Re: GCC-2.9.5 & Linux ( and Co. ) ...

Erik Mouw (J.A.K.Mouw@ITS.TUDelft.NL)
Fri, 22 Oct 1999 21:46:01 +0200


On Fri, 22 Oct 1999 20:14:26 +0200, Davide Libenzi wrote:
> my job has kept me very busy in the last 4 months, so I've loose a lot of
> topics ( mostly all ) of
> linux-kernel.
> Yesterday I've installed GCC 2.9.5-1 and I've tried to rebuild some packages
> of which, in my absence,
> has been released new versions.

You probably mean gcc 2.95.1 :-). It might be nice to know that gcc 2.95.2
is expected to be released this weekend (but don't be surprised if it
doesn't).

> As a result a lot of them failed to compile due to a more strict error
> checking and other issues and
> some of them got crashed after the rebuild step !
> Sad of this behaviour I've jumped to build the new kernels ( 2.2.13 &
> 2.3.22 ) that compile fine but got crashed !
> Now my questions are :
>
> 1) What's happened to GCC ?

The egcs team has taken over gcc development. See http://gcc.gnu.org/ .

> 2) Which way has they choosen to follow ?

They want to make a better compiler. I think they also want world
domination, just like Linux wants ;-).

> 3) If a new behaviour has been introduced, why not keep it as a choice ( at
> least for a bit of transition time ), leaving
> the maintainers the time to adjust code ?

But that was already possible: gcc 2.95.x is just egcs 1.2 with another
name badge. If your program compiled with egcs 1.1.x, it will probably
compile with gcc 2.95.1. Personally I had only minor problems when I moved
from gcc 2.7.2.3 (and gcc 2.8.1) to gcc 2.95.1. The C++ compiler is
different because the C++ standard itself changed.

> 4) When ( If ?!??! ) GCC will rejoin Linux kernel code ?

I don't know. AFAIK, the biggest problem is the strict aliasing check that
gives problems with Linux kernels, but using -fno-strict-aliasing solves
that. FYI: gcc 2.95.2 will default to compile without the strict alias
check, according to Jeffrey Law:

<quote>
From: Jeffrey A Law <law@cygnus.com>
To: gcc@gcc.gnu.org
Date: Thu, 21 Oct 1999 01:20:51 -0600
Subject: gcc-2.95.2 vs strict aliasing

This is an FYI, not an RFC -- I do not want to start another debate on the
strict aliasing issue _as it applies to the upcoming GCC 2.95.2 release_.

--

I have changed the branch to not perform strict aliasing optimizations by
default. This is a tactical retreat to give us more time to continue sorting
through the potential for warnings and avoiding optimization in cases where
we can detect that a program is likely violating the aliasing rules.

We will reevaluate the strict aliasing decision for future releases
independently of the decision to disable them for the GCC 2.95.2 release.
</quote>

Erik

-- 
J.A.K. (Erik) Mouw, Information and Communication Theory Group, Department
of Electrical Engineering, Faculty of Information Technology and Systems,
Delft University of Technology, PO BOX 5031,  2600 GA Delft, The Netherlands
Phone: +31-15-2785859  Fax: +31-15-2781843  Email J.A.K.Mouw@its.tudelft.nl
WWW: http://www-ict.its.tudelft.nl/~erik/

- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.rutgers.edu Please read the FAQ at http://www.tux.org/lkml/