From: "Ilya P. Ryzhenkov" <ilya AT spy DOT isp DOT nsc DOT ru>
Newsgroups: comp.os.msdos.djgpp
Subject: Re: Will djgpp optimize recursive function calls?
Date: Sun, 17 Nov 1996 12:43:47 -0800
Organization: Institute of Semiconductor Physics
Lines: 30
Distribution: world
Message-ID: <328F7903.5B8E@spy.isp.nsc.ru>
References: <Pine DOT A32 DOT 3 DOT 93 DOT 961115194737 DOT 37954A-100000 AT srv1 DOT freenet DOT calgary DOT ab DOT ca> <328E1F11 DOT 5F69 AT cs DOT com>
Reply-To: ilya AT spy DOT isp DOT nsc DOT ru
NNTP-Posting-Host: arab.isp.nsc.ru
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
To: djgpp AT delorie DOT com
DJ-Gateway: from newsgroup comp.os.msdos.djgpp

John M. Aldrich wrote:
> 
> I have read in some detail the gcc docs on optimization flags.  As I
> understand it, if gcc detects a recursive call, it will NOT inline the
> function that uses it.  It will still optimize the function itself, but
> will never convert it into inline code.
> 
> For more information, look in the gcc docs under Invoking GCC |
> Optimize.
Hmmmm.... It's not true perfectly ... Consider :
inline unsigned long fact(unsigned long i) /* I don't know if it will be
inlined without inline keyword, but i hope */
{ 
 return (i<2)?i:fract(i-1)*i;
}
void main()
{
 fract(0); fract(1); fract(2);
}
and compile this with gcc -O3 -o test.s test.c -S
and look into resulting test.s file
You'll see, that GCC DO optimizes (inline), but only for the FIRST call.
so you will actually get
 fract(x-1)*x for the call like fract(x) and a constant if you use
frac(0) or fract(1)

							Sincerely yours, Ilya
-----------------------------------------------------------------------------
mailto://ilya AT spy DOT isp DOT nsc DOT ru                           
http://spy.isp.nsc.ru