I'm certainly not that familiar enough with perl to generate that
script. However, I think the real issue is with Configure. It should
not force the -march=i486 as part of the fixed script, but rather have
it as on option or take the information from CFLAGS or CXXFLAGS as
appropriate.
Carter
Ca
On Thu, Jun 11, 2009 at 08:49:54AM -0700, Paul Rogers wrote:
> On Thu, 11 Jun 2009 09:00:21 -0400, "Carter Browne"
> said:
> > I'm not a perl monger either. I just changed the Linux-elf entry to
> > -march=x or -mtune=x for the more recent versions of gcc. That worked
> > without any problems f
On Thu, 11 Jun 2009 09:00:21 -0400, "Carter Browne"
said:
> I'm not a perl monger either. I just changed the Linux-elf entry to
> -march=x or -mtune=x for the more recent versions of gcc. That worked
> without any problems for me.
That was my first thought. If you saw my response to Kyle fly p
I'm not a perl monger either. I just changed the Linux-elf entry to
-march=x or -mtune=x for the more recent versions of gcc. That worked
without any problems for me.
Carter
Carter Browne
CBCS
cbro...@cbcs-usa.com
781-721-2890
Paul Rogers wrote:
> On Wed, 10 Jun 2009 13:17:48 -0400, "Carter
./Configure --prefix=/usr --openssldir=/etc/ssl shared os/compiler:"gcc
-march=i386" -lnsl -lsocket\
2>&1 | tee log.conf && exit $PIPESTATUS) &&
I have found the best way to pass non-std gcc flags is to set the
os/compiler to "gcc -flags". In this case -lnsl and -lsocket are not set
by defau
On Wed, 10 Jun 2009 13:17:48 -0400, "Carter Browne"
said:
> When I looked at the Configure script in the past, the only way I say to
> get an entry for x86 Linux was to modify the script itself. It does not
> look at CFLAGS or CSSFLAGS to get the CPU options.
The big problem seems to be that Co
When I looked at the Configure script in the past, the only way I say to
get an entry for x86 Linux was to modify the script itself. It does not
look at CFLAGS or CSSFLAGS to get the CPU options.
Carter
Carter Browne
CBCS
cbro...@cbcs-usa.com
781-721-2890
Paul Rogers wrote:
>> Did you run '.
> Did you run './config 386'?
OK, so I changed my script to try THAT:
# linux-elf is close, but we need -march=i386 instead of -m486
#cp Configure{,.backup} &&
#sed '/^"linux-elf"/s/486/arch=i386/' Configure.backup > Configure &&
#rm -f Configure.backup &&
#(Configure 386 linux-elf --prefix=/usr -
On Tue, 9 Jun 2009 22:54:41 -0700, "Kyle Hamilton"
said:
> Did you run './config 386'?
Mostly I was following a LFS page:
# linux-elf is close, but we need -march=i386 instead of -m486
cp Configure{,.backup} &&
sed '/^"linux-elf"/s/486/386/' Configure.backup > Configure &&
rm -f Configure.backup
Did you run './config 386'?
-Kyle H
On Tue, Jun 9, 2009 at 10:09 PM, Paul Rogers wrote:
> I found the "make report" and ran that on a pristine copy of my
> tarball with this result:
> OpenSSL self-test report:
>
> OpenSSL version: 0.9.7g
> Last change: Fixes for newer kerberos headers. NB:
I found the "make report" and ran that on a pristine copy of my
tarball with this result:
OpenSSL self-test report:
OpenSSL version: 0.9.7g
Last change: Fixes for newer kerberos headers. NB: the casts
are nee...
Options: no-krb5
OS (uname): Linux pandora 2.4.31 #2 Wed Feb 14
I have a NEED to build up a LIGHTWEIGHT version of an OLD Linux
system. It won't be in a perimeter, high threat situation, no
real NEED for OpenSSL, but it would be good to have. Here's the
outline, kernel-2.2.26, bash-2.04, perl-5.6.1, openssl-0.9.7g,
-march=i386. I've successfully installed 0.
12 matches
Mail list logo