From: Hans-Bernhard Broeker <broeker AT physik DOT rwth-aachen DOT de>
Message-Id: <9607310811.AA25549@axpmgr.physik.rwth-aachen.de>
Subject: Re: gcc -g -o
To: drupp AT cs DOT washington DOT edu (Douglas Rupp)
Date: Wed, 31 Jul 1996 10:11:17 +0200 (MET DST)
Cc: djgpp-workers AT delorie DOT com
In-Reply-To: <199607310628.XAA09018@june.cs.washington.edu> from "Douglas Rupp" at Jul 30, 96 11:28:25 pm
Content-Type: text

> > 
> > On Unix, gcc without -g does NOT strip the binary.  If ours will, that 
> > would be a deviation from the normal behavior, which might bite you if 
> > you use Makefiles that come from Unix.
> That's right, but what is normal behavior?  Different flavors of Unix treat
> the -g flag differently when it's passed to ld. Some ignore it, some don't.
> I'm trying to argue that it's reasonable for us to do something with it.

Do you have any idea how hard it might be to convince RMS to include
such changes into the distribution? In case you didn't know, FSF
standards (as read by RMS) tell you that all binaries should be
installed with 'at least minimal debug information'. That's why no FSF
makefile installs programs in a stripped way (no 'strip' call, no
'install -s'). If you compiled them with '-g', they will install a
version with full debug information. Breaking this rule might get us
into trouble with the FSF, like them refusing to incorporate future
changes from the DJGPP team.

And, given the fact that the '-s' flag works as it should, where's the
exact need for such a change? If you want it stripped on installation,
have Eli make sure that 'install -s' works, and there you are.

The one thing all this reminds me of is: DJGPP binutils (still) don't
understand stubbed .exe's... 

Hans-Bernhard Broeker (Aachen, Germany)