-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

> None of that is about bugs being ignored, but it might help explain
> why some of those bugs are so sticky...

I'm all in for a new way how facts are provided, like a more nested
datastructure (eg: hash). And I think at the point this is introduced
it might also be a good point to switch the meaning of their value.
Because if people want to use the new structure, they would have to
adapt their manifests anyway.

However, we are more or less still in the 1.6.x-series and the virtual
value used to work in 1.6.1 and is since 1.6.2 broken. 1.6 is now at
1.6.13 and the bugs reporting these issues are ~ 12 Months old.
And virtual is really one of the facts where people's eye begin to
shine with joy, when they see the possibilities they gain by using it
within their manifests.

And the issue I referred to with ipaddress is that 2 people logged
into the same machine as root might get totally different facts based
on their LOCALE. Yes, LOCALEs are hard to deal with and at this point
I'm actually not a 100% sure, but afair this haven't been the case in
some of the early 1.6 releases, at least we had never the problem
until we went to newer 1.6 releases. So I see that also as a regression.
Also if it isn't a regression: I still don't think that a core tool,
such as facter, should behave differently based on such a hidden and
not very well-known ENV-var such as LOCALE. You really have to know
facter a bit in detail, to get an idea what might get wrong if you get
a different output for 2 (nearly) identical sessions and combine it
with the knowledge that your co-worker rather likes to read german
than english (for whatever reasons). This is not that obvious and
hence people will see facter rather as being unpredictable and unstable.

In the past few months I bootstrapped a couple of new puppet
environments and each time I had to tell them, that they can't use the
official facter release and should take my patched RPM, because things
are broken, have been reported and will be fixed (at some point...).
Which have also been the case yesterday, which annoyed me and made me
complain.

So while I totally support your vision for a better datastructure and
more meaningful values within facter, I don't really see this as a
reason to not fix things that used to work and got broken and
especially things that are crucial for the value of facter as a tool:
Providing reliable and consistent values about a system.

~pete
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.11 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://www.enigmail.net/

iEYEARECAAYFAlB/HJcACgkQbwltcAfKi3+DjgCfY2Q3VaAnXcXwHOqTn9JtqFI/
OF0AnjHD3Tj71sJyQHheKKcbyTwM/eC0
=BZ7t
-----END PGP SIGNATURE-----

-- 
You received this message because you are subscribed to the Google Groups 
"Puppet Developers" group.
To post to this group, send email to [email protected].
To unsubscribe from this group, send email to 
[email protected].
For more options, visit this group at 
http://groups.google.com/group/puppet-dev?hl=en.

Reply via email to