Some questions about indexing speed?

[94kai]

What are the differences in indexing speed optimization between importing an Android project using kotlin-lsp and importing a project using Android Studio? I have a very large Android project with the classpath already prepared. Using Android Studio only takes a few minutes, but with kotlin-lsp, indexing hasn't finished even after about 20 minutes.

Additionally, when reopening the same project repeatedly, each indexing process takes nearly the same amount of time. Is there a caching mechanism for this?

[yamsergey]

I guess infinite loading of android project happening because it's not supported and seems to stay so for quite some time #26. Android projects are not usual Gradle or Java/Kotlin projects, they require Android Gradle Plugin to operate. However, I did an experiment by generating workspace.json and it works just fine while indexing doesn't take more than a couple of minutes in worst case scenarios, I'm preparing to open the repository with the generator.

[94kai]

Our project has 1000 + modules if a full workspace is generated， kotlin-lsp throw OOM (Max heap memory 100G). I tried to use a certain module as root, and the generated workspace contains 500 modules. Kotlin-lsp completed the index in 3 minutes, using 50G memory. Using androidstudio to index the whole project, only 8G memory is used.

[94kai]

I suspect this may be related to IntelliJ IDEA's indexing mechanism. IDEA likely employs a more intelligent preloading strategy that establishes fewer indexes during the initialization phase compared to kotlin-lsp. On the other hand, IDEA also maintains some index caching.
I hope the kotlin-lsp author can help explain it. Thanks

[adamtmorgan]

kotlin-lsp throw OOM

Hitting the same hurdle with a large gradle project. Of course, the same project loads fine in IntelliJ.

Here's the log entry for reference (running in neovim):

Error importing project:\nkotlinx.coroutines.JobCancellationException: Parent job is Cancelling; 
job=UndispatchedCoroutine{Cancelling}@108567cc\nCaused by: java.lang.OutOfMemoryError: Java heap space\n

[94kai]

You can modify the startup script of kotlin-lsp and configure JVM parameters to increase the heap memory. However, this can only address the symptoms rather than the root cause. The core issue is that kotlin-lsp consumes too much memory, and the author needs to explain the reason.

[JSpiner]

same happens to me.
My spring project is not that big and heavy. But I get OOM too.

[Error - 오후 2:57:11] Error importing project:
kotlinx.coroutines.JobCancellationException: Parent job is Cancelling; job=UndispatchedCoroutine{Cancelling}@55c0d7cb
Caused by: java.lang.OutOfMemoryError: Java heap space
	at java.base/java.util.LinkedHashMap.newNode(LinkedHashMap.java:281)
	at java.base/java.util.HashMap.putVal(HashMap.java:637)
	at java.base/java.util.HashMap.put(HashMap.java:618)
	at com.jetbrains.analyzer.index.InMemoryIndex$Companion.computeIndexData(InMemoryIndex.kt:235)
	at com.jetbrains.analyzer.index.InMemoryIndex$Companion.access$computeIndexData(InMemoryIndex.kt:71)
	at com.jetbrains.analyzer.index.InMemoryIndex.update(InMemoryIndex.kt:118)
	at com.jetbrains.ls.snapshot.api.impl.core.IndexerKt$processUpdates$2.invokeSuspend(Indexer.kt:160)
	at kotlin.coroutines.jvm.internal.BaseContinuationImpl.resumeWith(ContinuationImpl.kt:33)
	at kotlinx.coroutines.DispatchedTask.run(DispatchedTask.kt:100)
...

[qwwdfsad]

Should be addressed with 0.253.10629 -- index caches are now consistently pushed on-disk