From:	pavenis AT lanet DOT lv
To:	"Eli Zaretskii" <eliz AT is DOT elta DOT co DOT il>, djgpp-workers AT delorie DOT com
Date:	Sat, 28 Apr 2001 16:58:48 +0300
MIME-Version:	1.0
Subject:	Re: gcc-3.0 related patches for DJGPP CVS version
Message-ID:	<3AEAF6C8.19289.14FAA0A@localhost>
In-reply-to:	<4331-Fri27Apr2001162351+0300-eliz@is.elta.co.il>
References:	<3AE96E9B DOT 10268 DOT B1244D AT localhost> (pavenis AT lanet DOT lv)
X-mailer:	Pegasus Mail for Win32 (v3.12c)
Reply-To:	djgpp-workers AT delorie DOT com
Errors-To:	nobody AT delorie DOT com
X-Mailing-List:	djgpp-workers AT delorie DOT com
X-Unsubscribes-To:	listserv AT delorie DOT com

On 27 Apr 2001, at 16:23, Eli Zaretskii wrote:

> > From: pavenis AT lanet DOT lv
> > Date: Fri, 27 Apr 2001 13:05:31 +0300
> > 
> > 5) for zoneinfo/src/zic.c changed '#ifdef unix' to
> > '#if def ined(unix) || defined(__DJGPP__)'. At least current CVS version 
> > of gcc-3.0 doesn't define unix for DJGPP (unless I'm modifying specs)
> > 
> > 6) removed #ifdef __GO32__ (it was also not defined)
> 
> These two bother me.  Why doesn't GCC 3.0 define them?  There are
> ports out there which depend on these definitions.

I see 2 possible places where to add them:
	1) gcc/config/i386/djgpp.h
	2) sys/version.h in djdevXXX
For end user both are identical.

> >   -MD
> >   -O2
> > ! -m486
> >   -malign-loops=2
> >   -malign-jumps=2
> >   -malign-functions=2
> > --- 1,7 ----
> >   -MD
> >   -O2
> > ! -march=i386
> > ! -mcpu=pentium
> >   -malign-loops=2
> >   -malign-jumps=2
> >   -malign-functions=2
> 
> Shouldn't we revise the -malign-* switches as well?  I think these
> switches generate suboptimal code.

I think the best would be to remove them all. I built current CVS 
version both with and without -malign* switches (using gcc-3.0 
20010426). Size difference (read bloat ...) is very small (<1%). If 
somebody wants to do some profiling I can make archives available

Andris

- Raw text -