Re: [cc65] a question about the write function

On Sat, Feb 10, 2007 at 05:58:38PM -0500, Jac Ato wrote:
> In an attempt to implement the printf function on my embedded system,
> I wanted to reroute any queries that are sent through printf to my
> serial function: uart_send_char(char c). But I have failed to do
> exactly that. I've noticed that on the platform that I am working on,
> a modified NES with a serial port, I cannot get cc65 to generate a
> working binary file with the code below. It just wont get to the part
> of the write function. This was checked through my NES simulator. So
> what is the best way of doing the above idea?

The call hierarchy is:

    printf -> vfprintf -> _printf -> fwrite -> write

This means: printf uses fwrite for output, which in turn calls write. So I
would first try to call fwrite and check if your write function is called.

> int __fastcall__ write (int fd, const void* buf, unsigned count){
> 	char* temp=buf;
> 	char i=0;
> 	do{uart_send_char(temp[i++]);}
> 	while(count--);
> }
>
> I know that fastcall will get the right most parameter (is it count?)
> from a register instead of the stack, so that my implementation is
> probably wrong.

The implementation is wrong because write() must return the number of bytes
written, or EOF (-1) as an error code. Anyway, the function should be called
at least once. You may want to try something like this instead:

int __fastcall__ write (int fd, const void* buf, unsigned count)
{
    unsigned i = 0;
    while (i < count) {
        uart_send_char (((const char*) buf)[i]);
        ++i;
    }
    return count;
}

> Also, what does fd represents?

The file descriptor. For a quick solution, you can ignore it. Older runtime
library implementations for the Commodore machines did so in the times before
actual file I/O was implemented.

Regards


        Uz


-- 
Ullrich von Bassewitz                                  uz@musoftware.de
----------------------------------------------------------------------
To unsubscribe from the list send mail to majordomo@musoftware.de with
the string "unsubscribe cc65" in the body(!) of the mail.