[ILUG] Can anyone see any reason why a customer would need to compile a new compiler?

[Declan Moriarty]

On Mon, 2006-06-26 at 14:30 +0100, paul at clubi.ie wrote:
> On Mon, 26 Jun 2006, paul at clubi.ie wrote:
> 
> > On Mon, 26 Jun 2006, Declan Moriarty wrote:
> >
> >> On Sun, 2006-06-25 at 23:55 +0100, paul at clubi.ie wrote:
> >
> >>>  Binutils maybe. Kernel headers??? You thinking of glibc maybe?
> 
> Oh, if the point was too subtle: With respect to kernel headers, it's 
> the "/Linux/ kernel headers" I find hard to believe as being a gcc 
> build prerequisite. ;)

I didn't find it subtle, but adverserial.

The things needed to build gcc are direct requirements; the things
needed to put these requirements there are indirect requirements.
You're throwing nice shapes, Paul but convincing nobody with
generalities and probabilities and "In my vast experience" which,
incidentally I don't doubt.

The reason you're not convincing me is that the guys actually doing it
are coming up with different results. You should try installing gcc-4.1
sources on a vanilla rhel and see how far you get. I'll allow you the
odd sed on Makefile & configure script, but no patches or editor hacks,
and definitely no make -k. Oh, set SELinux on paranoid for the job. Most
of that compiling is on hardened systems, adding to the sweat.

If it's as straightforward as you seem to think it will be, that's when
I'll reach for the humble pie. If all these other things don't matter,
it should 'in all probability ' fall straight in :-)