Parsing a buffered binary data

**KianTern** · 05-23-2008

I'm trying to parse a binary file.
The file consists of multiple header->record pairs.
The record length and type are stored in the header.
I want to be able to read the entire header into a buffer and then split it into the variables.
Unfortunately I can't figure out how to do it.

Code:

#include <stdio.h>
#include <stdlib.h>
#define HEADER_SIZE 4

typedef struct rHead
{
    unsigned short int REC_LEN;
    unsigned char REC_TYP;
    unsigned char REC_SUB;
} recHead;


recHead readHeader(char *buffer)
{
    recHead hdr;
    /* no idea how to parse it */
    return hdr;
}

int main()
{
    FILE *fp;
    fp=fopen("1.std","rb");
    char *buffer;
    buffer = (char *)malloc(HEADER_SIZE+1);
    fread(buffer,HEADER_SIZE,1,fp);
    recHead header = readHeader(buffer);
    free(buffer);
    return 0;
}

**Elysia** · 05-23-2008

Why not read directly into the struct? You can do that, too. Or is there a reason you don't want to do that?

**KianTern** · 05-23-2008

Reading directly into the structure is OK for the header but it will cause problems for the record it self, as it contains some optional fields, so i will need to parse the buffer anyway.
The second reason is, I started using C yesterday (I'm a Perl programmer) and I want to learn.

**Elysia** · 05-23-2008

Then I think the easiest way is to use sscanf or pointers (the latter is the one I'm more familiar with).
A way might be like this:

Code:

typedef struct
{
    char c;
    short s;
    int i;
    long long l;
} mystruct;

void foo(const char* pBuffer)
{
    mystruct mys;
    mys.c = *(char*)pBuffer;
    pBuffer += sizeof(mys.c);
    mys.s = *(short*)pBuffer;
    pBuffer += sizeof(mys.s);
    mys.i = *(int*)pBuffer;
    pBuffer += sizeof(mys.i);
    mys.l = *(long long*)pBuffer;
    pBuffer += sizeof(mys.l);
    /* In case you don't want to assign something, just don't run the assign line */
}

Example (not tested).
Just watch out for endian problems. You should write the data in the same order as you read it and you can't port it to another endian machine without writing a new file first.

**KianTern** · 05-23-2008

Thanks it works great.

But I can't understand how this line works.
mys.c = *(char*)pBuffer;
Can you please explain or point me where can I read about this?

I'm currently not worried about the endian, it's just a project to learn
general algorithms and coding style in C.

**Elysia** · 05-23-2008

Although this

Code:

mys.c = *(char*)pBuffer;

...could just be simplified as this...

Code:

mys.c = *pBuffer;

...the secret behind these lines is that first casts the char pointer into a pointer to another type, and then dereferences that pointer. Memory itself contains no type information. Data is just treated by the compiler depending on what the type of variable it is. So we're just going to tell the compiler to treat the data differently.
To walk forward in the buffer, we simply add the size of the variable. Thus we skip forward X number of bytes, past the current data you just assigned. Then we repeat the process.

**KianTern** · 05-23-2008

Thank you very much for your help.

Thread: Parsing a buffered binary data

Thread Tools

Search Thread

Display

Parsing a buffered binary data

Similar Threads

xor linked list

Binary data handling

Binary data -> istream

Help with Reading and writing binary data

ReadFile and binary data