robert at linuxfromscratch.org
Fri Jul 23 21:24:58 PDT 2004
On July 23, 2004 11:44 pm, J. Scott Edwards wrote:
> On Wed, 21 Jul 2004, Robert Connolly wrote:
> > On July 21, 2004 11:16 am, Miguel Bazdresch wrote:
> > ...
> >>> sedwards at marconi tmp $ more
> >>> /mnt/lfs/sources/glibc-build/math/test-ifloat.out
> > I have a lot more test failures using gcc-3.3.4 compared to gcc-3.3.3
> > with both gcc and glibc's testsuites. I'm pretty sure I have had the
> > book's packages build without test failures but I need to retry it
> > because its been a while.
> > --
> Well, I've tried it now with glibc-2.3.2 and it won't even compile glibc
> in Ch. 6. In fact it won't compile anything, it complains that it can't
> find Scrt.o. I probably botched something trying to kludge in 2.3.2 in
> place of glibc-2.3.3-lfs-5.1.
> I am thinking now that I should go back to the original (non-hardened)
> book and see if I can get that to build ok. If it builds then I would
> know that my system is at least working I guess.
The pie patches depend on a new glibc-2.3.3 feature.
http://pax.grsecurity.net/ has a package with Scrt.o source (I think). Its
more complicated but it can work if you really want to use glibc-2.3.2.
More information about the hlfs-dev