Just a thought that popped into my head

Gerard Beekmans gerard at linuxfromscratch.org
Mon May 6 09:45:04 PDT 2002

On Mon, May 06, 2002 at 05:33:45PM +0100, Ian Molton wrote:
> LFS 4 will use gcc3.x?

Yes if all goes according to plan, gcc-3.1 will make it to CVS.
> that seems like a Really Bad Idea (tm) if what I am reading on the GCC mailinglists is true.
> 3.0.x has many regressions from 2.95.3
> 3.1.x has even more
> To me the GCC project looks to be in a state of extreme flux and not production grade at all.
> Im sure it'll be fine eventually, but right now I think gcc 3.x is a mistake.

I don't agree. I have to-date not seen one single problem with using
gcc-3.0.4 on my systems. The only thing that doesn't work is getting a java
plugin working with Mozilla due to those ABI problems. The work around is
simply compiling Mozilla (and Galeon in my case too) with gcc-2.95.3 and
all works perfectly.

I agree a hint needs to be written regardless how to install gcc-2.95.3 on
your system next to gcc-3 and that hint will be mentioned when installing
GCC in chapter 6 so you can take care of both at the same time if you want

I think a lot of the reported problems are only potential problems which
may explain why I can't find anything wrong here (but I use my systems in
different ways so I certainly am not saying that gcc-3 is good for
everybody). It sounds a lot like what the kernel guys were saying.
GCC-2.95.3 was not recommended to compile a kernel with for a very long
time, but literally tens of thousands of LFS'ers were doing it anyways and
I think I can count the reported problems with that compiler and old
kernels on one hand.

Perhaps I'm merely ignorant about the gcc-3 changes which will make it Very
Bad (tm) to put into LFS but unless I'm told otherwise with some evidence
to back it up (not just theory, I will want/need to see real-life examples
of failures) I may consider not adding gcc-3 to the book. But, based on
personal experiences it's not so bad as people may make it sound.

Gerard Beekmans

-*- 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