Thread: Examine hex number?

That depends what kind of datatype you (need to) use. Do you want to break these up into ints or something? You could just put the whole file into one big char buffer and parse it with read(), which has no problem with zero bytes -- although you could do this when you read the file in:

Code:

int fd=open("file", O_RDONLY), i++;
char byte, buffer[4096];
while ((read(fd,&byte,1))   {    /* true until EOF */
     [filter or look for bytes?]
     buffer[i]=byte; i++;
}

just a sketch

You can use that switch with (byte) in the above read() loop as part of a "parsing pipeline":

Code:

int fd=open("file", O_RDONLY), i++;
char byte, buffer[4096], tmp[256];
while ((read(fd,&byte,1))   {    /* true until EOF */
     switch (byte) {
           case (0xcd):  read(fd,&byte,1);
                    sprintf(tmp,"INT %d",(int)byte);
                    strcat(buffer,tmp);  break;
           case (0xf4) :      etc....
    
}

Yes, fread() is more portable than read, sorry.

You do it the same way but you use a FILE* stream -- fopen() -- instead of a file descriptor from open(). The fread() would look like this:

Code:

while ((fread(&byte,1,1,fstr))) {

I just noticed I have some unclosed parantheses in my previous work here, but anyway...

Originally Posted by Sly

My only problem is, how would I get it to recognize certian bytes as prefixes rather than ???. And how would I grab only a byte at a time.

Let's look at this more closely!

Code:

#include <stdio.h>

int main() {
        int halt=0;
        FILE *file=fopen("file", "r");
        unsigned char byte;
        while ((fread(&byte,1,1,file)))   {    /* true until EOF */
                switch (byte) {
                        case (0xcd):            /* notice: 1,1 means one byte at a time */
                                fread(&byte,1,1,file);  
                                printf("INT %d\n",(int)byte);
                                break;
                        case (0xf4): 
                                halt=1;
                                break;
                        default: break;       /* could use this for something too */
                }   
                if (halt==1) break;
                /* still more opportunities to work with "byte" */
        }   
        return 0;                           
}

I just noticed I get a gcc warning unless you use an unsigned char here, because 0xf4 would be an "illegal" value for a char.

Anyway, you get it to recognize a certain byte as a prefix using the switch statement -- but if you are worried that there will be a 0xcd that is not really a 0xcd, well...

why to use fread for reading 1 byte?
what is the problem with fgetc?

Originally Posted by vart

why to use fread for reading 1 byte?
what is the problem with fgetc?

I would say it doesn't matter much, but then fgetc returns an int.

Also, as the OP indicates, sh/e may in fact need to read more than one byte.

If you do that, you can use a small buffer and iterate thru each byte in a loop:

Code:

char buffer[256];
fread(buffer,1,256,file);
for (i=0; i<256; i++) {
        switch (buffer[i]) {
     ...etc.

Just keep in mind that "buffer" will not be null terminated.

OK. Thanks. But where would I put the x--?

Code:

        x = stat(argc[1],&fbuf);
        x = fbuf.st_size;

stat() is just as portable (or unportable) as open()/read()/etc. If you want to figure out the size of a file, open it, fseek() to the end, ftell() to get the position, and fseek()/rewind() back to the beginning of the file.

I don't know what you're asking with the "x--" question. If stat() returns the number of bytes in the file, loop in the range 0 to x-1.

But I don't see why you need to know the size of the file, anyway. MK27's code handles files of any size. Perhaps you need the count for something else, but you haven't said so. In fact, you haven't said what you're doing at all (though I'd guess some sort of disassembler). What are you doing?

I have a disassembler for x86, it's about 36KB in size.

It is table-driven, and although it say "disasm.cpp", it's not really C++ code.

It doesn't do the MMX/3DNow!/SSE instructions, or any other new instructions introduced in recent processors (branch prediction prefix for example).

But you are welcome to use it as a template, or simply play around with it.

Note that it is NOT an attempt on "most pretty code" - it does use a few different dirty tricks that may not be portable, it has at least one goto (that I spotted when doing a fast scan through the code), and I'm sure there are other things that aren't great either.

Due to the extension rules, the file is a renamed zip-file.

Edit: Perhaps I should have stated that I'm happy for anyone to use the source code, but you must mention that I wrote the original code - I think that's fair.

--
Mats