Bug#216466: libc6.1 breaks software far and wide
On Sun, Oct 19, 2003 at 07:45:13PM -0400, Daniel Jacobowitz wrote:
> On Sun, Oct 19, 2003 at 10:46:03AM +1000, Duraid Madina wrote:
> > Package: libc6.1
> > Version: 2.3.2-7
> > Severity: critical
> > Tags: sid
> > Justification: breaks unrelated software
> >
> > The current libc6.1 in unstable, (2.3.2.ds1-5) breaks various pieces of
> > software, such as:
> >
> > XFree86 (4.3.99.14)
> >
> > BEA's JVM (BEA WebLogic JRockit(R) Virtual Machine (build
> > 8.1sp1-1.4.1-viking-Load8-linux64-stheng03-20030709-1550, Native Threads,
> > Generational Concurrent Garbage Collector)
> >
> > the Intel C/C++ compiler
> >
> > Some of this stuff is binary-only, some of it is built from source.
> > Whatever the case, it doesn't matter, this libc _breaks stuff_. Here's an
> > example stack trace of XFree86 dying:
> >
> > Program received signal SIGSEGV, Segmentation fault.
> > 0x200000000018e6e1 in __strtoll_internal () from /lib/libc.so.6.1
> > (gdb) where
> > #0 0x200000000018e6e1 in __strtoll_internal () from /lib/libc.so.6.1
> > #1 0x40000000000a7a30 in xf86strtol ()
>
> Backtrace doesn't tell me much - but I learned a lot from the bit down
> at the bottom of your message that said ia64. I haven't tested on
> ia64, we'll have to investigate.
Every library upgrade has the potential to break stuff. Especially
with glibc this tends to be applications which abuse the published
interface; we work around these where possible but in the end they're
bugs elsewhere and that's all there is to it.
You didn't say which version of the Intel compiler broke. I'm told
that this compiler has known issues with recent versions of glibc so
I'm inclined to suspect that the compiler is at fault. JVMs are famous
for abusing glibc interfaces, also.
X is worrisome. Can you reproduce this without using a beta CVS
version of XFree86? Or any less monolithic application? You haven't
given us much to work with.
--
Daniel Jacobowitz
MontaVista Software Debian GNU/Linux Developer
Reply to: