[openstack-dev] [Heat] Template repository location (infra guidance needed)

Monty Taylor mordred at inaugust.com
Wed Apr 24 15:53:31 UTC 2013



On 04/24/2013 08:36 AM, Steven Hardy wrote:
> On Wed, Apr 24, 2013 at 12:04:22PM +0200, Zane Bitter wrote:
>> On 23/04/13 22:53, Steven Hardy wrote:
>>> Hi,
>>>
>>> We've had some discussions recently within the team about creating a
>>> "heat-templates" repo somewhere, for the following reasons:
>>>
>>> 1. Separate the existing example templates[1] from the main code repo
>>
>> It's worth pointing out that, for better or worse - and, candidly,
>> probably mostly the latter - many of the Heat "unit" tests rely on
>> reading templates from the source tree. So we'll probably have to
>> simplify those and pull them inline in the test code. Just throwing
>> that out there.
> 
> That's a very good point, which I'd forgotten ;)
> 
> I think it's mostly a dependency on the basic wordpress template, so I
> guess we could just move a copy of that template under the tests directory
> - I'll look into the easiest fix.
> 
>>> 2. Provide a central place for us to put/review new heat-native/dsl/HOT
>>>    templates when we define/implement the syntax [2]
>>> 3. Somewhere to put the oz TDLs we currently bundle with heat-jeos[3] so
>>>    that we can deprecate that tool and replace it with a simple shell-script
>>>    or documentation.
>>>
>>> So our discussions so far have revolved around creating a stackforge
>>> "heat-templates" repo, which seems OK, but my question (for the
>>> infra/release team) is, would it make more sense to make another repo under
>>> the main openstack org, since these templates will be directly related to
>>> Heat (although possibly not part of the coordinated release)?
>>>
>>> I'm happy to go ahead and create a stackforge repo if another repo under
>>> the main openstack github org is deemed a bad idea, but wanted
>>> clarification/guidance before proceeding.
>>
>> +1 for Stackforge
> 
> Ok, sounds like we have agreement on Stackforge, I'll set up a repo later
> today.

Nope! I dissent.

It sounds like these are a part of heat - which to me means they should
be in openstack/ just like python-heatclient. ESPECIALLY if the heat
unit (or any testing of heat) uses them.



More information about the OpenStack-dev mailing list