On Mon, Jul 27, 2009 at 11:13 AM, Luke Kanies<[email protected]> wrote:
>
> On Jul 27, 2009, at 8:05 AM, Nigel Kersten wrote:
>
>>
>> On Sun, Jul 26, 2009 at 9:05 PM, Luke Kanies<[email protected]> wrote:
>>>
>>> On Jul 26, 2009, at 8:41 AM, Nigel Kersten wrote:
>>>
>>>>
>>>> On Sun, Jul 26, 2009 at 8:33 AM, Brice
>>>> Figureau<[email protected]> wrote:
>>>>>
>>>>> On 23/07/09 18:02, Nigel Kersten wrote:
>>>>>> Signed-off-by: Nigel Kersten <[email protected]>
>>>>>> ---
>>>>>>  .../provider/nameservice/directoryservice.rb       |   36 ++++++
>>>>>> ++
>>>>>> +-----------
>>>>>>  1 files changed, 16 insertions(+), 20 deletions(-)
>>>>>>
>>>>>> diff --git a/lib/puppet/provider/nameservice/directoryservice.rb
>>>>>> b/
>>>>>> lib/puppet/provider/nameservice/directoryservice.rb
>>>>>> index 9daed17..f4c9d59 100644
>>>>>> --- a/lib/puppet/provider/nameservice/directoryservice.rb
>>>>>> +++ b/lib/puppet/provider/nameservice/directoryservice.rb
>>>>>> @@ -108,18 +108,14 @@ class DirectoryService <
>>>>>> Puppet::Provider::NameService
>>>>>>              return @macosx_version_major
>>>>>>          end
>>>>>>          begin
>>>>>> -            product_version =
>>>>>> Facter.value(:macosx_productversion)
>>>>>> -            if product_version.nil?
>>>>>> -                raise Puppet::Error, "Could not determine OS X
>>>>>> version from Facter"
>>>>>> -            end
>>>>>> -            product_version_major = product_version.scan(/(\d+)\.
>>>>>> (\d+)./).join(".")
>>>>>> +            product_version_major =
>>>>>> Facter.value(:macosx_productversion_major)
>>>>>
>>>>> My Facter (version 1.5) doesn't have this fact (but has
>>>>> macosx_productversion).
>>>>>
>>>>> Could it be possible to have a degraded mode, as right now I get
>>>>> tons of
>>>>> failing (old) tests?
>>>>>
>>>>> I know I should upgrade (and will), but supporting (not so) older
>>>>> versions of Facter should be great, or at least produce a warning
>>>>> saying
>>>>> that I should upgrade.
>>>>
>>>> We need to pick a minimum Facter version for 0.25.0 compatibility.
>>>>
>>>> The reason I made this patch was because we had a bug in the Facter
>>>> code that essentially did the same thing, and fixing it in two
>>>> separate spots seems silly...
>>>>
>>>> We've been doing this in install.rb, and I feel like that is the
>>>> correct place to check the Facter version.  We should patch that
>>>> when
>>>> we've decided what the minimum Facter version for 0.25.0 is.
>>>>
>>>> It's the job of the packagers for various distros to ensure that
>>>> external dependencies are correct imho, and it seems like a waste of
>>>> effort to go through working out degraded modes for all the
>>>> different
>>>> Facter features to me....
>>>
>>
>>> I think that, at the least, it's reasonable to produce an upgrade
>>> notice or something.  Just having a failure because someone hasn't
>>> upgraded isn't very good.  Plenty of people use hand-rolled packages,
>>> or git, or tarballs.
>>>
>>> If it's reasonable to support the old version but produce a
>>> deprecation notice, that would be best.
>>
>> So I have a couple of options for the actual provider, and am happy to
>> take direction on any of them:
>>
>> a) continue to duplicate the Facter code for major version inside the
>> provider without a Facter version warning.
>> b) continue to duplicate the Facter code for major version inside the
>> provider with a Facter version warning.
>> c) Fail with a deprecation warning if the Facter version is below
>> whatever version first provided this fact.
>>
>> I still think we should be specifying a minimum Facter version for
>> Puppet 0.25.x somewhere, whether it be in the install.rb script, or in
>> the Puppet code, or both.
>>
>> Whichever approach people are happy with, I'll make the tests do the
>> same thing I guess.
>

> I would prefer option b here.  I agree on specifying a minimum
> version; you willing to also provide a patch to install.rb that checks
> the version?

ok. Re-doing the patch with option b) and will submit a separate patch
to install.rb that checks.



>
> --
> A motion to adjourn is always in order. --Robert Heinlein
> ---------------------------------------------------------------------
> Luke Kanies | http://reductivelabs.com | http://madstop.com
>
>
> >
>



-- 
Nigel Kersten
[email protected]
System Administrator
Google, Inc.

--~--~---------~--~----~------------~-------~--~----~
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