Several folks say I should ignore my friend, and that gcc 2.96 is just fine.
Wrong.
About a week ago, *I* was compiling a new kernel, 2.4.19 (seems to be the
real stable release). I was getting SEGVs, alternating with undefined
functions or variables. I finally succeded only by doing the make bzImage
about six times, sequentially. Each time, it had no trouble getting through
the problem it had before, but croaked later.
My take on it was that on the SEGVs, I'm guessing there was a stack overflow,
and on the undefineds,
1) the linker in Linux is a single-pass linker;
2) unlike, for example, Solaris' linker, it does *not* allow one library to
be specified multiple times, and unless you use a rather obscure
switch, and list libraries with that switch, rather than the usual,
it will *not* look in them again in one session.
For example, if it links function A from library A, then doesn't find
function B in that lib, it goes on to check lib B; if it then comes to lib
C, which happens to be defined in lib A, it will not find it.
And yes, I have identified this happening in the past.
This is my take on the undefined.
Further, in the README, that came with the sources, it says:
> COMPILING the kernel:
> - Make sure you have gcc 2.95.3 available. gcc 2.91.66 (egcs-1.1.2) may
> also work but is not as safe, and *gcc 2.7.2.3 is no longer supported*.
Ok? No FUD, personal knowledge, and documented in the source.
mark
________
Mark Roth
Unix/Linux systems administrator & software developer
HUSO Support, a division of Prometheus Unbound: affordable Linux
technical support and training for the home user and small business.
For more information, please contact us at:
email <[EMAIL PROTECTED]>
phone +1 773 274 2584
--
redhat-list mailing list
unsubscribe mailto:redhat-list-request@;redhat.com?subject=unsubscribe
https://listman.redhat.com/mailman/listinfo/redhat-list
