Coda File System

did I hit that compiler bug?

From: Matthias Drochner <M.Drochner_at_fz-juelich.de>
Date: Tue, 26 Aug 2003 09:54:49 -0400
Hi -
Now I've decided to give 6.0.2 a try.
It just crashed (signal 11):
(gdb) where
#0  0x482eb527 in _sys___sigsuspend14 () from /usr/lib/libc.so.12
#1  0x80b093d in SigChoke ()
#2  0x483237c4 in __sigtramp_sigcontext_1 () from /usr/lib/libc.so.12
#3  0x805abe0 in fso_vol_iterator::operator() ()
#4  0x80a443c in repvol::ValidateFSOs ()
#5  0x80a3097 in repvol::GetVolAttr ()
#6  0x808dfef in volent::Enter ()
#7  0x80a609c in vproc::Begin_VFS ()
[...]

Now I'm seeing in the 6.0.1 announcement:
    Fixes a miscompilation of a code fragment with some gcc compilers.
    This lead to various crashes mostly in and around various list
    iterators. This affected pretty much anyone who built Coda with
    an not so bleeding edge version of the gnu C++ compiler (2.95.3,
    3.0, 3.1, 3.2.1?)

My compiler is 2.95.3.
It _might_ be possible that the NetBSD pkgsrc build environment
thwarts such a fix by passing "-O2" through the environment.
So before I'm looking into the wrong direction, just a question:
Is that the bug I hit, and how was it fixed?

best regards
Matthias
Received on 2003-08-26 10:03:14