Resilient enum for extension metadata & remove 'code' tag from resteasy-reactive-xxx

[ia3andy]

• Adding new enum values in the extension metadata will not break compatibility after this change, it will now use the default if an invalid value is found.

• Currently, in code.quarkus.io all the resteasy-reactive extensions have the code tag. This is because we needed a way to avoid having the classic resteasy starter code with those extensions. As a workaround we used the core kind which isn't showing the code tag but still referencing the codestart. In a few versions we will be able to introduce default-codestart without much risk.

[ia3andy]

@aloubyansky If I get it right, to avoid any compatibility problem, we should introduce this with a minor version change (and not a patch). By doing this:

• It won't fail using a older version of a Quarkus tool as they will not automatically use a different minor version.
• It will fail if manually specifying a more recent version with an older Quarkus tool.
• We will need to update code.quarkus.io tooling before updating the registry (which is ok as it uses the core bom)

[aloubyansky]

No, our command line tools are consuming whatever the registry serves.

[aloubyansky]

I am still not sure what makes it a breaking change.

[ia3andy]

the legacy toEnum will throw an exception because it doesn't know default-codestart :-/

[maxandersen]

we should not be using hardcoded enums for things that are meant to be extendable.
is there no way we can introduce this without breaking things ? or at least make a patch to 2.2.x that won't break when this is out ?

[maxandersen]

besides that - im not following why not showing "code" tag is important ? like what is this change actually for ?

[ia3andy]

@maxandersen

we should not be using hardcoded enums for things that are meant to be extendable.
is there no way we can introduce this without breaking things ? or at least make a patch to 2.2.x that won't break when this is out ?

This patch will make it safe in the future.

besides that - im not following why not showing "code" tag is important ? like what is this change actually for ?

Because they don't provide code (fake news :-p)

I might have a way to make it non breaking. It will change the PR.

Also, as we might be forced to break some day, we should introduce an api-version that we increment on breaking change (and that we keep in the stream metatdata). When the api-version is lower in the client, we display a warning to tell the user it may fail and they should update to a recent version.

[ia3andy]

I created an issue for #20323

[maxandersen]

This patch will make it safe in the future.

thats great but with 2.2 being what we support now long term we do need to limit what we break here. At least we need to quite careful.

besides that - im not following why not showing "code" tag is important ? like what is this change actually for ?

Because they don't provide code (fake news :-p)

but code is not about them providing code starts just that a codestart is available for them,right?

Maybe my question is what does default-codestart mean compared to what we already have?

I might have a way to make it non breaking. It will change the PR.

Great - let me know when ready for review.

[ia3andy]

This patch will make it safe in the future.

thats great but with 2.2 being what we support now long term we do need to limit what we break here. At least we need to quite careful.

besides that - im not following why not showing "code" tag is important ? like what is this change actually for ?

Because they don't provide code (fake news :-p)

but code is not about them providing code starts just that a codestart is available for them,right?

Maybe my question is what does default-codestart mean compared to what we already have?

Well for example, resteasy-qute provides a code, while resteasy-reactive-qute only provide the resteasy-react code (default), like any other extension btw (which provides the resteasy code by default)

I might have a way to make it non breaking. It will change the PR.

Great - let me know when ready for review.

it's ready for review now.

[quarkus-bot]

Failing Jobs - Building 8b1e0fc

Status	Name	Step	Failures	Logs	Raw logs
✔️	JVM Tests - JDK 11
⌛	JVM Tests - JDK 16	Build	⚠️ Check →	Logs	Raw logs
✖	MicroProfile TCKs Tests	Verify	Failures	Logs	Raw logs

Full information is available in the Build summary check run.

Failures

⚙️ MicroProfile TCKs Tests #

- Failing: tcks/resteasy-reactive 

📦 tcks/resteasy-reactive

✖ Failed to execute goal org.codehaus.mojo:exec-maven-plugin:3.0.0:exec (test) on project quarkus-tck-resteasy-reactive: Command execution failed.

📦 tcks/resteasy-reactive/target/testsuite/tests

✖ com.sun.ts.tests.jaxrs.platform.container.completioncallback.JAXRSClient0164.argumentIsNullWhenRegistredClassTest line 209 - More details - Source on GitHub

com.sun.ts.tests.jaxrs.common.JAXRSCommonClient$Fault: Unexpected response content No name has been set yet expecting NULL
	at com.sun.ts.tests.jaxrs.platform.container.completioncallback.JAXRSClient0164.assertString(JAXRSClient0164.java:458)
	at com.sun.ts.tests.jaxrs.platform.container.completioncallback.JAXRSClient0164.argumentIsNullWhenRegistredClassTest(JAXRSClient0164.java:209)