I had two of them in mind (not sure about the second):
1) if you assign more than four directives in a rule (e.g. five facilities),
metalog aborts with the following glibc error:
*** glibc detected *** /usr/sbin/metalog: realloc(): invalid next size:
0x0804e230 ***
======= Backtrace: =========
/lib/libc.so.6[0xa7e23b00]
/lib/libc.so.6[0xa7e268ed]
/lib/libc.so.6(__libc_realloc+0xff)[0xa7e27a7f]
/usr/sbin/metalog[0x804ac0b]
/lib/libc.so.6(__libc_start_main+0xd8)[0xa7dd57c8]
/usr/sbin/metalog[0x80493e1]
======= Memory map: ========
08048000-0804c000 r-xp 00000000 08:02 746326 /usr/sbin/metalog
0804c000-0804d000 rw-p 00004000 08:02 746326 /usr/sbin/metalog
0804d000-0806e000 rw-p 0804d000 00:00 0 [heap]
a7c00000-a7c21000 rw-p a7c00000 00:00 0
a7c21000-a7d00000 ---p a7c21000 00:00 0
a7dbf000-a7dc0000 rw-p a7dbf000 00:00 0
a7dc0000-a7ee4000 r-xp 00000000 08:02 1717013 /lib/libc-2.5.so
a7ee4000-a7ee5000 r--p 00124000 08:02 1717013 /lib/libc-2.5.so
a7ee5000-a7ee7000 rw-p 00125000 08:02 1717013 /lib/libc-2.5.so
a7ee7000-a7eea000 rw-p a7ee7000 00:00 0
a7eea000-a7f09000 r-xp 00000000 08:02 1718261 /lib/libpcre.so.0.0.1
a7f09000-a7f0a000 rw-p 0001e000 08:02 1718261 /lib/libpcre.so.0.0.1
a7f10000-a7f1a000 r-xp 00000000 08:02 720148 /usr/lib/libgcc_s.so.1
a7f1a000-a7f1b000 rw-p 00009000 08:02 720148 /usr/lib/libgcc_s.so.1
a7f1b000-a7f1d000 rw-p a7f1b000 00:00 0
a7f1d000-a7f1e000 r-xp a7f1d000 00:00 0 [vdso]
a7f1e000-a7f38000 r-xp 00000000 08:02 1716992 /lib/ld-2.5.so
a7f38000-a7f39000 r--p 00019000 08:02 1716992 /lib/ld-2.5.so
a7f39000-a7f3a000 rw-p 0001a000 08:02 1716992 /lib/ld-2.5.so
afba9000-afbbe000 rw-p afba9000 00:00 0 [stack]
I can reproduce this with different glibc versions (the output is from 2.5).
It happens
both with 0.7 and with 0.8-rc1 (this is from 0.8-rc1).
May be that this simply means that one should not use more than four
directives in a rule (it is always possible to split them in two rules gaining
the same result), but this should be declared in the docs, while the docs
simply say that you can use multiple facilities or programs or regex.
2) http://bugs.gentoo.org/show_bug.cgi?id=26694
I can reproduce this one, but I understand that it can be considered
independent of metalog.
Thanks for your attention, regards
Giorgio Lando
On 06:06 Sat 06 Jan , Mike Frysinger wrote:
> On Saturday 06 January 2007 05:57, Giorgio Lando wrote:
> > There are a couple of well-known, non serious bugs
>
> could you be a little more specific ?
> -mike
|
