On Mon, Mar 22, 2010 at 3:43 PM, Luke Kanies <[email protected]> wrote:
> On Mar 22, 2010, at 2:03 PM, Markus Roberts wrote:
>
>>> I don't think there's code ready, but #2604 (Ruby 1.8.1's known bad
>>> syck causes core dumps) would be great to see fixed in 025.5 rather
>>> than wait a month or so for 0.25.6, even if it's just a half-decent
>>> work-around.
>>>
>>> I believe this makes using puppet on RHEL/CentOS 4 rather difficult.
>>> Fortunately, there don't seem to be that many folks doing so, as there
>>> were very few complaints when we had 0.25.x in epel-testing for a good
>>> month or so.
>>
>> It isn't just a matter of waiting a month or so.  Short of
>> internalizing yaml production the way we did with JSON, there's
>> nothing we can do about it.  The 6+ year old yaml library in
>> RHEL/CentOS 4 has existing, known problems and all the workarounds for
>> the string length problem just led to other problems.  Options are:
>>
>> * Upgrade
>> * Use PSON
>> * Convince us (meaning Luke) that we should internalize yaml
>> * Avoid things like templates that produce long strings in the catalog
>> as much as possible and hope for the best.
>
> I'm fine pulling a yaml version into Puppet; we've already done so for json,
> and it's totally obvious that the Ruby version of it will never stop
> sucking.
>
> Not sure we can comfortably do that in a 0.x.y release cycle, though, right?

Well, we did PSON under the same conditions and as I recall the yaml
replacement was a significantly easier patch (I think I've even got it
stashed away someplace).  I'll see if I can dig it up.

-- Markus

-- 
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