flv文件的onMetaData部分解析

参考flv的spec文档，写了个解析onMetaData，代码没有整理

/*
 * Copyright (c) 2013 Steven Liu
 *
 *
 * 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 <stdio.h>
#include <stdlib.h>
#include <string.h>
#include <unistd.h>
#include <fcntl.h>
#include <sys/types.h>
#include <sys/stat.h>
#include <errno.h>

#define TAG_FLV_F 'F'
#define TAG_FLV_L 'L'
#define TAG_FLV_V 'V'
#define TAG_TYPE_SCRIPTDATA     0x12
int script_type_parse(unsigned char *data);

union av_intfloat64 {
    uint64_t i;
    double   f;
};

enum script_data_type {
    Number = 0,
    Boolean,
    String,
    Object,
    MovieClip,
    Null,
    Undefined,
    Reference,
    EcmaArray,
    ObjectEndMarker,
    StringArray,
    Date,
    LongString,
};
typedef enum script_data_type script_type_t;

struct script_string {
    char len[2];
};
typedef struct script_string String_t;

struct flv_header {
    unsigned char tag_f;
    unsigned char tag_l;
    unsigned char tag_v;
    unsigned char version;
    unsigned char audio_and_video;
    unsigned char DataOffset[4];
};
typedef struct flv_header header_t;

struct flv_tag {
    unsigned char tagtype;
    unsigned char size[3];
    unsigned char timestamp[3];
    unsigned char extent_timestamp;
    unsigned char stream_id[3];
};
typedef struct flv_tag tag_t;

struct flv_body {
    tag_t tag;
    // unsigned char PreviousTagSize[2];
};
typedef struct flv_body body_t;

struct object_name {
    unsigned short length;
    unsigned char name;
};
typedef struct object_name object_name_t;

struct script_object {
    object_name_t name;
};
typedef struct script_object_name script_object_t;


/**
 * Reinterpret a 64-bit integer as a double.
 */
static double int2double(uint64_t i)
{
    union av_intfloat64 v;
    v.i = i;
    return v.f;
}

/*
 * get_header_audio check if the flv have audio stream
 * info, a byte with video and audio bit flag
 *
 * return 0 or 1
*/
int get_header_audio(char info)
{
    return (info & 0x4) >> 2;
}

/*
 * get_header_video check if the flv have video stream
 * info, a byte with video and audio bit flag
 *
 * return 0 or 1
*/
int get_header_video(char info)
{
    return info & 0x1;
}

/*
 * flv_header_parse just parse the FLV VERSION tag
 * fd, file describe handle
 * return flv version
*/
int flv_header_parse(int fd)
{
    int ret = 0;
    header_t header;

    memset(&header, 0, sizeof(header));
    ret = read(fd, &header, sizeof(header));
    if (ret <= 0) {
        fprintf(stderr, "read header error, %d\n", -EINVAL);
        return -EINVAL;
    }


    fprintf(stdout, "flv = [%c %c %c] version = [%x]\n"
            "TypeFlagsAudio = [%x], TypeFlagsVideo = [%x], DataOffset = [%x %x %x %x]\n",
            header.tag_f, header.tag_l, header.tag_v, header.version,
            get_header_audio(header.audio_and_video), get_header_video(header.audio_and_video),
            header.DataOffset[0], header.DataOffset[1], header.DataOffset[2], header.DataOffset[3]);

    return (int)header.version;
}


/*
 * flv_do_tag just parse all the flv tag
 * fd, file describe handle
 * size, the size of the flv tag body
 *
 * return tag type
*/
int flv_do_tag(int fd, unsigned long *size)
{
    int ret = 0;
    char *p = NULL;
    char data_size[11] = {0};
    body_t body;


    ret = read(fd, &body, sizeof(body));
    if (ret <= 0) {
        fprintf(stderr, "Can not ret the fd errno [%d]\n", -EINVAL);
        return -EINVAL;
    }

    fprintf(stdout, "size = %lu tag->type = [%x]"
            "tag->size = [%0.2x %0.2x %0.2x], tag->timestamp = [%0.2x %0.2x %0.2x], "
            "tag->extent_timestamp = [%0.2x], tag->streamid = [%0.2x %0.2x %0.2x]\n",
            sizeof(body), body.tag.tagtype,
            body.tag.size[0], body.tag.size[1],body.tag.size[2], body.tag.timestamp[0], body.tag.timestamp[1],
            body.tag.timestamp[2], body.tag.extent_timestamp, body.tag.stream_id[0],
            body.tag.stream_id[1], body.tag.stream_id[2]);
    memcpy(data_size, body.tag.size, 3);
    p = data_size;
    snprintf(data_size, sizeof(data_size), "0x%x%x%x", body.tag.size[0], body.tag.size[1], body.tag.size[2]);
    *size = strtoul(p, NULL, 16);
    fprintf(stdout, "size = [%lu] \n", *size);

    return body.tag.tagtype;
}

/*
 * get_string_len get the key-value value string length
 * data, the data will be parsed
 *
 * return the String length by int
*/
int get_string_len(unsigned char *data)
{

    unsigned char *p = data;
    unsigned char len_char[8];
    int len_int = 0;

    memset(len_char, 0, sizeof(len_char));
    snprintf((char *)len_char, sizeof(len_char), "0x%x%x", data[0], data[1]);
    p = len_char;

    len_int = strtoul((char *)p, NULL, 16);

    return len_int;
}

/*
 * get_bool_value get bool type value (key-value value)
 * data, the data will be parsed
 *
 * return bool value 0 or 1
*/
int get_bool_value(unsigned char *data)
{
   return *data;
}

/*
 * get_key_len the key-value key string length unit: four bytes
 * data, the data will be parsed
 *
 * return the key length by int
*/
int get_key_len(unsigned char *data)
{
    char len_char[16];
    char *p = len_char;

    memset(len_char, 0, sizeof(len_char));
    snprintf(len_char, sizeof(len_char), "0x%x%x%x%x", data[0], data[1], data[2], data[3]);

    return strtoul(p, NULL, 16);
}

int process_ecma_array_end(unsigned char *data)
{
    return 0;
}

/*
 * process_ecma_array The onMetaData sub type named ECMA Array
 * data, the data will be parsed
 *
 * return the offset of the data
*/
int process_ecma_array(unsigned char *data)
{
    int ecma_array_len = 0;
    int keyname_len = 0;
    unsigned char keyname[32];
    unsigned char *p = data;
    int i = 0;

    ecma_array_len = get_key_len(p);
    p += 4;

    for (i = 0; i < ecma_array_len; i++) {
        keyname_len = get_string_len(p);
        p += 2;
        memset(keyname, 0, sizeof(keyname));
        strncpy((char *)keyname, (const char *)p, keyname_len);
        fprintf(stdout, "keyname = [%s]\n", keyname);
        p += keyname_len;
        p += script_type_parse(p);
    }

    if (*p == 0 && *(p + 1) == 0 && *(p + 2) == 9) {
        p += 3;
    }
    return p - data;
}

/*
 * get_double get the double number by long long type
 * data, the data will be parsed
 *
 * return the parse out number of long long
*/
unsigned long long get_double(unsigned char *data)
{
    unsigned char double_number[64];
    unsigned char *p = double_number;

    memset(double_number, 0, sizeof(double_number));
    snprintf((char *)double_number, sizeof(double_number), "0x%0.2x%0.2x%0.2x%0.2x%0.2x%0.2x%0.2x%0.2x",
             *data, *(data + 1), *(data + 2), *(data + 3), *(data + 4), *(data + 5), *(data + 6), *(data + 7));

    return strtoull((char *)p, NULL, 16);

}

/*
 * script_type_parse parse the script object type and parse the data
 * data, the data will be parsed
 *
 * return the data offset
*/
int script_type_parse(unsigned char *data)
{
    int ret = 0;
    unsigned char *p = data;
    unsigned char string_output[512];
    unsigned long long number = 0;
    double double_number = 0.0;

    switch (*p) {
    case Number:
        p++;
        number = get_double(p);
        double_number = int2double(number);
        fprintf(stdout, "number = [%.2f]\n", double_number);
        p += 8;
        break;

    case Boolean:
        p++;
        ret = get_bool_value(p);
        p++;
        break;

    case String:
        p++;
        memset(string_output, 0, sizeof(string_output));
        ret = get_string_len(p);
        p += 2;
        strncpy((char *)string_output, (const char *)p, ret);
        fprintf(stdout, "String = [%s]\n", string_output);
        p += ret;
        break;

    case Object:
        p++;
        break;

    case MovieClip:
        p++;
        break;

    case Null:
        p++;
        break;

    case Undefined:
        p++;
        break;

    case Reference:
        p++;
        break;

    case EcmaArray:
        p++;
        ret = process_ecma_array(p);
        p += ret;
        break;

    case ObjectEndMarker:
        p++;
        break;

    case StringArray:
        p++;
        break;

    case Date:
        p++;
        break;

    case LongString:
        p++;
        break;

    default:

        break;
    }
    return p - data;
}

/*
 * do_tag_onMetaData parse the onMetaData tags
 * fd, file describe handle
 * size, will read size of the data, unit: byte
 *
 * return offset of the data
*/
int do_tag_onMetaData(int fd, int size)
{
    int ret = 0;
    unsigned char *p = NULL;

    unsigned char *onMetaData = malloc(size);
    fprintf(stdout, "size = [%d]\n", size);
    memset(onMetaData, 0, size);
    ret = read(fd, onMetaData, size);
    fprintf(stdout, "ret = [%d]\n", ret);

    int fuck = open("./xxx.dat", O_RDWR);
    write(fuck, onMetaData, ret);
    close(fuck);
    p = onMetaData;

    while (p) {
        if (p - onMetaData >= size) {
            break;
        }
        ret = script_type_parse(p);
        fprintf(stdout, "in while 1 offset = [%d]\n", p - onMetaData);
        p += ret;
    }
    return p - onMetaData;
}

/*
 * dump_flv_info just print the flv onMetaData info
 * filename, the flv filename
 *
 * return always right
*/
int dump_flv_info(char *filename)
{
    int fd = 0;
    int tag_type = -1;
    int ret = 0;
    unsigned long size = 0;
    unsigned char PreviousTagSize[4];

    fd = open(filename, O_RDONLY);
    if (fd < 0) {
        fprintf(stdout, "Can not open the file [%s]\n", filename);
        return -ENOENT;
    }
    fprintf(stdout, "filename = [%s]\n", filename);

    flv_header_parse(fd);
    ret = read(fd, PreviousTagSize, sizeof(PreviousTagSize));
    tag_type = flv_do_tag(fd, &size);

    switch(tag_type) {
    case TAG_TYPE_SCRIPTDATA:
        fprintf(stdout, "data_size = [%lu]\n", size);
        do_tag_onMetaData(fd, size);

        break;
    }
    return 0;
}

void usage(char **argv)
{
    fprintf(stderr, "Usage:\n"
            "\t%s flv_file_name\n", argv[0]);
    return ;
}

int main(int argc, char *argv[])
{
    char *filename = NULL;

    if (argc < 2) {
        usage(argv);
        return -EINVAL;
    }

    filename = argv[1];

    dump_flv_info(filename);
    return 0;
}

Makefile为

all:
    gcc -O2 -Wall -g dump_flv_info.c -o flv_info

clean:
    rm -rf flv_info

执行结果如下：

2.flv的内容：

主要参考文档为：
video_file_format_spec_v10_1_20131118111845.pdf

根据参考文档可以看出来

解析步骤如下

首先解析文件头

根据上面的内容可以看出来，FLV和Version占用了4个字节，而Version总是1,其中的DataOffset一般来在flv的version是1的时候总是9，如图

464c 5601为四个字节，就是FLV 和Version部分，接下来的05为一个字节，这个字节为TypeFlagsReserved, TypeFlagsAudio,TyopeFlagsReserved, TypeFlagsVideo的内容，可以将05转换为2进制，就是00000101, 也就是video是1，audio也是1，接下爱就是DataOffset，DataOffset一次读取32位，就是4个字节，内容为00 0000 09，也就是上面参考中的内容，The DataOffset field usually has a value of 9；

接下来就开始解析FLV的File Body:

部分，这部分排列方式位PreviousTagSizeN（N >= 0）,如图

占用了4个字节，全部为0.在前面的spec里面可以看到，接下来的就是FLVTAG，这个FLVTAG为另外一个结构，结构如下：

通过结构可以看出，Reserved,Filter,TagType占用了一个字节，也就是前面的12，这里说的12其实不是十进制的12，而是16进制的0x12，将其转成十进制就是18，可以转成2进制进行对比，最终TagType来判断这个FLVTAG是音频，还是视频，还是脚本数据(ScriptData)，如果TagType是8就是audio，如果TagType是9就是video，如果TagType是18，就是ScriptData；接下来继续读取24个字节，来判断将要读取得FLVTAG对应的数据字节数，例如onMetaData这部分是ScriptData，那么读取ScriptData的数据的时候，就要读取DataSize这么大，如图：

可以看到，需要读取0x000125这么多数据，转换成十进制的话，就是293个字节的数据，这个数据我们暂且记下，后面会用到，千万要记住！！！

接下来就是读取TimeStamp（3个字节）和Extent Timestamp（1个字节），两个字段加起来，占用了4个字节，全部为0，具体参考上(孙悟空 )面的spec可以看到为什么，过了四个字节以后，进入到StreamID部分，StreamID占用了3个字节，总是0，Spec中已经可以看到，接下来就是0200 0a6f ....部分，这部分在前面的spec中可以看到说明，由于前面我们看到的是TagType是18，那么这个地方就做ScriptData相关的处理办法，直接按照ScriptData来处理：

从上面spec中可以看到，ScriptData里面使用了ScriptTagBody，而ScriptTagBody的定义为一个Type和ScriptDataValue，其中的DataValue是根据Type来决定的，我们读数据读到offset为0x0000010 + 8位置，这时下一个字节为02,也就是说，Type为SCRIPTDATASTRING，这个时候会处理SCRIPTDATASTRING,

SCRIPTDATASTRING份两部分，一个是2两个字节的StringLength，一个是StringData，这个StringData部分占用的字节多少由StringLength部分决定，那么就继续上面的位置读取2字节，为00 0a,然后接下来读取00 0a个字节，作为String内容，也就是看到的onMetaData这个字符串，这样，onMetaData这个字符串读取 (孙悟空 ) 完毕，也就是前面提到的ScriptTagBody的Name部分读取完毕。接下来内容进入到了高潮：读取ECMA ARRAY部分，同样使用前面的办法读取对应的内容，如图：

其中08为ECMAARRAY的类型，接下来的0000 000d为ECMAARRAY中的结构的ECMAArrayLength，一次读取4个字节，得到的结果为13(0x0d)。

也就是说，ECMAARRAY由13个对应的项，可以建立一个递归或者循环，去读取这13个Variables，接下来开始解析第一个Variable，也就是SCRIPTDATAOBJECTPROPERTY。

接下来就是解析PropertyName，也是SCRIPTDATASTRING类型，对应解析办法前面有提到过，一次读两个字节，然后获得PropertyName的字符串，然后获得对应的PropertyData数据，要注意的是PropertyData里面的数据类型，读到的是00的话，就是Number方法，这个Number存的值就是个DOUBLE,通过前 (孙悟空 ) 面的spec参考可以看到对应的信息，下图是对应的字节：

接下来可以按照前面的方法继续循环或者递归，这样，就将onMetaData信息全部读取出来了，读取最后的时候不要忘记有三个字节，0000 09,为SCRIPTDATAOBJECTEND，

这样，FLV的onMetaData部分解析完毕