Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Support Java 21 main launch protocol #36164

Merged
merged 1 commit into from Oct 2, 2023
Merged

Conversation

geoand
Copy link
Contributor

@geoand geoand commented Sep 26, 2023

This solution is based on transforming the main class so call-sites don't need to change

Closes: #36154
Alternative to: #36158

@geoand geoand changed the title Support Java 21 main launch protocol in prod and dev mode Support Java 21 main launch protocol Sep 26, 2023
@geoand geoand marked this pull request as ready for review September 27, 2023 12:55
@geoand
Copy link
Contributor Author

geoand commented Sep 27, 2023

P.S. we probably want to add some tests for this, but I am very short on time this week. If I do have find some, I'll add them in a follow up PR

@Ladicek
Copy link
Contributor

Ladicek commented Sep 27, 2023

Interesting thing about this is that we'll support the new styles of writing main even on older JDKs :-)

(We could also easily support private main methods, but I agree we shouldn't.)

@geoand
Copy link
Contributor Author

geoand commented Sep 27, 2023

Interesting thing about this is that we'll support the new styles of writing main even on older JDKs :-)

Yeah, I thought that was funny as well. I'll add the noteworthy-label for that reason alone :)

@geoand
Copy link
Contributor Author

geoand commented Sep 27, 2023

P.S. we probably want to add some tests for this, but I am very short on time this week. If I do have find some, I'll add them in a follow up PR

Actually, I see we have similar QuarkusProdModeTest type tests, so hopefully I can add some tests in a few minutes Done

@quarkus-bot

This comment has been minimized.

@Ladicek
Copy link
Contributor

Ladicek commented Sep 28, 2023

I just realized there's one or two things this PR doesn't do and at least one of them it should.

One, the main method can be protected. That's easy to solve, since changing modifiers is just a bunch of bitwise operations on an int, so you just do removeModifiers(PROTECTED) in addition to addModifiers(PUBLIC) and there's no need for additional check.

Two, an instance main can be inherited from a superclass. Not sure if we wanna do anything about that, I guess we can just leave that unsupported. There's an interesting interaction between static main declared directly on the class and an instance main inherited from a superclass, so it's probably better to just ignore inherited main than implement it incorrectly.

@geoand
Copy link
Contributor Author

geoand commented Sep 28, 2023

so you just do removeModifiers(PROTECTED) in addition to addModifiers(PUBLIC) and there's no need for additional check.

Good point!

an instance main can be inherited from a superclass

Ah darn... I do see that https://openjdk.org/jeps/445 mentions a superclass as well... Let me chek what we can do about it..

@geoand
Copy link
Contributor Author

geoand commented Sep 28, 2023

@Ladicek WDYT now? I think the superclass case should be handled now

@geoand
Copy link
Contributor Author

geoand commented Sep 28, 2023

One thing I am not 100% certain of is the proper order of lookups when there are both main(String[] args) and main() and superclasses involved.
However I think that's an extreme corner case that we can fix later if need be.

@quarkus-bot

This comment has been minimized.

@quarkus-bot

This comment has been minimized.

@Ladicek
Copy link
Contributor

Ladicek commented Sep 29, 2023

The JEP says:

When launching a class, the launch protocol chooses the first of the following methods to invoke:

  • A static void main(String[] args) method of non-private access (i.e., public, protected or package) declared in the launched class,
  • A static void main() method of non-private access declared in the launched class,
  • A void main(String[] args) instance method of non-private access declared in the launched class or inherited from a superclass, or, finally,
  • A void main() instance method of non-private access declared in the launched class or inherited from a superclass.

That seems quite straightforward to me:

  1. If the class declares static void main(String[]), that's the right one.
  2. Otherwise, if the class declares static void main(), that's the right one.
  3. Otherwise, if the class declares or inherits void main(String[]), that's the right one.
  4. Otherwise, if the class declares or inherits void main(), that's the right one.
  5. Otherwise, there's no main.

@geoand
Copy link
Contributor Author

geoand commented Sep 29, 2023

Okay, let's try that

Copy link
Contributor

@Ladicek Ladicek left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I would structure the code differently, but it seems OK. The only objection I would have is that there's no test for a combination of multiple valid main methods (where the correct one must be picked).

@geoand
Copy link
Contributor Author

geoand commented Sep 29, 2023

Yeah, currentry it does not behave correctly, so I'll need to fix it

@geoand
Copy link
Contributor Author

geoand commented Sep 29, 2023

Should be fixed now (I also added a test to demonstrate it)

@quarkus-bot

This comment has been minimized.

@quarkus-bot

This comment has been minimized.

@geoand
Copy link
Contributor Author

geoand commented Oct 2, 2023

@Ladicek ok to merge?

@Ladicek
Copy link
Contributor

Ladicek commented Oct 2, 2023

If I'm reading the code correctly, if the main class defined these 2 methods:

  • void main(String[])
  • static void main()

We would prefer the 1st, even though we really should prefer the 1st, as specified in https://docs.oracle.com/javase/specs/jls/se21/preview/specs/unnamed-classes-instance-main-methods-jls.html#jls-12.1.4

@geoand
Copy link
Contributor Author

geoand commented Oct 2, 2023

Ah, that's an interesting edge case indeed...

@geoand
Copy link
Contributor Author

geoand commented Oct 2, 2023

I'm gonna rethink how this is done

@Ladicek
Copy link
Contributor

Ladicek commented Oct 2, 2023

I would just write an extremely stupid code like this:

ClassVisitor result = findStaticMainWithStringArgs();
if (result != null) {
    return result;
}
result = findStaticMainWithoutArgs();
if (result != null) {
    return result;
}
result = findInstanceMainWithStringArgs();
if (result != null) {
    return result;
}
result = findInstanceMainWithoutArgs();
if (result != null) {
    return result;
}
throw ...

@geoand
Copy link
Contributor Author

geoand commented Oct 2, 2023

That's where I am headed - at least unless I find any other problems :)

This solution is based on transforming the main class
so call-sites don't need to change

Closes: quarkusio#36154
Alternative to: quarkusio#36158
@geoand
Copy link
Contributor Author

geoand commented Oct 2, 2023

It was slighly more complicated, but not much.

I think that anything other than perhaps extreme edge cases should now be covered

Copy link
Member

@rsvoboda rsvoboda left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Solves #36154

Thanks @Ladicek and @geoand for the thorough discussion about possible cases + nice impl with attached tests

@geoand
Copy link
Contributor Author

geoand commented Oct 2, 2023

Thanks for reporting and checking!

Copy link
Contributor

@Ladicek Ladicek left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I'm not a fan of renaming those private mains to $originalMain$, that seems fairly risky and I think it would be better to bail out if we detect a collision. But other than that, it seem fine.

@geoand
Copy link
Contributor Author

geoand commented Oct 2, 2023

I went back and forth on that.
The reason I did it was because I could not find anywhere in the JEP what is supposed to happen in this case.

Bailing out would definitely have been easier from an implementation standpoint 😎

@quarkus-bot
Copy link

quarkus-bot bot commented Oct 2, 2023

Failing Jobs - Building fd9f134

Status Name Step Failures Logs Raw logs Build scan
Native Tests - Messaging1 Build Failures Logs Raw logs

Full information is available in the Build summary check run.

Failures

⚙️ Native Tests - Messaging1 #

- Failing: integration-tests/reactive-messaging-kafka 

📦 integration-tests/reactive-messaging-kafka

io.quarkus.it.kafka.KafkaConnectorIT.testDataForKeyed - More details - Source on GitHub

java.lang.RuntimeException: 
java.lang.RuntimeException: io.quarkus.builder.BuildException: Build failure: Build failed due to errors
	[error]: Build step io.quarkus.kafka.client.deployment.DevServicesKafkaProcessor#startKafkaDevService threw an exception: java.lang.RuntimeException: org.testcontainers.containers.ContainerLaunchException: Container startup failed for image docker.io/vectorized/redpanda:v22.3.4

@geoand geoand merged commit 089bd28 into quarkusio:main Oct 2, 2023
50 of 51 checks passed
@geoand geoand deleted the #36154-take2 branch October 2, 2023 15:43
@quarkus-bot quarkus-bot bot added this to the 3.5 - main milestone Oct 2, 2023
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Projects
None yet
Development

Successfully merging this pull request may close these issues.

Quarkus and Java 21 Instance Main Methods is not working
3 participants