> Greg came up with a pretty good solution for DIY (IMO). Basically, by
> passing CC="gcc -B/usr/bin/" during binutils-pass1 and gcc-pass1, gcc
> will continue to use the host linker until we get our gcc built.
Maybe this hack work for FC6 (i686), but it isn't actually sufficient
on bi-arch (multi
And two remarks on the LFS-6.2.
1. Section 6.9.1. Installation of Glibc
[QUOTE]
When running make install, a script called test-installation.pl
performs a small sanity test on our newly installed Glibc. However,
because our toolchain still points to the /t
Fix wrote:
> And two remarks on the LFS-6.2.
>
> 1. Section 6.9.1. Installation of Glibc
>
> [QUOTE]
> When running make install, a script called test-installation.pl
> performs a small sanity test on our newly installed Glibc. However,
> because our toolchain sti
Chris Staub wrote:
>> Running "make install" in section 6.9.1 I have discovered that perl
>> binary installed in /tools/bin tries to locate its
>> extensions in /usr/local/* directories, however they were installed
>> in /tools/share/perl5/*. Before entering the chroot
>> environment
On 3/19/07, Chris Staub <[EMAIL PROTECTED]> wrote:
> No, the "test" it is referring to is done at the end of "make install".
> It even specifies this at the beginning of the text you just
> copied-and-pasted from the book. If you *are* getting a large number of
> testsuite errors, something else i
Fix wrote:
> On 3/19/07, Chris Staub <[EMAIL PROTECTED]> wrote:
>
>> No, the "test" it is referring to is done at the end of "make install".
>> It even specifies this at the beginning of the text you just
>> copied-and-pasted from the book. If you *are* getting a large number of
>> testsuite error
On 3/19/07, Chris Staub <[EMAIL PROTECTED]> wrote:
> >
> LFS does not work for 64-bit systems.
Yes, I know. But on a i386 system, these tests should be linked
against /tools/lib/ld-linux.so.2 or against just compiled new linker
that resides somewhere in glibc-build directory?
Fix
--
http://linux
Fix wrote:
> On 3/19/07, Chris Staub <[EMAIL PROTECTED]> wrote:
>> LFS does not work for 64-bit systems.
>
> Yes, I know. But on a i386 system, these tests should be linked
> against /tools/lib/ld-linux.so.2 or against just compiled new linker
> that resides somewhere in glibc-build directory?
>
On 3/19/07, Chris Staub <[EMAIL PROTECTED]> wrote:
> Well, as I mentioned before, it should never have been created anyway.
> The actual problem needs to be fixed, not simply worked around with a
> symlink.
For sure. I've said "a temporary" workaround.
Fix
--
http://linuxfromscratch.org/mailman
On 3/19/07, Chris Staub <[EMAIL PROTECTED]> wrote:
> The command in the book should change it so that the test uses the
> newly-installed libraries in /lib.
About what a test are you saying, that which is execute after (in the
process of) "make install", or about bunch of the small test programs
Fix wrote:
>
> About what a test are you saying, that which is execute after (in the
> process of) "make install", or about bunch of the small test programs
> that are executed by "make check"? I mean second, and at the time they
> are executing NO libraries are installed in /lib. If, naturally, y
11 matches
Mail list logo
