Code review comment for lp:~doanac/utah/rsyslog-v4

Revision history for this message
Andy Doan (doanac) wrote :

On 03/11/2013 01:26 PM, Max Brustkern wrote:
> 1. I'm still seeing ubiquity (such as precise desktop amd64) installs where the last line is like:
> Mar 11 14:44:37 utah-1657-precise-amd64 rsyslogd: [origin software="rsyslogd" swVersion="5.8.6" x-pid="1652" x-info="http://www.rsyslog.com"] exiting on signal 15.
> I think that should be one of the options.

I left that one out because one we always hit one of the other current
regex's before this ever shows. However, its still a good safe-guard to
keep us from waiting forever.

> 2. I think the configuration should be in the package. Since all installs will now be assuming it exists, I think it should be in config.py. The reason everything is in config.py is so that if a user has an empty utah config file, we can still theoretically have some defaults and work. I think that design decision is something we want to evaluate, per some email discussions, but for now, I think we should continue doing things in that manner until we actually move away from it.

I agree. I wasn't sure if we wanted to start putting new fields like
this in conf.d instead of config.py so that things don't break when we
update the package. However, I guess apt will warn us, right?

> Finally, I'm seeing some weird things on mini installs like this:

That's interesting. I wonder if this happens on trunk?

« Back to merge proposal