Re: egcs and 2.0.XX (was: Re: 5000 emails later, compressed summary ;))

Greg Zornetzer (gaz@andrew.cmu.edu)
Wed, 3 Jun 1998 14:58:14 -0400 (EDT)


On Wed, 3 Jun 1998, Jimmie Farmer wrote:
<SNIPPED>
>
> I hate to be contrary, but I have had no problems compiling 2.0.33
> here using egcs-1.0.3:
Hi Jimmie,

That's right. You can compile 2.0.33 just fine with egcs or gcc2.8.
But ioport.c has code that depends on a bug in gcc2.7 for it to work.
Many people who compiled kernels in this way found that X-Windows started
killing the machine. The solution was to backport the ioport.c from the
2.1 series, which has been fixed. For some reason, the developers have
not moved these changes back to 2.0. This was probably done to require
users of the 2.0 stable series use a final stable version of the gcc
compiler. Yes, it does make numerous compilation errors, but it has been
the standard c compiler for linux machines for a while now. Gcc8 and Egcs
have had less time to prove themselves, and probably should not be
included in production-level machines.

>
> Linux version 2.0.33 (root@malchick) (gcc version egcs-2.90.29 980515
> (egcs-1.0.3 release)) #24 Thu May 28 01:54:31 PDT 1998
>
> I get a LOT of warnings generated during the compile, but it
> works. =-)

I believe there's a patch out for 2.1 that fixes many of these warnings.

>
> I haven't tried the pre2.0.34's yet, though... I was just
> basically waiting on it to be released before trying it out. If something
> like the above suggestion was added, it would greatly irritate me that I
> would have to comment that out to get a kernel to compile here. =-\
<SNIPPED>

It will really annoy the kernel hackers to see another bogus report of 2.0
locking under X because the wrong compiler was used to compile the kernel.

Of course, everything said here, except the warning of crashes under X, is
just my opinion.

Greg Zornetzer - gaz+@andrew.cmu.edu
"Light shines brightest in the darkest night"
http://www.contrib.andrew.cmu.edu/~gaz

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu