Hi Adrien,
On 18.02.2014 18:50, Adrien Thebo wrote:> I should have responded sooner
to this thread, but now that the horse
> has clearly left the barn I might as well shut the gate.
>
> We've been working on getting Facter 2 ready for release for about two
> months now, and right now we're right on the cusp of shipping an RC. The
> goal for the initial release is to make the necessary changes to the
> internal API to permit structured data and minimize the amount of
> changes otherwise. The hope is that we can finally ship Facter 2, the
> initial release will be stable and will be as free of surprises
> possible, and then we can work on releasing new facts that can now be
> composed of structured data.
to be honest the role of the master and facter-2 branch are still not
clear to me. What I've did so far
1) The original pull request was based on master, so I've rebased on
facter-2
2) I've already merged the pull request in facter-2 yesterday (sorry for
that, I did not notice the "The reload to facter 2" thread and your
answer comes to late for me now)
3) I've already merged facter-2 into master
If I did understand you correctly, I'd rather have merged it directly
into master and hoped someone would backport it to facter-2, once facter
2.0 is out the door?
> The partition UUID fact is especially problematic because it's a
> fantastic candidate to use as part of a structured fact for partition
> information.
You are right that the new uuid facts represent structured data in a
flat, unstructured way, but that is also true for all the other
blockdevice facts. Are there actually any structured facts in the
facter-2 codebase? When I run `facter` on the `facter-2` branch, I do
not see a single array or hash, just strings. Having a few examples
might be helpful to actually write a proper facter2 fact (or should all
current facts stay-as-is and only new facts should make use of the new
API? Personally I'd be confused if some blockdevice facts where
unstructured while others in the same realm were structured)
> Stefan, how adamant are you about including these facts for Facter 2.0?
> Is there any way that we can postpone this till Facter 2.1?
Personally I do not have any use for that fact. I just noticed the pull
request did not receive any reaction for a long time and thought I might
help out. So if you plan on reverting the change to have a clean
facter-2 codebase, that's better answered by Chris Portman than myself
(merged pull request for reference:
https://github.com/puppetlabs/facter/pull/560)
-Stefan