[PATCH 0/3] Event API back end for syslog

Matt Wagner matt.wagner at redhat.com
Thu Dec 1 00:01:56 UTC 2011


On Wed, Nov 30, 2011 at 03:38:05PM -0500, Jason Guiditta wrote:
> This patchset implements the feature in $subject, Redmine feature:
> https://www.aeolusproject.org/redmine/issues/2509
> 
> I am resending patch 1 due to no feedback on previous send.  There is 
> a minor change to calling process on an event, in that if you wish to
> specify a converter, you pass in an object, rather than a magic string.
> However, this causes no change on the default call (syslog), it merely
> opens up an easy way to later pass in custom-configured converters. For
> instance, we may want to allow the app/user to specify a different facility
> or ident which is different from the default we are using now.
> 
> Once again, tests can be run with:
> rspec spec/aeolus/event/*.rb --format doc
> 
> [PATCH conductor 1/3] RM #2822: Add way to retrieve full list of
> [PATCH conductor 2/3] Redmine #2514: Write and test transformation
> [PATCH conductor 3/3] Redmine #2515: Write and test actual logging

Conditional ACK.

I found one (critical) bug in this, and it also exposed a few bugs in my
own code. I'm sending patches that get everything working.

I commented out the entirety of my EventSpec test because I don't
understand what's going wrong. It's blowing up in event.rb with the
just-created Cidr or Cddr record being nil. I _think_ it's because
should_receive in the spec tests implicitly stubs out the method, so nil
is returned. I've poked and prodded to no avail, but it _looks_ like a
problem with the test and not the code, so I commented the test out for
now. (I just added it the other day, before we had end-to-end working,
so it's not like it's a time-tested test.)

So my ACK is conditional on either my patches being pushed (I'm not
going to push them myself at this hour), or someone doing something
better that gets tests passing. If you're okay with my patches, go ahead
and push the whole set.

-- Matt



More information about the aeolus-devel mailing list