Re: With a wacky macro before it, 'global' appears to miss a function

[Ezolt, Phillip]

Shigio,

On Apr 24, 2011, at 2:23 AM, "Shigio YAMAGUCHI" <address@hidden> wrote:

>> That seems a little dangerous to me.  I could easily see someone adding an e=
>> xtra line after a function definition by accident,  and being completely con=
>> fused when that function doesn't show up. (leading to another bug report..)
>> ...
>> 
>> Alternatively, would it be possible to use the semicolon at the end of the S=
>> TRUCT_DEF_MACRO to stop the parsing of that function? (so we move on to the 
>> n=
>> ext function..)
>> 
>> The rule should be something like: If we hit a semicolon,  and we don't have=
>> any open braces, then consider the function to be complete..
> 
> It is a good rule.
> However, may I throw away K&R style?

Ouch. Yeah... we don't want to do that. I think we need k&r support.

> 
> 1. ANSI style
> -------------
> 
> int
> main(int argc, char **argv) {
>    printf("Hello\n");
> }
> 
> 2. K&R(old) style
> -----------------
> 
> int
> main(argc, argv)
> int argc;
> char **argv;
> {
>    printf("Hello\n");
> }
> 
> If we need not support K&R, I would like to adopt your rule.
> What do you think?

Ok.. Ok. How about THIS rule.. (if this even possible?)

If we see a semicolon, AND then we see an open paren BEFORE we see an open 
brace, we know the token before the paren is a function and the previous 
function is complete?

That would catch this:
A() bob;
int foo() <==== this paren means the previous symbol is complete
{
}
...
But not k and r:
A(bar)
int bar;
{
}

> --
> Shigio YAMAGUCHI <address@hidden>
> PGP fingerprint: D1CB 0B89 B346 4AB6 5663  C4B6 3CA5 BBB3 57BE DDA3

Cheers,
--Phil