On Thu, Feb 04, 2010 at 11:45:50PM +0100, Bernhard Reutner-Fischer wrote: >On Thu, Feb 04, 2010 at 09:10:16PM +0100, Laurent Bercot wrote: >>> I'm not convinced it is actually that important to have the weak >>> behaviour of _stdio_init, for a number of reasons: > >> Hello, >> >> I do have a complete system that does not use stdio. > >>For various reasons, I'm switching over to uClibc. >> I'm currently bulding an initramfs with static binaries for an embedded >>device. (I'm not using busybox for now, although that could change.) >> >> I would very much appreciate NOT pulling in stdio into said binaries. > >please try the workaround (it's not really our bug but in the toolchain) >from bugs.uclibc.org/1033 it should have the desired effect for all.
I assume you have DOPIC set, do you? If you turn it off then everything works as expected, without the 1033 patch, right? Sounds a bit like we are bending PIC rules a tiny bit too much, and as such it's our bug and not the toolchain's. Mike, do you remember why we build all archive members with PIC in the first place (not arguing about nonshared where it's obviously ok)? _______________________________________________ uClibc mailing list [email protected] http://lists.busybox.net/mailman/listinfo/uclibc
