Debian Bug report logs - #26436
[done with gcc-2.96] please recompile with -fhonor-std

Toggle useless messages

Message #5 received at submit@bugs.debian.org (full text, mbox, reply):

From: <wichert@wiggy.ml.org>

To: submit@bugs.debian.org

Subject: libstdc++2.9: please recompile with -fhonor-std

Date: Sat, 5 Sep 1998 20:38:01 +0200

Package: libstdc++2.9
Version: 2.91.56-1

Looking at /usr/doc/g++/NEWS.gz namespace support is now finally in egcs,
but the libstdc++ has been compiled without -fhonor-std, so it still does
not work.

Please recompile libstdc++, so I can finally port some stuff written
in msvc to Linux.

Wichert.

-- System Information
Debian Release: 2.0
Kernel Version: Linux pc135a 2.0.33 #2 Fri May 22 20:39:06 CEST 1998 i686 unknown

Versions of the packages libstdc++2.9 depends on:

Message #10 received at 26436@bugs.debian.org (full text, mbox, reply):

From: Matthias Klose <doko@cs.tu-berlin.de>

To: <wichert@wiggy.ml.org>, 26436@bugs.debian.org

Subject: Bug#26436: libstdc++2.9: please recompile with -fhonor-std

Date: Sun, 6 Sep 1998 12:50:01 +0200 (MET DST)

wichert@wiggy.ml.org writes:
> Package: libstdc++2.9
> Version: 2.91.56-1
> 
> Looking at /usr/doc/g++/NEWS.gz namespace support is now finally in egcs,
> but the libstdc++ has been compiled without -fhonor-std, so it still does
> not work.
> 
> Please recompile libstdc++, so I can finally port some stuff written
> in msvc to Linux.

Would this libstdc++ still be binary compatible with the one compiled
without -fhonor-std? Sorry for my ignorance but I'm not a C++
developer. Would the compilation with -fhonor-std be conformant with
the LCS?

Message #15 received at 26436@bugs.debian.org (full text, mbox, reply):

From: Wichert Akkerman <wichert@wiggy.ml.org>

To: Matthias Klose <doko@cs.tu-berlin.de>, 26436@bugs.debian.org

Subject: Re: Bug#26436: libstdc++2.9: please recompile with -fhonor-std

Date: Sun, 6 Sep 1998 22:38:22 +0200

[Message part 1 (text/plain, inline)]

Previously Matthias Klose wrote:
> Would this libstdc++ still be binary compatible with the one compiled
> without -fhonor-std? Sorry for my ignorance but I'm not a C++
> developer. Would the compilation with -fhonor-std be conformant with
> the LCS?

No, since it changes the mangled names of all symbols in the standard
C++ library. But the C++ standard dictates all symbols in the library
must be in the std namespace, so we should use -fhonor-std anyway.
If we don't switch now we'll have to make the switch later, which might
be more painful.

Wichert.

[Message part 2 (application/pgp-signature, inline)]

Message #20 received at 26436@bugs.debian.org (full text, mbox, reply):

From: jdassen@wi.leidenuniv.nl

To: wichert@wiggy.ml.org, 26436@bugs.debian.org

Cc: galenh-egcs@debian.org

Subject: Re: Bug#26436: libstdc++2.9: please recompile with -fhonor-std

Date: Tue, 8 Sep 1998 08:38:56 +0200

On Sun, Sep 06, 1998 at 10:38:22PM +0200, Wichert Akkerman wrote:
> Previously Matthias Klose wrote:
> > Would this libstdc++ still be binary compatible with the one compiled
> > without -fhonor-std? Sorry for my ignorance but I'm not a C++ developer.
> > Would the compilation with -fhonor-std be conformant with the LCS?
> 
> No, since it changes the mangled names of all symbols in the standard C++
> library. But the C++ standard dictates all symbols in the library must be
> in the std namespace, so we should use -fhonor-std anyway. If we don't
> switch now we'll have to make the switch later, which might be more
> painful.

I've considered your request, but decided against -fhonor-std. I realise
that this means our standard C++ library isn't conforming to the C++ standard, 
but the GNU one hasn't been conformant so far anyway. I know it is moving
towards ISO C++ conformance, and that's a good thing. 
Breaking inter-Linux binary compatibility in order for us to be somewhat
more conformant somewhat earlier is a bad thing. It'll only give us a
reputation for being "the distribution whose C++ libraries are broken".

We'll switch to honouring the STD namespace when egcs does so by default
(and then there'll be a so number change to indicate the incompatibility
with previous versions).

Ray
-- 
Cyberspace, a final frontier. These are the voyages of my messages, 
on a lightspeed mission to explore strange new systems and to boldly go
where no data has gone before. 

Message #29 received at 26436@bugs.debian.org (full text, mbox, reply):

From: Matthias Klose <doko@cs.tu-berlin.de>

To: 26436@bugs.debian.org, wichert@wiggy.ml.org

Date: Tue, 17 Nov 1998 11:24:14 +0100 (MET)

Please see http://www.cygnus.com/ml/egcs/1998-Nov/0704.html and
followups (http://www.cygnus.com/ml/egcs/1998-Nov/0720.html) for an
explanation.

Is this a wishlist item?

Message #44 received at 26436-done@bugs.debian.org (full text, mbox, reply):

From: Wichert Akkerman <wichert@wiggy.net>

To: Adam Lazur <zal@debian.org>

Cc: 26436-done@bugs.debian.org, 26028-done@bugs.debian.org

Subject: Re: reopenning bug 26028 with working email address

Date: Mon, 17 Dec 2001 22:01:00 +0100

Previously Adam Lazur wrote:
> Wichert, I've taken the liberty of reopenning bug #26028 under a working
> email address. (bugs 24783 and 26436 also seem to still use your old
> address... I didn't migrate those though)

26436 is fixed with g++-3, 26028 seems to be fixed. I haven't run into
27783 anymore, but then again I haven't been reading much usenet lately

Wichert.

-- 
  _________________________________________________________________
 /wichert@wiggy.net         This space intentionally left occupied \
| wichert@deephackmode.org            http://www.liacs.nl/~wichert/ |
| 1024D/2FA3BC2D 576E 100B 518D 2F16 36B0  2805 3CB8 9250 2FA3 BC2D |

Send a report that this bug log contains spam.