[OASIS Issue Tracker] (ODATA-1130) JSON Batch and continue-on-error

[OASIS Issues Tracker <workgroup_mailer@lists.oasis-open.org>]

     [ https://issues.oasis-open.org/browse/ODATA-1130?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Michael Pizzo updated ODATA-1130:
---------------------------------

    Environment: Proposed
       Proposal: allow "continue-on-error=false" and "continue-on-error=true".  "continue-on-error" implies true.  If continue-on-error is not specified, then default is false in multipart mixed, true in json. In multipartmixed, continue-on-error=false means stop if any statement fails. In json it means that a best effort is requested not to spend cycles on further processing if a failure occurs outside of dependency groups because the client is not interested in the result.

> JSON Batch and continue-on-error
> --------------------------------
>
>                 Key: ODATA-1130
>                 URL: https://issues.oasis-open.org/browse/ODATA-1130
>             Project: OASIS Open Data Protocol (OData) TC
>          Issue Type: Improvement
>          Components: JSON Format
>    Affects Versions: V4.01_CSD03
>         Environment: Proposed
>            Reporter: Ralf Handl
>             Fix For: V4.01_CS01
>
>
> The JSON Format specifcation states in http://docs.oasis-open.org/odata/odata-json-format/v4.01/csprd03/odata-json-format-v4.01-csprd03.html#sec_ProcessingaBatchRequest that
> The service MAY process the individual requests and atomicity groups within a batch request in any order that is compatible with the dependencies expressed with the dependsOn name/value pair. Processing stops on the first error unless the continue-on-error preference is specified.
> This makes batch processing somewhat non-deterministic.
> Also it is not clear what "stop processing" means: 
> - If the service chose to process independent groups of requests in parallel, will only groups stop that contain a failing individual request?
> - What is contained in the array of responses for "skipped"/"unprocessed" individual requests?
> Do we need continue-on-error at all for JSON Batch requests?
> If yes: does it only apply to subsequent requests within "dependency chains"?
> Or do we intentionally leave this up to the service implementation and thus non-deterministic?



--
This message was sent by Atlassian JIRA
(v6.2.2#6258)