[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]

Re: querying non-existing cartridge is not erroring?



On 01/29/2014 05:03 PM, Clayton Coleman wrote:
Let me answer with a question - what do you need the data for obsolete cartridges for?  If it's for apps that already have that cart I would recommend using the app cartridges list.

I have no clue about the value of deprecated cartridges. I got a PR on the openshift-java-client where I was told that the reasoning for it was to be able to query deprecated cartridges that would not show up in the list of available cartridges:
https://github.com/openshift/openshift-java-client/pull/109

If I misunderstood the reasoning then I'd love to know what the reason for adding this is? The client lib tries to avoid unnecessary queries and caches the list of available cartridges. If this PR is only about getting a cartridge by name then I'd change the suggested change to query the cached list instead of querying the backend.


On Jan 29, 2014, at 10:43 AM, André Dietisheim <adietish redhat com> wrote:

I'd like to verify an assumption which I'm not sure is right:

The curl (/cartridges?id=<cartname>) is as far as I can understand querying cartridges by name. I was assuming that deprecated cartridges (I was told that you'll deprecate carts in the future) could be queried in this way while they would not show up in /cartridges.
Is this correct?

On 01/29/2014 04:15 PM, Clayton Coleman wrote:

On Jan 29, 2014, at 10:13 AM, André Dietisheim <adietish redhat com> wrote:

On 01/29/2014 04:09 PM, Clayton Coleman wrote:
We shouldn't change legacy behavior though - cartridges/:id should behave like old code, but cartridge/:id should behave like new, and if we need a search behavior we should use a different attribute (like name, since carts don't have public ids today).  Did cartridges/:id change for old clients in a breaking way?
for openshift-java-client there's no breakage since this was not implemented before (I got this via a new patch I was verifying and writing tests for).
We really shouldn't change that old behavior - I have some changes landing soon that restore the old behavior and add ?name= prefix search under /cartridges.

On Jan 29, 2014, at 10:01 AM, Jordan Liggitt <jliggitt redhat com> wrote:

Typically, when using a filter which can return multiple results (e.g. https://openshift.redhat.com/broker/rest/cartridges.json?id=nodejs), an empty result set is expected rather than a 404.




On 01/29/2014 09:51 AM, André Dietisheim wrote:
Hi LIli, hi all

I was writing tests for the new query to get infos about a cartridge. I stumbled upon the following: When I do

curl -H "Accept: application/json; version=1.2" --user adietish redhat com:1q5T9o3E$ https://openshift.redhat.com/broker/rest/cartridges?id=bogus -X GET | json_reformat

I dont get 404 as I had expected but the following:

{
    "api_version": 1.2,
    "data": [

    ],
    "messages": [
        {
            "exit_code": 0,
            "field": null,
            "index": null,
            "severity": "info",
            "text": "List bogus cartridges"
        }
    ],
    "status": "ok",
    "supported_api_versions": [
        1.0,
        1.1,
        1.2,
        1.3,
        1.4,
        1.5,
        1.6
    ],
    "type": "cartridges",
    "version": "1.2"
}

Is this expected or a bug as I tend to think? If this is a bug, was this filed or should I?

Thanks
André

_______________________________________________
dev mailing list
dev lists openshift redhat com
http://lists.openshift.redhat.com/openshiftmm/listinfo/dev
_______________________________________________
dev mailing list
dev lists openshift redhat com
http://lists.openshift.redhat.com/openshiftmm/listinfo/dev


[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]