unmerge 576945
reassign 576945 g++
tag 576945 - pending
thank you
Cloning failed because I had submitted two reports by mistake and
merged them; fixing.
l
--
To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.de
clone 576944 -1
reassign -1 g++
tag 576944 + pending
thanks
On Sat, Apr 10, 2010 at 05:22:45PM +0200, Luca Niccoli wrote:
> I've recently submitted bug #576944 and could use some help to trace
> it down better, and understand if it's really a bug.
> Some more things I've n
Hi,
I've recently submitted bug #576944 and could use some help to trace
it down better, and understand if it's really a bug.
Some more things I've noticed:
If I set /proc/cpu/alignment to 0 (ignore) or 3 (fixup+warn) I see
three alignment traps each time _xapian.so is loaded (
> I've actually noticed that rebuilding python-xapian is enough to get
> rid of the odd dynamic relocation entries in the SO and of the
> alignment traps with them.
I was wrong.
What fixed the problem was passing CXXFLAGS="-g -O2 -Wcast-align" to debuild.
More precisely, by default debian/rules bu
I've actually noticed that rebuilding python-xapian is enough to get
rid of the odd dynamic relocation entries in the SO and of the
alignment traps with them.
Maybe a bug on the buildd toolchain? (I rebuilt the package with an
up-to-date sid box).
Cheers,
Luca
--
To UNSUBSCRIBE, email to debia
Package: python-xapian
Version: 1.0.18-1
Severity: important
After setting the kernel to fix and report alignement faults, I found this
warnings in my dmesg:
[1032087.934414] Alignment trap: update-apt-xapi (16364) PC=0x4000b1c0
Instr=0xe59a3000 Address=0x4099c336 FSR 0x001
[1032087.947502] Alig
6 matches
Mail list logo
