[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]
Subject: Re: [odata] RE: [EXTERNAL] [odata] Status of the Rapid
> From my part, I'd like to better understand how a GraphQL client could call an OData service.Â There have been a few investigations at Microsoft and in the industry, such as the use of "GraphQL Mesh", which has an OData handler (https://graphql-mesh.com/docs/handlers/odata/).
Mesh and other solutions like this are using GraphQL resolvers (separate gateway server) as a proxy and OData is just used here as REST API that is called to get results.
This is used to move OData APIs to GraphQL in situations where developers cannot invest to rewrite those to use GraphQL resolvers.
They getting GraphQL API with the cost of the extra server to maintain, but do not need to rewrite their data layers.
A more interesting example was done by Microsoft (https://github.com/microsoftgraph/graphql-demo)
This is actually doing transformation on client. This approach uses only OData server (OData first - comparing to Mesh which is GraphQL first)
It looks like it was abandoned because main dev moved to Facebook :)
More info and ideas here:Âhttps://github.com/oasis-open/odata-rapid/issues/7
The previous two meetings have been largely focused on aggregation extensions.
Today we discussed binary formats for about 30 minutes and spent the rest of the time walking through the RSDL.Â Christof has a PR that he made against the standard-api repo that he'll copy over to the oasis repo and update based on today's discussion.
As far as the existing packages, it may still be interesting to be able to generate CSDL from a GraphQL compatible syntax, not as a core part of RAPID but as part of our "integration with other technologies".Â What do other folks think?
Both Christof and I are out next week, but assuming that Ralf can chair and we have quorum the main agenda item should be going through the backlog of items in the RAPID repo to prioritize and figure out who can work on what moving forward.
From my part, I'd like to better understand how a GraphQL client could call an OData service.Â There have been a few investigations at Microsoft and in the industry, such as the use of "GraphQL Mesh", which has an odata handler (https://graphql-mesh.com/docs/handlers/odata/).
Does that help?
Due to the conflicting call, I have at the time (GraphQL Spec group and Managers meeting) I could not attend fully last couple calls.Â
Do we have any TL:DR on what is the direction Rapid is going and how individuals like me can help.
I seen we have PR created but there is not much traction to get it merged.
From my point of view ideally, we should be:
- Deprecating existing packages (those on NPM)
- Clear repositoryÂfrom the packages that were created
- Review and Push C# based parser - I assume that this is 100% go?
- Put new plan into the repo and ramp up progress.