Open Stack

On Thu, Dec 12, 2013 at 7:11 AM, Ryan Petrello
<ryan.petrello at dreamhost.com>wrote:

> Hello,
>
> I’ve spent the past week experimenting with using Pecan for Nova’s API,
> and have opened an experimental review:
>
> https://review.openstack.org/#/c/61303/6
>
> …which implements the `versions` v3 endpoint using pecan (and paves the
> way for other extensions to use pecan).  This is a *potential* approach
> I've considered for gradually moving the V3 API, but I’m open to other
> suggestions (and feedback on this approach).  I’ve also got a few open
> questions/general observations:
>
> 1.  It looks like the Nova v3 API is composed *entirely* of extensions
> (including “core” API calls), and that extensions and their routes are
> discoverable and extensible via installed software that registers itself
> via stevedore.  This seems to lead to an API that’s composed of installed
> software, which in my opinion, makes it fairly hard to map out the API (as
> opposed to how routes are manually defined in other WSGI frameworks).  I
> assume at this time, this design decision has already been solidified for
> v3?
>

Yes, from an implementation view everything is an "extension", even core
functionality. One issue with the V2 API is that because core is hard coded
and separate from the plugin framework there were things you could do in
core API code that you couldn't do in extensions and other things which you
could do in both, but had to do in different ways. Which is bad from a
maintainability/readability point of view. And inevitably we ended up with
extension specific code sitting in what should have been only core code. So
we ended up deciding to make everything a plugin to consistency of how API
code is written and also ensured that the framework didn't treat core API
code in any special way.


>
> 2.  The approach in my review would allow us to translate extensions to
> pecan piecemeal.  To me, this seems like a more desirable and manageable
> approach than moving everything to pecan at once, given the scale of Nova’s
> API.  Do others agree/disagree?  Until all v3 extensions are translated,
> this means the v3 API is composed of two separate WSGI apps.
>
>
Yes, I think this is the way to go. Attempting to get a big-bang patch
merged would be rather challenging.



> 3.  Can somebody explain the purpose of the wsgi.deserializer decorator?
>  It’s something I’ve not accounted for yet in my pecan implementation.  Is
> the goal to deserialize the request *body* from e.g., XML into a usable
> data structure?  Is there an equivalent for JSON handling?  How does this
> relate to the schema validation that’s being done in v3?
>
>
Yes the deserializer decorator specifies and XML deserializer which is used
when the default one is not suitable. schema validation is done on the
deserialized output so essentially covers both JSON and XML (eg XML is not
directly validated, but what we end up interpreting in the api code is).

Chris
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstack.org/pipermail/openstack-dev/attachments/20131212/ae7032f6/attachment.html>