From owner-freebsd-current Wed May 27 09:24:01 1998 Return-Path: Received: (from majordom@localhost) by hub.freebsd.org (8.8.8/8.8.8) id JAA02377 for freebsd-current-outgoing; Wed, 27 May 1998 09:24:01 -0700 (PDT) (envelope-from owner-freebsd-current@FreeBSD.ORG) Received: from ns1.yes.no (ns1.yes.no [195.119.24.10]) by hub.freebsd.org (8.8.8/8.8.8) with ESMTP id JAA02348 for ; Wed, 27 May 1998 09:23:57 -0700 (PDT) (envelope-from eivind@bitbox.follo.net) Received: from bitbox.follo.net (bitbox.follo.net [195.204.143.218]) by ns1.yes.no (8.8.7/8.8.7) with ESMTP id QAA01013; Wed, 27 May 1998 16:23:43 GMT Received: (from eivind@localhost) by bitbox.follo.net (8.8.8/8.8.6) id SAA03746; Wed, 27 May 1998 18:23:37 +0200 (MET DST) Message-ID: <19980527182337.47243@follo.net> Date: Wed, 27 May 1998 18:23:37 +0200 From: Eivind Eklund To: Robert Nordier Cc: current@FreeBSD.ORG Subject: Re: Fix for undefined "__error" and discussion of shared object versioning References: <19980527155446.50625@follo.net> <199805271551.RAA11565@ceia.nordier.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii X-Mailer: Mutt 0.89.1i In-Reply-To: <199805271551.RAA11565@ceia.nordier.com>; from Robert Nordier on Wed, May 27, 1998 at 05:51:11PM +0200 Sender: owner-freebsd-current@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.ORG On Wed, May 27, 1998 at 05:51:11PM +0200, Robert Nordier wrote: > Eivind Eklund wrote: > > > On Wed, May 27, 1998 at 08:15:40AM +0200, Ollivier Robert wrote: > > > According to Eivind Eklund: > > > > I think making it support e.g. 'long long' would need modifications of > > > > the frontend. > > > > > > lcc, now at 4.0, still has the same problem I think. > > > > That's basically because "long long" is in total violation of the ISO > > standard, IIRC. > > A type long long int is part of the ISO standard presently under > consideration (C9X). I know - I reviewed this as part of the review team for Norway :-) > It is also not a "total violation" of the current ISO (C89/C94) > standard in any meaningful sense. It invokes undefined behaviour, and requires a diagnostic, and I believe it require a (minor) re-definition of some other parts of the grammar, and it break the semantics for 'long'. It is about as total a violation as you can get. The only reason it was allowed it into C9X was that it already was used too many places to be reasonably deniable :-( > TenDRA can and already does support type long long int. How do you set it up to allow this? I tried to find some way of doing it, but didn't. > > I've sent off a mail to the contact person for the TenDRA team; I'll > > wait and see that they say about the feasibility of using TenDRA as an > > OS compiler. > > This was the subject of an Open Group research study (involving UnixWare > and Linux) around 1995. (Fairly obviously, it is feasible.) I was thinking more of compiler speed and how much support is planned in the future. I know of the above studies (though I've not read them yet). > The critical issue, however, is really whether it is feasible to > compile *FreeBSD* with a non-GCC compiler. Whether it can be made feasible :-) It isn't doable at the moment, at least. > Any real problems will be due to FreeBSD (GCC quirks in the code) > not to the compiler. (There are no fundamental difficulties > inherent in compiling UNIX-like systems). It think the difficult issues here are linker sets and assembly. There are a few places where the code use GCC extensions to the preprocessor, but removing this has been on my TODO list for some time. Eivind. To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-current" in the body of the message