[augeas-devel] Possible bug in lokkit lens

David Lutterkort lutter at redhat.com
Fri Oct 9 17:30:15 UTC 2009


On Fri, 2009-10-09 at 13:31 +0100, Richard W.M. Jones wrote:
> I don't know how serious this is, but libvirtd --verbose prints these
> messages every time I do a 'virsh dominfo' request:

I think it's coming from netcf, and
if /proc/sys/net/bridge/bridge-nf-call-iptables is 0, it doesn't matter;
if it's 1, you're in trouble. So it's good to report this, anyway.

I just pushed a patch that fixes the issue.

> warning: augeas initialization had errors
> please file a bug with the following lines in the bug report:
> /augeas/files/etc/sysconfig/system-config-firewall/error = "parse_failed"
> /augeas/files/etc/sysconfig/system-config-firewall/error/pos = "59"
> /augeas/files/etc/sysconfig/system-config-firewall/error/line = "4"
> /augeas/files/etc/sysconfig/system-config-firewall/error/char = "0"
> /augeas/files/etc/sysconfig/system-config-firewall/error/lens = "/usr/share/augeas/lenses/dist/lokkit.aug:76.10-.32"
> /augeas/files/etc/sysconfig/system-config-firewall/error/message = "Iterated lens matched less than it should"
...
> I don't know where the lokkit file is that Augeas is trying to parse.

Errors connected to file I/O are always reported
under /augeas/files/$FILE/error (where FILE is the absolute path of the
file)

The entries in the error message also tell you where Augeas got stuck:
line 4, char 0 (or 59 characters into the file, the 'pos' node)

If you wanted to investigate this further, you could write a little test
module /tmp/t.aug

        module T =
        
          let s = Sys.read_file "/etc/sysconfig/system-config-firewall"
        
          test Lokkit.lns get s = ?

and run that with augparse. Then try to pinpoint and fix the lokkit lens
to get it to parse that file. (Not that I'd expect that, the bug report
+ file where it fails was enough to go on)

David





More information about the augeas-devel mailing list