Conversation
There was a problem hiding this comment.
Running this through a semantic diff tool, there is a bunch of added exampleSetFlag items, and the description items from the moar-openapi.json file above.
|
So with the layout of the tests right now, the only way to test the plugin is to either release or install the plugin, and then run the tests with that version. The latter means that we can get rid of the separate test for the plugin and change the instructions for the main tests to We could restructure the plugin so it's part of the same versioning and parent pom as the other items in the repository, like the Any thoughts? I'm going on holidays at the end of this week so won't be able to do anything after that. |
|
@rbygrave I'm still on holidays but will be back soon. What are your thoughts on making the plugin versioned alongside the rest of the project? We can then also spin the jsonb stuff for the OpenAPI schema into a seperate module and then the |
|
This PR is still DRAFT per se ... but yeah I'm pretty happy with it.
Maybe sort out this PR first?
Yes, could be a good followup. |
|
The thing blocking this PR is (mostly) the lack of testing, and that requires the plugin be versioned alongside the rest of the project. But I'll take that as approval to do that in order to get the tests going 😁 |
Resolves #51
I've marked this as draft in case you want to change the approach taken, and because I've yet to fully finish the testing. The remaining changes are:
testsmodule that builds off oftest-javalin-jsonbso we can check it correctly pulls in from librariesI'm also not sure if this should be a separate goal or just part of the existing one.
I was also thinking possibly (as a second PR) of moving the json serialisation of the OpenAPI spec to a library, and the
http-generator-corecan use that to serialise/deserialiseAnother thing I thought of as I was working on this is that this change works for libraries that have OpenAPI definitions, but not for ones in which there's no javadoc available for the annotation processor to parse... Is it worth adding the ability to parse
Schemaannotations as well?