From: pavenis AT lanet DOT lv
Message-ID: <B0000101030@stargate.astr.lu.lv>
To: djgpp-workers AT delorie DOT com
Date: Thu, 9 Sep 1999 20:45:57 +0300
MIME-Version: 1.0
Content-type: text/plain; charset=US-ASCII
Content-transfer-encoding: 7BIT
Subject: Re: gcc-2.95
References: <199909091439 DOT QAA20898 AT propus DOT lu DOT erisoft DOT se>
In-reply-to: <Pine.LNX.3.93.990909182429.8972J-100000@acp3bf>
X-mailer: Pegasus Mail for Win32 (v3.12a)
Reply-To: djgpp-workers AT delorie DOT com

On 9 Sep 99, at 18:30, Hans-Bernhard Broeker wrote:

> On Thu, 9 Sep 1999, Martin Stromberg wrote:
> 
> > Andris said:
> > > ------------------  assembler code generated by gcc-2.95
> > >         leal -16(%ecx),%ebx
> > >         movl $-1,%eax
> > >         leal 8(%ebx),%edx
> > > /APP
> > >         .byte 0xf0, 0x0f, 0xc1, 0x02
> > > /NO_APP
> > >         cmpl $1,%eax
> > >         jne L712
> 
> The marks /APP and /NO_APP usually enclose assembly coming from inline
> assembly found in the C++ source. I.e. this particular line may well exist
> as a verbatim 'asm' statement, in the compiled source. 
> 

No need to demangle names: Grepping contents of $DJDIR/lang/cxx 
and subdirectories gives the source of this instruction:
	lang/cxx/tsd/bastring.h           lines 84-87

So even if compiler itself will not generate instructions which don't run 
on i386 or i486, some other package can use macros __i486__, 
__i586__ to permit use on some additional instructions so defaulting
to -march=pentium could be dangerous even if compiler itself don't 
generate instructions which are absent on i386.

Andris