On Oct 27, 9:45 pm, Nigel Kersten <[email protected]> wrote:
> On Wed, Oct 27, 2010 at 5:57 PM, donavan <[email protected]> wrote:
> > Just to follow up on 5112 I have a dirty patch that appears to work.
> > Nominally tested it on 10.4, 10.5, & 10.6. 10.4 now applies catalogs
> > instead of failing. All versions successfully manage a test services
> > state as well.
>
> > Does anyone have a better suggestion than '-o /dev/stdout'? Seems a
> > mite hacky to me.
>
> I actually think you've come up with a reasonably clean solution given
> the hackiness of the underlying tools :)

I'll try to get some actual testing in tomorrow. At first glance it
does fix #5112 without a lot of effort. And I suppose it's reasonable
to expect /dev/stdout to be around.

That said it still leaves two issues:

Your objection to launchd provider start, stop, enable, & disable
always writing as xml regardless of original format. I assume this is
part of #4841? The binary -> xml behavior isnt as import as getting
#5112 fixed, to me.

As I mentioned in #5112 it looks like Puppet::Error.new was raised,
but the reports did not contain the error. Because of that all my
catalogs were passing, though applying 0 resources. New issue or am I
missing something obvious?

> > Also I think that the 10.4 machines are going to a have a \ ( slash )
> > file in whatever puppets working dir was. plutil seems to have been
> > interpreting as literal file name.
>
> Can you elaborate? I'm not understanding the problem here.

Markus got it right for me. I meant dash not slash. Looks like /-
( filename dash in root directory ) was created on my affected
machines. Dunno if that's worth an errata against 2.6.2 or not.

-- 
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