Follow @Openwall on Twitter for new release announcements and other news
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20150520174303.GZ17573@brightrain.aerifal.cx>
Date: Wed, 20 May 2015 13:43:03 -0400
From: Rich Felker <dalias@...c.org>
To: musl@...ts.openwall.com
Subject: Re: Re: musl xclock?

On Wed, May 20, 2015 at 07:00:17PM +0200, Szabolcs Nagy wrote:
> * Rich Felker <dalias@...c.org> [2015-05-20 12:05:02 -0400]:
> > On Wed, May 20, 2015 at 11:28:50AM -0400, John Mudd wrote:
> > > /usr/lib/gcc/i686-linux-gnu/4.8/. Clock.o xclock.o -lX11 -lXaw7 -lXft
> > 
> > ...and the only mildly suspicious thing is this last -L. Is it
> > possible that your distro has glibc libc.so in the same directory as
> > the gcc libs dir, /usr/lib/gcc/i686-linux-gnu/4.8/?
> > 
> 
> it seems to be an ubuntu gcc
> 
> libc.so should not be in that dir but libstdc++.so is there
> (which has libc.so.6 and ld-linux-x86-64.so.2 dependency among
> others)
> 
> if the problem is what i think it is then an empty ld-linux-x86.so.2
> may fix it

But how is /lib being searched at all? It's not in the -L path. 

> > > -lxkbfile -lXmu -lXrender -lXt /usr/lib/gcc/i686-linux-gnu/4.8/libgcc.a
> > > /usr/lib/gcc/i686-linux-gnu/4.8/libgcc_eh.a -lc
> > > /usr/lib/gcc/i686-linux-gnu/4.8/libgcc.a
> > > /usr/lib/gcc/i686-linux-gnu/4.8/libgcc_eh.a
> > > /usr/lib/gcc/i686-linux-gnu/4.8/crtend.o
> > > /home/mudd/musl/musl-1.1.8.install/lib/crtn.o
> > > /usr/bin/ld: h_errno: TLS definition in //lib/i386-linux-gnu/libc.so.6
> > > section .tbss mismatches non-TLS definition in
> > > /home/mudd/musl/musl-1.1.8.install/lib/libc.so section .bss
> > > //lib/i386-linux-gnu/libc.so.6: error adding symbols: Bad value
> > > collect2: error: ld returned 1 exit status
> > 
> > It's odd that //lib appears in the above output since that has nothing
> > to do with the -L pathnames above and it has a gratuitous double slash
> > at the beginning.
> > 
> > What distro are you using?
> > 
> 
> the first / comes from the default sysroot (gcc was built with
> --with-sysroot=/) and i think gcc falls back to that when it
> does not find something in the specified places..

Uhg, is there any way around that? If not, the musl-gcc wrapper is
useless with sysrooted toolchains. :(

Rich

Powered by blists - more mailing lists

Confused about mailing lists and their use? Read about mailing lists on Wikipedia and check out these guidelines on proper formatting of your messages.