since
it pretty much kills the puppet run if the autoload fails. I think
supporting anything in the 3.x.x line is a reasonable thing to shoot for.
Eric
On Monday, August 18, 2014 9:24:17 AM UTC-5, Rob Reynolds wrote:
>
>
>
>
> On Sun, Aug 17, 2014 at 8:16 PM, badgerious >
I'd like to define an FFI feature in a module so it can die gracefully on
older puppets, but the question I have is this: Is it better to
A) create a feature at 'puppet/feature/ffi.rb' and do something like:
Puppet.features.add(:ffi, :libs => ['ffi'])
B) create a feature at 'puppet/feature/my_
I'm bringing this thread back to life because I've found myself with some
time to work on puppet things, and plan to make use of the
post_resource(s)_eval hook. I notice it still has the original name as of
3.6. So I've come to ask that the name not be changed. It's pretty annoying
to have API
Trevor,
Cool! Something like an augmented version of the flowchart you've done up
would be handy.
Eric
On Friday, March 7, 2014 9:36:39 AM UTC-6, Trevor Vaughan wrote:
>
> Eric,
>
> I attempted to do some of that for the 2.7 series
> https://www.onyxpoint.com/type-and-provider-execution-walkt
variable, you'll need to make sure that they're all unique across ALL
> types.
>
> Trevor
>
>
> On Thu, Mar 6, 2014 at 5:59 PM, badgerious
> > wrote:
>
>> Not sure if this will work as expected for your custom type, but a
>> thought might be to
There was a fair amount of bumbling to get this hook named in the first the
place, and clearly the name isn't satisfactory still. I think it's
difficult because different names make sense depending on the perspective
you look at the hook from (as Brice suggested). Maybe it would be useful to
dr
Not sure if this will work as expected for your custom type, but a thought
might be to set some (provider) class variable in #flush and then, in
::post_resource_eval, check if this variable is set and do the restart (so
each flush does no actual work other than setting the flag).
Eric
On Thurs
On Thursday, January 9, 2014 1:14:56 PM UTC-6, Trevor Vaughan wrote:
>
>
> I'm not certain if there is a portion of the provider code that gets run
> every time at the end without fail. There might be but I can't check the code
> right now.
>
>
In Puppet 3.4+, providers can define a ::post_resou
>
> If we support this functionality and there is a failure during the catalog
> run after a reboot at the end has been requested, what would be your
> expectation for the system.
>
> Would you expect it to:
>
> A) still reboot
>
B) not reboot
>
C) something else (please comment)
I vote A.
>
>
> We've set YARD up so that documentation is only generated for methods and
> classes with an '@api' tag. When we switched to using YARD we decided to
> take that approach because there were so many comments interpreted as
> documentation that were either blatantly wrong or not documentati
Re: the yard docs, I think they are generally pretty good for API docs, and
it looks like most of the stuff is in fact documented in the code, but not
appearing in generated output for some reason. I took a quick look but
didn't see any obvious reason for this. I'd be willing to donate some spar
(This is an expansion of the discussion on this PR:
https://github.com/puppetlabs/puppet/pull/1851)
The topic of interest is the naming of 'hook' methods (e.g. 'prefetch' on a
provider, 'validate' on a type).
>From the virtual lips of Adrien Thebo:
Regarding the name, I would like us to se
On Monday, July 15, 2013 11:53:10 AM UTC-5, Trevor Vaughan wrote:
>
> I just went to +1 this on #3946. I ran into a very similar situation
> recently and was quite annoyed to find that the Providers didn't have a
> 'finish' equivalent to the Types.
>
> Nan has a very good point that it needs to s
On Monday, July 15, 2013 10:05:16 AM UTC-5, Nan Liu wrote:
>
> I considered that option when I tackled this problem. The problem I ran
> into is when a puppet resource fails the last resource (due to dependency)
> does not get applied. So a failed puppet run did not clean up the
> connections. I
r option
> for this as well.
>
Thanks Nan. Here is the ticket I think you were referring to:
http://projects.puppetlabs.com/issues/3946 . Appears to be in languish mode
at the moment.
I started poking at another solution
(https://github.com/badgerious/puppet-finalizable/blob/master/lib
15 matches
Mail list logo
