sipkab / jvm-tail-recursion Goto Github PK

View Code? Open in Web Editor NEW

121.0 4.0 1.0 446 KB

Optimizer library for tail recursive calls in Java bytecode

License: GNU General Public License v3.0

Java 100.00%

jvm-bytecode tail-recursion optimization-tools saker-build

jvm-tail-recursion's People

Contributors

Stargazers

Watchers

Forkers

zeta1999

jvm-tail-recursion's Issues

Benchmark

Are there any benchmarks showing the effect of applying the tail-recursion optimizations?

Optimize recursive calls when unoptimizable branches are present in the return path

There may be cases when a tail recursive call has multiple execution paths after the call itself. The optimization should be performed in cases where it is possible.

One example is such:

public static int count(int n) {
	if (n == 0) {
		return 0;
	}
	int v = count(n - 1);
	if (n == 0) {
		callSomeOtherMethod();
	}
	return v;
}

Should be optimized to:

public static int count(int n) {
start:
	if (n == 0) {
		return 0;
	}
	
	if (n == 0) {
		int v = count(n - 1);
		callSomeOtherMethod();
		return v;
	}
	n = n - 1;
	goto start;
}

Current state

Currently we allow branches to be present in the return path, but they all need to be side-effect free. This issue requires individual modifications and analysis to the subsequent execution paths.

Any advantage of using this with Kotlin?

I wonder how does this compare to kotlin tailrec keyword implementation.

Implement as a SpotBugs detector

Seems like this would be a good spotbugs check to identify recursion that could easily be improved by the programmer. Spotbugs doesn't modify the output but can produce a warning when such a condition exists which would be helpful.

https://spotbugs.github.io/

ASM error: Class versions V1_5 or less must use F_NEW frames.

I'm getting this error:

Exception in thread "main" java.lang.IllegalArgumentException: Class versions V1_5 or less must use F_NEW frames.
        at sipka.jvm.tailrec.thirdparty.org.objectweb.asm.MethodWriter.visitFrame(MethodWriter.java:779)
        at sipka.jvm.tailrec.thirdparty.org.objectweb.asm.tree.FrameNode.accept(FrameNode.java:141)
        at sipka.jvm.tailrec.thirdparty.org.objectweb.asm.tree.InsnList.accept(InsnList.java:145)
        at sipka.jvm.tailrec.thirdparty.org.objectweb.asm.tree.MethodNode.accept(MethodNode.java:752)
        at sipka.jvm.tailrec.thirdparty.org.objectweb.asm.tree.MethodNode.accept(MethodNode.java:650)
        at sipka.jvm.tailrec.thirdparty.org.objectweb.asm.tree.ClassNode.accept(ClassNode.java:466)
        at sipka.jvm.tailrec.TailRecursionOptimizer.optimizeMethods(TailRecursionOptimizer.java:114)
        at sipka.jvm.tailrec.TailRecursionOptimizer.optimizeMethods(TailRecursionOptimizer.java:92)
        at sipka.jvm.tailrec.MainCommand.optimizeJar(MainCommand.java:208)
        at sipka.jvm.tailrec.MainCommand.optimizeFile(MainCommand.java:182)
        at sipka.jvm.tailrec.MainCommand.call(MainCommand.java:139)
        at sipka.jvm.tailrec.Main.lambda$parse$1(Main.java:166)
        at sipka.jvm.tailrec.Main.callCommand(Main.java:195)
        at sipka.jvm.tailrec.Main.main(Main.java:177)

Version: v0.8.1
Java version used to run: Java 15

The target JAR uses Java 8 as far as I know, but it has some libraries that should be backwards compatible or such for my understanding, otherwise who would on the earth still uses Java 5? There is Java 15!

sipkab / jvm-tail-recursion Goto Github PK

jvm-tail-recursion's People

Contributors

Stargazers

Watchers

Forkers

jvm-tail-recursion's Issues

Benchmark

Optimize recursive calls when unoptimizable branches are present in the return path

Any advantage of using this with Kotlin?

Implement as a SpotBugs detector

ASM error: Class versions V1_5 or less must use F_NEW frames.

Recommend Projects

React

Vue.js

Typescript

TensorFlow

Django

Laravel

D3

Recommend Topics

javascript

web

server

Machine learning

Visualization

Game

Recommend Org

Facebook

Microsoft

Google

Alibaba

D3

Tencent