Pending GCC-3.1 upgrade - do or don't

Gerard Beekmans gerard at linuxfromscratch.org
Wed May 8 08:03:25 PDT 2002


On Wed, May 08, 2002 at 10:13:08AM +0100, Ian Molton wrote:
> I KNOW in the past we have held off upgrades (A recent version of yacc
> and auto-something we held off, because it gave problems with several
> packages).

A base LFS system couldn't be compiled properly so we didn't do that one
upgrade. More importantly, we knew a new version was going to be released
very soon afterwards so I didnt' feel like upgrading (it was bison) to only
upgrade it again two weeks later.
 
> The compiler is the //foundation// of the system. it should have the
> LONGEST waiting period of all before acceptance.

Released is released. It doesnt' matter how long you let gcc-2.95.3 wait,
it won't get any better or worse. The code is released and static. So the
only reason to wait very long is if people are very slow to build an LFS
system with some additions. I can duplicate my workstation build within 24
hours so I really quickly put gcc-3 through all kinds of motions, tests and
more tests.

> 2.7.2.3 does NOT compile everything fine. I've personally written code
> it f*cked up.

It will compile most things fine. Just like 2.95.3 will compile most things
fine, but 3.1 will compile other things better.
 
> well, what is your idea of an acceptable delay time for 3.1 inclusion
> following its release ?

However long it will take me to re-install my entire workstation and
servers and test every app a bunch of times. I can do it in a week if I
take the time to do it (building software itself in a day, testing the
software to remainder of a week).
 
> anything. Its a HUGE and COMPLEX piece of code.   

Yep I totally agree. But theoretical failures don't count with me I'm
afraid.

-- 
Gerard Beekmans
www.linuxfromscratch.org

-*- If Linux doesn't have the solution, you have the wrong problem -*-
-- 
Unsubscribe: send email to listar at linuxfromscratch.org
and put 'unsubscribe lfs-dev' in the subject header of the message



More information about the lfs-dev mailing list