Re: Problem with libc6-pic
- To: Ben Collins <email@example.com>
- Cc: firstname.lastname@example.org
- Subject: Re: Problem with libc6-pic
- From: Rene Mayrhofer <email@example.com>
- Date: Tue, 31 Oct 2000 12:07:13 +0100
- Message-id: <39FEA7E1.C0806134@vianova.at>
- References: <39FD415B.B82324E8@vianova.at> <20001030074250.L10561@visi.net> <39FDAB75.D791F355@vianova.at> <20001030122129.N10561@visi.net> <39FDB0A7.CB0D9200@vianova.at> <20001030124550.O10561@visi.net> <39FDBCB9.9367739A@vianova.at> <20001030134610.P10561@visi.net> <39FDC429.1AE20A3A@vianova.at> <20001030141721.Q10561@visi.net>
Ben Collins wrote:
> On Mon, Oct 30, 2000 at 07:55:37PM +0100, Rene Mayrhofer wrote:
> > Ben Collins wrote:
> > > objdump --dynamic-syms bin/busybox | grep lxstat
> > >
> > > Send me the output from that please.
> > 080496cc DF *UND* 00000137 GLIBC_2.0 __lxstat
> Strange, busybox is showing it doesn't use lxstat64 (which is what I
> expected, but it is still strange :)
> > Does this mean that it only needs glibc 2.0 ? There are only 2 symbols tagged
> > with glibc 2.1:
> > 0804979c DF *UND* 0000016a GLIBC_2.1 fclose
> > 0804993c DF *UND* 0000009f GLIBC_2.1 fopen
> Ok, this is from libc.so.6:
> 00000000000c800c g DF .text 0000000000000100 (GLIBC_2.1) __lxstat64
> 00000000000c7c34 g DF .text 00000000000000fc GLIBC_2.0 __lxstat
> 00000000000c800c g DF .text 0000000000000100 GLIBC_2.2 __lxstat64
> So the most recent lxstat64 is the GLIBC_2.2 one, which is the one that
> should get used if anything links with it. Now the question is, where does
> this come from? __lxstat64@GLIBC_2.1 is a weak symbol (only there for
> backward compatibility).
> IMO, this might just be a bug in mkshlibs.sh
Does anybody besides Markus Brinkmann know the script well enough to debug it ?
I certainly do not. Does it work on your system ? Can you build a minimized
libc6 for busybox ?
Am I the only one experiencing this problem at the moment ?