DevUI pages for Hibernate ORM

[TomasHofman]

• provide information about persistence units and their managed entities
• allow user to generate create-schema script

Fixes #16553

[TomasHofman]

@Sanne I linked a PR with the first draft of the DevUI pages.

Vision: I think the useful things to show in the DevUI could be:

• a list of existing persistence units,
• a detail page for each PU showing a list of entities managed by given PU and a create-schema script.

Problems:

In order to generate the schema via the SchemaExport tool, I need to obtain the Metadata and ServiceRegistry instances of the persistence provider. I spend some time to figure out how to get them. I think my approach might not be entirely wrong but there are some further issues that need to be worked out.

• The implementation creates a new instance of FastBootHibernatePersistenceProvider to generate the schema and adds some non API methods to this class.
• The PreconfiguredServiceRegistryBuilder#buildNewServiceRegistry() method fails when called repeatedly (to generate new PersistenceProvider instance). I worked around it and added a comment to the place, but not sure if this is OK.
• The PersistenceUnitsHolder#popRecordedState() throws away the metadata after the PU are created, which makes it impossible to generate the PersistenceProvider repeatedly. Can the metadata be kept in dev-mode?

[Sanne]

@TomasHofman ! sorry for the delay, I've been off - need to catch up on things and then I'll have a better look at this. Thanks!

[TomasHofman]

@Sanne thanks and no problem, I'm catching up on other stuff too...

[TomasHofman]

I did some more work on the HTML, attaching screenshots:

[geoand]

I can't comment on any of this stuff, but having a DevUI entry for Hibernate ORM with such info is an awesome idea!

[Sanne]

It's starting to look cool :) But the ORM initialization code is a bit complex, I'll try to find some time to help with that.

I'm thinking there's a lot of functionality which we could add, but it's probably best to split it in smaller tasks.
For example, we could have:

• also the "drop" script
• a button to reset the DB state (actually runs the drop script, followed by the create and optionally the import.sql if there's any)

But IMO such additional things could be done via follow-up PRs, not least to get you some help from other people as well.
Let's focus first on the ORM boostrap changes and get the basics integrated first.

The "Names Query" table probably needs a column to show to which persistence unit it belongs; or alternatively maybe it's best to show multiple tables, one each for each PU.

[TomasHofman]

I can squash the commits when we are done, but I will keep adding updates in separate commits for now to make it easier to review.

[TomasHofman]

The "Names Query" table probably needs a column to show to which persistence unit it belongs; or alternatively maybe it's best to
show multiple tables, one each for each PU.

Good idea, I would use the separate-table-for-each-PU approach for "Managed entities" page too, so that the pages are consistent.

[yrodiere]

@TomasHofman @Sanne I pushed a commit to change how we retrieve the information: instead of being pulled at runtime, it's now generated on bootstrap. This works around all the problems related to generation of the schema after bootstrap.

Theoretically it should only have an impact in dev mode (not in tests nor in production), but let's wait for the build results :)

I also took the liberty of rebase this PR on the main branch since I don't think it makes sense to work on 1.13 right now, but let me know if that was a mistake.

@TomasHofman I know it's been a while, but do you think you could address the remaining comments? Feel free to ping me if you need help.

I believe y'all wanted to move to "one page per persistence unit", but after that's done I think it's going to be a great start and we can merge it?

[TomasHofman]

@yrodiere that sounds great. I will take a look hopefully next week when I'm back from PTO and try to move it forward. Thanks!

Definitely the master branch - I had some problems with it at the time, which is why I started to work on 1.13.

[Sanne]

Sounds great, many thanks @TomasHofman

[TomasHofman]

Hello, I updated the PR with additional changes.

• Data for each PU are shown in separate tables now.
• Showing both create and drop SQL scripts for each PU.
• Some formatting changes.

I think this should cover all the requested changes, apart from providing a button to reset the database, which I would leave to a follow up PR.

Attaching current screenshots:

[yrodiere]

Thanks!

I pushed two more commits to address minor problems (one about dependencies, one about the javascript) and rebased it all on main.

Good to merge as soon as CI finishes the build.

Addressed concerns.

[Sanne]

Awesome, many thanks @TomasHofman and @yrodiere ! Lots of people will find this useful.

[gsmet]

Would it be a problem to add a small test as done here: https://github.com/quarkusio/quarkus/pull/18738/files ?

The idea is mostly to check we don't break the pages with a 500 at some point.

[TomasHofman]

I will have a look at the test.

[quarkus-bot]

This workflow status is outdated as a new workflow run has been triggered.

Failing Jobs - Building 5bec537

Status	Name	Step	Test failures	Logs	Raw logs
✔️	JVM Tests - JDK 11
✖	JVM Tests - JDK 11 Windows	Build	Test failures	Logs	Raw logs
✔️	JVM Tests - JDK 16

Full information is available in the Build summary check run.

Test Failures

⚙️ JVM Tests - JDK 11 Windows #

📦 extensions/hibernate-orm/deployment

✖ io.quarkus.hibernate.orm.sql_load_script.IntroducingDefaultImportScriptTestCase.testAddNewImportSql - More details - Source on GitHub

java.lang.RuntimeException: 
java.lang.RuntimeException: java.lang.RuntimeException: java.lang.RuntimeException: java.lang.RuntimeException: java.lang.RuntimeException: io.quarkus.builder.BuildException: Build failure: Build failed due to errors
	[error]: Build step io.quarkus.netty.deployment.NettyProcessor#build threw an exception: java.lang.OutOfMemoryError: Metaspace
	at java.base/java.lang.ClassLoader.defineClass1(Native Method)
	at java.base/java.lang.ClassLoader.defineClass(ClassLoader.java:1017)
	at io.quarkus.bootstrap.classloading.QuarkusClassLoader.loadClass(QuarkusClassLoader.java:445)
	at io.quarkus.bootstrap.classloading.QuarkusClassLoader.loadClass(QuarkusClassLoader.java:405)
	at io.netty.buffer.UnpooledUnsafeDirectByteBuf._setByte(UnpooledUnsafeDirectByteBuf.java:183)
	at io.netty.buffer.AbstractByteBuf.writeByte(AbstractByteBuf.java:985)
	at io.netty.handler.codec.http.HttpObjectEncoder.<clinit>(HttpObjectEncoder.java:53)
	at java.base/java.lang.Class.forName0(Nati...

[TomasHofman]

I added test test and a fix for a problem where the import.sql file name was not determined correctly. Waiting for CI to complete.

[gsmet]

@TomasHofman small detail: could we change the ordering with a comparator to get the <default> one first and then the named ones in alphabetical order?

[gsmet]

Also I wonder if I wouldn't make the scripts in the persistence units view foldable (and folded by default). It sorta works for a small db but will be more challenging with a bunch of entities. I think we have an example in the score page of the RESTEasy Reactive Dev UI.

I would add a simple icon to fold/unfold. It would work well with your existing UI, you could put the copy button and the folding button at the end of the Create/Drop Script line, that way, we could copy even if folded.

[quarkus-bot]

Failing Jobs - Building 66e1b37

Status	Name	Step	Test failures	Logs	Raw logs
✖	Native Tests - Misc4	Build	Test failures	Logs	Raw logs

Full information is available in the Build summary check run.

Test Failures

⚙️ Native Tests - Misc4 #

📦 integration-tests/gradle

✖ io.quarkus.gradle.nativeimage.CustomNativeTestSourceSetIT.runNativeTests line 18 - More details - Source on GitHub

org.opentest4j.AssertionFailedError: 

expected: "SUCCESS"
 but was: null
	at java.base/jdk.internal.reflect.NativeConstructorAccessorImpl.newInstance0(Native Method)
	at java.base/jdk.internal.reflect.NativeConstructorAccessorImpl.newInstance(NativeConstructorAccessorImpl.java:62)
	at java.base/jdk.internal.reflect.DelegatingConstructorAccessorImpl.newInstance(DelegatingConstructorAccessorImpl.java:45)
	at io.quarkus.gradle.nativeimage.CustomNativeTestSourceSetIT.runNativeTests(CustomNativeTestSourceSetIT.java:18)
	at java.base/jdk.internal.reflect.NativeMethodAccessorImpl.invoke0(Native Method)
	at java.base/jdk.internal.reflect.NativeMethodAccessorImpl.invoke(NativeMethodAccessorImpl.java:62)
	at java.base/jdk.internal.reflect.DelegatingMethodAccessorImpl.invoke(DelegatingMethodAccessorImpl.java:43)
	at java.base/java.lang.reflect.Method.invoke(Method.java:566)
	at org.junit.platform.commons.util.ReflectionUtils.invokeMethod(ReflectionUtils.java:688)
	at org.junit.jupiter.engine....

[gsmet]

Let's merge this, it's a very good step forward. We can tweak things in follow-up PRs. Thanks!

[TomasHofman]

Also I wonder if I wouldn't make the scripts in the persistence units view foldable (and folded by default). It sorta works for a small db but will be more challenging with a bunch of entities. I think we have an example in the score page of the RESTEasy Reactive Dev UI.

I would add a simple icon to fold/unfold. It would work well with your existing UI, you could put the copy button and the folding button at the end of the Create/Drop Script line, that way, we could copy even if folded.

Sounds like a good idea. I will create a new PR for that in couple of days after finish other work.

Thanks!