[openstack-dev] Is WSME really suitable? (Was: [nova] Autogenerating the Nova v3 API specification)

Mac Innes, Kiall kiall at hp.com
Tue Aug 6 09:35:05 UTC 2013


So,

 From experimenting with, and looking at the WSME code - raising a 
status with `pecan.abort(404)` etc doesn't actually work.

WSME sees that, and helpfully swaps it out for a HTTP 500 ;)

The author of WSME even says there is currently no way to return a 404. 
So, ceilometer must be either not using anything but http 400 and http 
500, or have replaced WSMEs error handling :/

I'll have to have a look a ceilometers API to see if they ran into/fixed 
the issue..

Thanks,
Kiall

On 06/08/13 07:56, Mike Perez wrote:
> On Mon, Aug 5, 2013 at 11:03 AM, Mac Innes, Kiall <kiall at hp.com
> <mailto:kiall at hp.com>> wrote:
>
>     While the topic of WSME is open - Has anyone actually tried using it?
>
> <snip>
>
>
>     I would be very cautious about assuming WSME can support anything we
>     need when the absolute fundamentals of building a REST API are
>     totally MIA.
>
>
> There was a whole thread about this a couple of months ago [1].
>
> tl;dr Ceilometer is already using it. I have a rough patch for what
> would be v3 of Cinder using Pecan/WSME for J if we decide we need a bump
> [2]. Neutron will likely be using it when it switches to Pecan.
>
> For Cinder, WSME raises a 400 on type checking in the body as I need it
> to. Everything else I have raised in the controller as needed.
>
> Thanks,
> Mike Perez
>
> [1] -
> http://lists.openstack.org/pipermail/openstack-dev/2013-June/009824.html
> [2] -
> http://lists.openstack.org/pipermail/openstack-dev/2013-June/010857.html



More information about the OpenStack-dev mailing list