RE: [avr-gcc-list] Comma Operator in #define

[Tom Harris]

I think that commas only seperates expressions, which are evaluated and the
result is the result of the last expression, so that 

a = ++b, c;

sets a to the value of c, after incrementing b. I have never seen the point
of using it, it seems to lead to hard to understand code.

The `asm` is a statement, not an expression, so it must be terminated by a
semicolon. 

Why not define an inline function for your code? It will have the same
effect as a #define, but be safer.

(in a header file):
extern inline void set_clock(uint8_t x) { 
        asm volatile("nop\n\t"::);
        PORTB |= BIT3;
}

TomH <tomh(AT)optiscan(DOT)com>

-----Original Message-----
From: Keith Gudger [mailto:address@hidden
Sent: Wednesday, 23 July 2003 7:58 AM
To: E. Weddington
Cc: address@hidden
Subject: Re: [avr-gcc-list] Comma Operator in #define


Eric:

PLease check your ANSI C manual - a 'comma' operator is a perfectly legal
C operator for separating 'C' statements.  It should work in a define if
the pre-processor is working correctly.

Keith

On Tue, 22 Jul 2003, E. Weddington wrote:

> > Maybe I'm missing something, but I *did* check the ANSI C 
> manual, and this
> > should be OK, I think.
> > 
> > I want to use a comma operator in a macro (#define) 
> statement:
> > 
> > #define set_clock(x) asm volatile("nop\n\t"::), (PORTB |= 
> BIT3)
> > 
> > avr-gcc gives the following error when I use this macro:
> > 
> >   set_clock();
> > 
> > error: parse error before ',' token
> > 
> > If I just make it a ';', then everything works OK.
> > 
> > Any ideas?  What am I doing wrong?  Or is this a bug?
> > 
> 
> AFAIK, the define body is composed of 2 seperate C 
> statements which should be seperated by semicolon. A comma 
> does not seperate C statements; a semicolon seperates C 
> statements. (Sometimes in English too. ;-))
> 
> It's not a matter of whether a comma is allowed in macro, 
> as so much whether the macro is defined to be legal C.
> 
> 
> 
> 
> 



_______________________________________________
avr-gcc-list mailing list
address@hidden
http://www.avr1.org/mailman/listinfo/avr-gcc-list