
> +}
> +
> +enum ReadFromURLMode {
> + READ_NORMAL,
> + READ_COMPLETE,
> +};
> +
> +static int read_from_url(struct representation *pls, struct fragment *seg,
> + uint8_t *buf, int buf_size,
> + enum ReadFromURLMode mode)
> +{
> + int ret;
> +
> + /* limit read if the fragment was only a part of a file */
> + if (seg->size >= 0)
> + buf_size = FFMIN(buf_size, pls->cur_seg_size - pls->cur_seg_offset);
> +
> + if (mode == READ_COMPLETE) {
> + ret = avio_read(pls->input, buf, buf_size);
> + if (ret < buf_size) {
> + av_log(pls->parent, AV_LOG_WARNING, "Could not read complete fragment.\n");
> + }
> + } else {
> + ret = avio_read(pls->input, buf, buf_size);
> + }
> + if (ret > 0)
> + pls->cur_seg_offset += ret;
> +
> + return ret;
> +}
> +
> +static int open_input(DASHContext *c, struct representation *pls, struct fragment *seg)
> +{
> + AVDictionary *opts = NULL;
> + char url[MAX_URL_SIZE];
> + int ret;
> +
> + // broker prior HTTP options that should be consistent across requests
> + av_dict_set(&opts, "user-agent", c->user_agent, 0);
> + av_dict_set(&opts, "cookies", c->cookies, 0);
> + av_dict_set(&opts, "headers", c->headers, 0);
> + if (c->is_live) {
> + av_dict_set(&opts, "seekable", "0", 0);
> + }
Code repeated from somewhere else in this patch.
> + if (seg->size >= 0) {
> + /* try to restrict the HTTP request to the part we want
> + * (if this is in fact a HTTP request) */
> + av_dict_set_int(&opts, "offset", seg->url_offset, 0);
> + av_dict_set_int(&opts, "end_offset", seg->url_offset + seg->size, 0);
> + }
> +
> + ff_make_absolute_url(url, MAX_URL_SIZE, c->base_url, seg->url);
> + av_log(pls->parent, AV_LOG_VERBOSE, "DASH request for url '%s', offset %"PRId64", playlist %d\n",
> + url, seg->url_offset, pls->rep_idx);
> + ret = open_url(pls->parent, &pls->input, url, c->avio_opts, opts, NULL);
> + if (ret < 0) {
> + goto cleanup;
> + }
> +
> + /* Seek to the requested position. If this was a HTTP request, the offset
> + * should already be where want it to, but this allows e.g. local testing
> + * without a HTTP server. */
> + if (!ret && seg->url_offset) {
> + int64_t seekret = avio_seek(pls->input, seg->url_offset, SEEK_SET);
> + if (seekret < 0) {
> + av_log(pls->parent, AV_LOG_ERROR, "Unable to seek to offset %"PRId64" of DASH fragment '%s'\n", seg->url_offset, seg->url);
> + ret = (int) seekret;
> + ff_format_io_close(pls->parent, &pls->input);
> + }
> + }
> +
> +cleanup:
> + av_dict_free(&opts);
> + pls->cur_seg_offset = 0;
> + pls->cur_seg_size = seg->size;
> + return ret;
> +}
> +
> +static int update_init_section(struct representation *pls)
> +{
> + static const int max_init_section_size = 1024*1024;
> + DASHContext *c = pls->parent->priv_data;
> + int64_t sec_size = 0;
> + int64_t urlsize = 0;
> + int ret = 0;
> +
> + /* read init section only once per representation */
> + if (!pls->init_section || pls->init_sec_buf) {
> + return 0;
> + }
> +
> + ret = open_input(c, pls, pls->init_section);
> + if (ret < 0) {
> + av_log(pls->parent, AV_LOG_WARNING, "Failed to open an initialization section in playlist %d\n", pls->rep_idx);
> + return ret;
> + }
> +
> + if (pls->init_section->size >= 0) {
> + sec_size = pls->init_section->size;
> + } else if ((urlsize = avio_size(pls->input)) >= 0) {
> + sec_size = urlsize;
> + } else {
> + sec_size = max_init_section_size;
> + }
> + av_log(pls->parent, AV_LOG_DEBUG, "Downloading an initialization section of size %"PRId64"\n", sec_size);
> + sec_size = FFMIN(sec_size, max_init_section_size);
> + av_fast_malloc(&pls->init_sec_buf, &pls->init_sec_buf_size, sec_size);
Doesn't that need an error check or so?
> + ret = read_from_url(pls, pls->init_section, pls->init_sec_buf, pls->init_sec_buf_size, READ_COMPLETE);
> + ff_format_io_close(pls->parent, &pls->input);
> + if (ret < 0)
> + return ret;
> +
> + if (pls->fix_multiple_stsd_order && pls->rep_idx > 0) {
> + uint8_t **stsd_entries = NULL;
> + int *stsd_entries_size = NULL;
> + int i = 4;
> +
> + while (i <= (ret - 4)) {
> + // find start stsd atom
> + if (!memcmp(pls->init_sec_buf + i, "stsd", 4)) {
> + /* 1B version
> + * 3B flags
> + * 4B num of entries */
> + int stsd_first_offset = i + 8;
> + int stsd_offset = 0;
> + int j = 0;
> + uint32_t stsd_count = AV_RB32(pls->init_sec_buf + stsd_first_offset);
> + stsd_first_offset += 4;
> + if (stsd_count != pls->rep_count) {
> + i++;
> + continue;
> + }
> + // find all stsd entries
> + stsd_entries = av_mallocz_array(stsd_count, sizeof(*stsd_entries));
> + stsd_entries_size = av_mallocz_array(stsd_count, sizeof(*stsd_entries_size));
> + for (j = 0; j < stsd_count; ++j) {
> + /* 4B - size
> + * 4B - format */
> + stsd_entries_size[j] = AV_RB32(pls->init_sec_buf + stsd_first_offset + stsd_offset);
> + stsd_entries[j] = av_malloc(stsd_entries_size[j]);
> + memcpy(stsd_entries[j], pls->init_sec_buf + stsd_first_offset + stsd_offset, stsd_entries_size[j]);
> + stsd_offset += stsd_entries_size[j];
> + }
> + // reorder stsd entries
> + // as first put stsd entry for current representation
> + j = pls->rep_idx;
> + stsd_offset = stsd_first_offset;
> + memcpy(pls->init_sec_buf + stsd_offset, stsd_entries[j], stsd_entries_size[j]);
> + stsd_offset += stsd_entries_size[j];
> + for (j = 0; j < stsd_count; ++j) {
> + if (j != pls->rep_idx) {
> + memcpy(pls->init_sec_buf + stsd_offset, stsd_entries[j], stsd_entries_size[j]);
> + stsd_offset += stsd_entries_size[j];
> + }
> + av_free(stsd_entries[j]);
> + }
> + av_freep(&stsd_entries);
> + av_freep(&stsd_entries_size);
> + break;
> + }
> + i++;
> + }
GAAAAAAH. Is that really necessary, and what does it do?
> + }
> +
> + av_log(pls->parent, AV_LOG_TRACE, "pls[%p] init section size[%d]\n", pls, (int)ret);
> + pls->init_sec_data_len = ret;
> + pls->init_sec_buf_read_offset = 0;
> +
> + return 0;
> +}
> +
> +static int64_t seek_data(void *opaque, int64_t offset, int whence)
> +{
> + struct representation *v = opaque;
> + if (v->n_fragments && !v->init_sec_data_len) {
> + return avio_seek(v->input, offset, whence);
> + }
> +
> + return AVERROR(ENOSYS);
> +}
> +
> +static int read_data(void *opaque, uint8_t *buf, int buf_size)
> +{
> + int ret = 0;
> + struct representation *v = opaque;
> + DASHContext *c = v->parent->priv_data;
> +
> +restart:
> + if (!v->input) {
> + free_fragment(&v->cur_seg);
> + v->cur_seg = get_current_fragment(v);
> + if (!v->cur_seg) {
> + ret = AVERROR_EOF;
> + goto end;
> + }
> +
> + /* load/update Media Initialization Section, if any */
> + ret = update_init_section(v);
> + if (ret)
> + goto end;
> +
> + ret = open_input(c, v, v->cur_seg);
> + if (ret < 0) {
> + if (ff_check_interrupt(c->interrupt_callback)) {
> + goto end;
> + ret = AVERROR_EXIT;
> + }
> + av_log(v->parent, AV_LOG_WARNING, "Failed to open fragment of playlist %d\n", v->rep_idx);
> + v->cur_seq_no++;
> + goto restart;
Apparently a third retry loop... this time without av_usleep, and with
a goto instead of a loop.
> + }
> + }
> +
> + if (v->init_sec_buf_read_offset < v->init_sec_data_len) {
> + /* Push init section out first before first actual fragment */
> + int copy_size = FFMIN(v->init_sec_data_len - v->init_sec_buf_read_offset, buf_size);
> + memcpy(buf, v->init_sec_buf, copy_size);
> + v->init_sec_buf_read_offset += copy_size;
> + ret = copy_size;
> + goto end;
> + }
> +
> + /* check the v->cur_seg, if it is null, get current and double check if the new v->cur_seg*/
> + if (!v->cur_seg) {
> + v->cur_seg = get_current_fragment(v);
> + }
> + if (!v->cur_seg) {
> + ret = AVERROR_EOF;
> + goto end;
> + }
> + ret = read_from_url(v, v->cur_seg, buf, buf_size, READ_NORMAL);
> + if (ret > 0)
> + goto end;
> +
> + if (!v->is_restart_needed)
> + v->cur_seq_no++;
> + v->is_restart_needed = 1;
> +
> +/*
> + ff_format_io_close(v->parent, &v->input);
> + v->cur_seq_no++;
> + goto restart;
> +*/
> +end:
> + return ret;
> +}
> +
> +static int save_avio_options(AVFormatContext *s)
> +{
> + DASHContext *c = s->priv_data;
> + const char *opts[] = { "headers", "user_agent", "user-agent", "cookies", NULL }, **opt = opts;
> + uint8_t *buf = NULL;
> + int ret = 0;
> +
> + while (*opt) {
> + if (av_opt_get(s->pb, *opt, AV_OPT_SEARCH_CHILDREN, &buf) >= 0) {
> + if (buf[0] != '\0') {
> + ret = av_dict_set(&c->avio_opts, *opt, buf, AV_DICT_DONT_STRDUP_VAL);
> + if (ret < 0)
> + return ret;
> + }
> + }
> + opt++;
> + }
> +
> + return ret;
> +}
> +
> +static int nested_io_open(AVFormatContext *s, AVIOContext **pb, const char *url,
> + int flags, AVDictionary **opts)
> +{
> + av_log(s, AV_LOG_ERROR,
> + "A DASH playlist item '%s' referred to an external file '%s'. "
> + "Opening this file was forbidden for security reasons\n",
> + s->filename, url);
> + return AVERROR(EPERM);
> +}
> +
> +static int reopen_demux_for_component(AVFormatContext *s, struct representation *pls)
> +{
> + DASHContext *c = s->priv_data;
> + AVInputFormat *in_fmt = NULL;
> + AVDictionary *in_fmt_opts = NULL;
> + uint8_t *avio_ctx_buffer = NULL;
> + int ret = 0;
> +
> + if (pls->ctx) {
> + /* note: the internal buffer could have changed, and be != avio_ctx_buffer */
> + av_freep(&pls->pb.buffer);
> + memset(&pls->pb, 0x00, sizeof(AVIOContext));
> + pls->ctx->pb = NULL;
> + avformat_close_input(&pls->ctx);
> + pls->ctx = NULL;
> + }
> + if (!(pls->ctx = avformat_alloc_context())) {
> + ret = AVERROR(ENOMEM);
> + goto fail;
> + }
> +
> + avio_ctx_buffer = av_malloc(INITIAL_BUFFER_SIZE);
> + if (!avio_ctx_buffer ) {
> + ret = AVERROR(ENOMEM);
> + avformat_free_context(pls->ctx);
> + pls->ctx = NULL;
> + goto fail;
> + }
> + if (c->is_live) {
> + ffio_init_context(&pls->pb, avio_ctx_buffer , INITIAL_BUFFER_SIZE, 0, pls, read_data, NULL, NULL);
> + } else {
> + ffio_init_context(&pls->pb, avio_ctx_buffer , INITIAL_BUFFER_SIZE, 0, pls, read_data, NULL, seek_data);
> + }
> + pls->pb.seekable = 0;
> +
> + if ((ret = ff_copy_whiteblacklists(pls->ctx, s)) < 0)
> + goto fail;
> +
> + pls->ctx->flags = AVFMT_FLAG_CUSTOM_IO;
> + pls->ctx->probesize = 1024 * 4;
> + pls->ctx->max_analyze_duration = 4 * AV_TIME_BASE;
> + ret = av_probe_input_buffer(&pls->pb, &in_fmt, "", NULL, 0, 0);
> + if (ret < 0) {
> + av_log(s, AV_LOG_ERROR, "Error when loading first fragment, playlist %d\n", (int)pls->rep_idx);
> + avformat_free_context(pls->ctx);
> + pls->ctx = NULL;
> + goto fail;
> + }
> +
> + pls->ctx->pb = &pls->pb;
> + pls->ctx->io_open = nested_io_open;
> +
> + // provide additional information from mpd if available
> + ret = avformat_open_input(&pls->ctx, "", in_fmt, &in_fmt_opts); //pls->init_section->url
> + av_dict_free(&in_fmt_opts);
> + if (ret < 0)
> + goto fail;
> + if (pls->n_fragments) {
> + ret = avformat_find_stream_info(pls->ctx, NULL);
> + if (ret < 0)
> + goto fail;
> + }
> +
> +fail:
> + return ret;
> +}
> +
> +static int open_demux_for_component(AVFormatContext *s, struct representation *pls)
> +{
> + int ret = 0;
> + int i;
> +
> + pls->parent = s;
> + pls->cur_seq_no = calc_cur_seg_no(s, pls);
> + pls->last_seq_no = calc_max_seg_no(pls);
> +
> + ret = reopen_demux_for_component(s, pls);
> + if (ret < 0) {
> + goto fail;
> + }
> + for (i = 0; i < pls->ctx->nb_streams; i++) {
> + AVStream *st = avformat_new_stream(s, NULL);
> + AVStream *ist = pls->ctx->streams[i];
> + if (!st) {
> + ret = AVERROR(ENOMEM);
> + goto fail;
> + }
> + st->id = i;
> + avcodec_parameters_copy(st->codecpar, pls->ctx->streams[i]->codecpar);
> + avpriv_set_pts_info(st, ist->pts_wrap_bits, ist->time_base.num, ist->time_base.den);
> + }
> +
> + return 0;
> +fail:
> + return ret;
> +}
> +
> +static int dash_read_header(AVFormatContext *s)
> +{
> + void *u = (s->flags & AVFMT_FLAG_CUSTOM_IO) ? NULL : s->pb;
> + DASHContext *c = s->priv_data;
> + int ret = 0;
> + int stream_index = 0;
> +
> + c->interrupt_callback = &s->interrupt_callback;
> + // if the URL context is good, read important options we must broker later
> + if (u) {
> + update_options(&c->user_agent, "user-agent", u);
> + update_options(&c->cookies, "cookies", u);
> + update_options(&c->headers, "headers", u);
> + }
> +
> + if ((ret = parse_manifest(s, s->filename, s->pb)) < 0)
> + goto fail;
> +
> + if ((ret = save_avio_options(s)) < 0)
> + goto fail;
> +
> + /* If this isn't a live stream, fill the total duration of the
> + * stream. */
> + if (!c->is_live) {
> + s->duration = (int64_t) c->media_presentation_duration * AV_TIME_BASE;
> + }
> +
> + /* Open the demuxer for curent video and current audio components if available */
> + if (!ret && c->cur_video) {
> + ret = open_demux_for_component(s, c->cur_video);
> + if (!ret) {
> + c->cur_video->stream_index = stream_index;
> + ++stream_index;
> + } else {
> + free_representation(c->cur_video);
> + c->cur_video = NULL;
> + }
> + }
> +
> + if (!ret && c->cur_audio) {
> + ret = open_demux_for_component(s, c->cur_audio);
> + if (!ret) {
> + c->cur_audio->stream_index = stream_index;
> + ++stream_index;
> + } else {
> + free_representation(c->cur_audio);
> + c->cur_audio = NULL;
> + }
> + }
> +
> + if (!stream_index) {
> + ret = AVERROR_INVALIDDATA;
> + goto fail;
> + }
> +
> + /* Create a program */
> + if (!ret) {
> + AVProgram *program;
> + program = av_new_program(s, 0);
> + if (!program) {
> + goto fail;
> + }
> +
> + if (c->cur_video) {
> + av_program_add_stream_index(s, 0, c->cur_video->stream_index);
> + }
> + if (c->cur_audio) {
> + av_program_add_stream_index(s, 0, c->cur_audio->stream_index);
> + }
> + }
> +
> + return 0;
> +fail:
> + return ret;
> +}
> +
> +static int dash_read_packet(AVFormatContext *s, AVPacket *pkt)
> +{
> + DASHContext *c = s->priv_data;
> + int ret = 0;
> + struct representation *cur = NULL;
> +
> + if (!c->cur_audio && !c->cur_video ) {
> + return AVERROR_INVALIDDATA;
> + }
> + if (c->cur_audio && !c->cur_video) {
> + cur = c->cur_audio;
> + } else if (!c->cur_audio && c->cur_video) {
> + cur = c->cur_video;
> + } else if (c->cur_video->cur_timestamp < c->cur_audio->cur_timestamp) {
> + cur = c->cur_video;
> + } else {
> + cur = c->cur_audio;
> + }
> +
> + if (cur->ctx) {
> + while (!ff_check_interrupt(c->interrupt_callback) && !ret) {
> + ret = av_read_frame(cur->ctx, pkt);
> + if (ret >= 0) {
> + /* If we got a packet, return it */
> + cur->cur_timestamp = av_rescale(pkt->pts, (int64_t)cur->ctx->streams[0]->time_base.num * 90000, cur->ctx->streams[0]->time_base.den);
> + pkt->stream_index = cur->stream_index;
> + return 0;
> + }
> + if (cur->is_restart_needed) {
> + while (!ff_check_interrupt(c->interrupt_callback)) {
> + cur->cur_seg_offset = 0;
> + cur->init_sec_buf_read_offset = 0;
> + if (cur->input)
> + ff_format_io_close(cur->parent, &cur->input);
> + ret = reopen_demux_for_component(s, cur);
> + if (c->is_live && ret) {
> + av_usleep(1000*1000);
> + continue;
A 4th retry loop! This time it's following the pattern of the 1st and
2nd loop.
> + }
> + break;
> + }
> + cur->is_restart_needed = 0;
> + }
> +
> + }
> + }
> + return AVERROR_EOF;
> +}
> +
> +static int dash_close(AVFormatContext *s)
> +{
> + DASHContext *c = s->priv_data;
> + if (c->cur_audio) {
> + free_representation(c->cur_audio);
> + }
> + if (c->cur_video) {
> + free_representation(c->cur_video);
> + }
> +
> + av_freep(&c->cookies);
> + av_freep(&c->user_agent);
> + av_dict_free(&c->avio_opts);
> + av_freep(&c->base_url);
> + return 0;
> +}
> +
> +static int dash_seek(AVFormatContext *s, struct representation *pls, int64_t seek_pos_msec, int flags)
> +{
> + int ret = 0;
> + int i = 0;
> + int j = 0;
> + int64_t duration = 0;
> +
> + av_log(pls->parent, AV_LOG_VERBOSE, "DASH seek pos[%"PRId64"ms], playlist %d\n", seek_pos_msec, pls->rep_idx);
> +
> + // single fragment mode
> + if (pls->n_fragments == 1) {
> + pls->cur_timestamp = 0;
> + pls->cur_seg_offset = 0;
> + ff_read_frame_flush(pls->ctx);
> + return av_seek_frame(pls->ctx, -1, seek_pos_msec * 1000, flags);
> + }
> +
> + if (pls->input)
> + ff_format_io_close(pls->parent, &pls->input);
> +
> + // find the nearest fragment
> + if (pls->n_timelines > 0 && pls->fragment_timescale > 0) {
> + int64_t num = pls->first_seq_no;
> + av_log(pls->parent, AV_LOG_VERBOSE, "dash_seek with SegmentTimeline start n_timelines[%d] "
> + "last_seq_no[%"PRId64"], playlist %d.\n",
> + (int)pls->n_timelines, (int64_t)pls->last_seq_no, (int)pls->rep_idx);
> + for (i = 0; i < pls->n_timelines; i++) {
> + if (pls->timelines[i]->t > 0) {
> + duration = pls->timelines[i]->t;
> + }
> + duration += pls->timelines[i]->d;
> + if (seek_pos_msec < ((duration * 1000) / pls->fragment_timescale)) {
> + goto set_seq_num;
> + }
> + for (j = 0; j < pls->timelines[i]->r; j++) {
> + duration += pls->timelines[i]->d;
> + num++;
> + if (seek_pos_msec < ((duration * 1000) / pls->fragment_timescale)) {
> + goto set_seq_num;
> + }
> + }
> + num++;
> + }
> +
> +set_seq_num:
> + pls->cur_seq_no = num > pls->last_seq_no ? pls->last_seq_no : num;
> + av_log(pls->parent, AV_LOG_VERBOSE, "dash_seek with SegmentTimeline end cur_seq_no[%"PRId64"], playlist %d.\n",
> + (int64_t)pls->cur_seq_no, (int)pls->rep_idx);
> + } else if (pls->fragment_duration > 0) {
> + pls->cur_seq_no = pls->first_seq_no + ((seek_pos_msec * pls->fragment_timescale) / pls->fragment_duration) / 1000;
> + } else {
> + av_log(pls->parent, AV_LOG_ERROR, "dash_seek missing fragment_duration\n");
> + pls->cur_seq_no = pls->first_seq_no;
> + }
> + pls->cur_timestamp = 0;
> + pls->cur_seg_offset = 0;
> + pls->init_sec_buf_read_offset = 0;
> + ret = reopen_demux_for_component(s, pls);
> +
> + return ret;
> +}
> +
> +static int dash_read_seek(AVFormatContext *s, int stream_index, int64_t timestamp, int flags)
> +{
> + int ret = 0;
> + DASHContext *c = s->priv_data;
> + int64_t seek_pos_msec = av_rescale_rnd(timestamp, 1000,
> + s->streams[stream_index]->time_base.den,
> + flags & AVSEEK_FLAG_BACKWARD ?
> + AV_ROUND_DOWN : AV_ROUND_UP);
> + if ((flags & AVSEEK_FLAG_BYTE) || c->is_live)
> + return AVERROR(ENOSYS);
> + if (c->cur_audio) {
> + ret = dash_seek(s, c->cur_audio, seek_pos_msec, flags);
> + }
> + if (!ret && c->cur_video) {
> + ret = dash_seek(s, c->cur_video, seek_pos_msec, flags);
> + }
> + return ret;
> +}
> +
> +static int dash_probe(AVProbeData *p)
> +{
> + if (!av_stristr(p->buf, "<MPD"))
> + return 0;
> +
> + if (av_stristr(p->buf, "dash:profile:isoff-on-demand:2011") ||
> + av_stristr(p->buf, "dash:profile:isoff-live:2011") ||
> + av_stristr(p->buf, "dash:profile:isoff-live:2012") ||
> + av_stristr(p->buf, "dash:profile:isoff-main:2011")) {
> + return AVPROBE_SCORE_MAX;
> + }
> + if (av_stristr(p->buf, "dash:profile")) {
> + return AVPROBE_SCORE_MAX / 2;
> + }
> +
> + return 0;
> +}
> +
> +#define OFFSET(x) offsetof(DASHContext, x)
> +#define FLAGS AV_OPT_FLAG_DECODING_PARAM
> +static const AVOption dash_options[] = {
> + {NULL}
> +};
> +
> +static const AVClass dash_class = {
> + .class_name = "dash",
> + .item_name = av_default_item_name,
> + .option = dash_options,
> + .version = LIBAVUTIL_VERSION_INT,
> +};
> +
> +AVInputFormat ff_dash_demuxer = {
> + .name = "dash",
> + .long_name = NULL_IF_CONFIG_SMALL("Dynamic Adaptive Streaming over HTTP"),
> + .priv_class = &dash_class,
> + .priv_data_size = sizeof(DASHContext),
> + .read_probe = dash_probe,
> + .read_header = dash_read_header,
> + .read_packet = dash_read_packet,
> + .read_close = dash_close,
> + .read_seek = dash_read_seek,
> + .flags = AVFMT_NO_BYTE_SEEK,
> +};
More information about the ffmpeg-devel
mailing list