Index out of bound in extensionless file

[JStronkhorst]

What version of OpenRewrite are you using?

I am using Maven plugin v5.14.0

How are you running OpenRewrite?

I am using the Maven plugin, and my project is a single module project.

 mvn -U -X org.openrewrite.maven:rewrite-maven-plugin:run -D rewrite.recipeArtifactCoordinates=org.openrewrite.recipe:rewrite-migrate-java:RELEASE -D rewrite.activeRecipes=
org.openrewrite.java.migrate.UpgradeToJava21

What is the full stack trace of any errors you encountered?

Error while visiting scripts\jenkins-promote\dev-tst\Jenkinsfile_promoteToTST: java.lang.IndexOutOfBoundsException: Index 0 out of bounds for length 0
  java.base/jdk.internal.util.Preconditions.outOfBounds(Preconditions.java:100)
  java.base/jdk.internal.util.Preconditions.outOfBoundsCheckIndex(Preconditions.java:106)
  java.base/jdk.internal.util.Preconditions.checkIndex(Preconditions.java:302)
  java.base/java.util.Objects.checkIndex(Objects.java:385)
  java.base/java.util.ArrayList.get(ArrayList.java:427)
  org.openrewrite.java.JavaTemplate$Matcher.parameter(JavaTemplate.java:103)
  org.openrewrite.java.migrate.net.URLConstructorsToURIRecipes$URLSingleArgumentConstructorRecipe$1.visitExpression(URLConstructorsToURIRecipes.java:68)
  org.openrewrite.java.migrate.net.URLConstructorsToURIRecipes$URLSingleArgumentConstructorRecipe$1.visitExpression(URLConstructorsToURIRecipes.java:57)

[timtebeek]

Thanks for your report @JStronkhorst ! it seems there's an issue with parsing Jenkinsfile_promoteToTST in a recipe we generate out of URLConstructorsToURI. Can you share any lines in Jenkinsfile_promoteToTST that relate to calling new URL(String)? That would help create an reproducer test for our groovy parser.

[timtebeek]

Meanwhile you can also use exclusions to skip parsing that particular file, such that you can at least apply the other changes while work on a fix for the Groovy parser or recipe.

Or use this modified recipe that does not contain the recipe that fails:

type: specs.openrewrite.org/v1beta/recipe
name: org.openrewrite.java.migrate.Java21
displayName: Migrate to Java 21
description: Excluding the URL constructor recipes.
recipeList:
  - org.openrewrite.java.migrate.UpgradeToJava17
  - org.openrewrite.java.migrate.JavaVersion21
  - org.openrewrite.java.migrate.lang.ThreadStopUnsupported
  - org.openrewrite.java.migrate.util.SequencedCollection
  - org.openrewrite.java.migrate.util.UseLocaleOf
  - org.openrewrite.staticanalysis.ReplaceDeprecatedRuntimeExecMethods
  - org.openrewrite.github.SetupJavaUpgradeJavaVersion

[JStronkhorst]

Hi @timtebeek, thanks for your time. As a work around I temporarly removed that file and ran the recipe and it worked. The problem with the new URL() is in a method from a groovy script in a different repository.

@Library('jenkins-pipeline-library@main') _

import org.svb.PipelineBuilder
import org.svb.GlobalVariables

// Create instance of PipelineBuild to help out with general tasks
def helper = new PipelineBuilder(this)
withCredentials([string(credentialsId: credentialId, variable: 'apiToken')]) {
    response = helper.RestRequestGet("http://${GlobalVariables.ArtifactoryApiBaseUrl_P}/api/docker/${sourceRepository}/v2/${applicationName}/tags/list", apiToken)
    sourceTags = response.get("tags")
}

That calls this method:

def RestRequestGet(String url, String apiToken){
        def responseText = ""
        def conn
        try {
            echo "Creating new connection"
            conn = new URL(url).openConnection()
            conn.setRequestMethod("GET")
            conn.setRequestProperty("Authorization", "Bearer ${apiToken}")
            def postRC = conn.getResponseCode()
            responseText = conn.getInputStream().getText()
            steps.println "Got:" + postRC + "\n" + responseText
        } finally {
            conn.disconnect()
        }
        return new groovy.json.JsonSlurperClassic().parseText(responseText)
    }

[timtebeek]

I've tried to replicate this with the following in GroovyParserTest, but no luck yet; this just works and is similar to what we generate for the URLConstructorsToURI, with the exception of the Semantics.expression used there.

@Test
@Issue("https://github.com/openrewrite/rewrite/issues/3737")
void groovyFirstConstructorArgument() {
    rewriteRun(
      spec -> spec.recipe(toRecipe(() -> (TreeVisitor<?, ExecutionContext>)
        new GroovyIsoVisitor<ExecutionContext>() {
            final JavaTemplate urlConstructor = JavaTemplate
              .builder("new java.net.URL(#{spec:any(java.lang.String)})")
              .build();

            @Override
            public Expression visitExpression(Expression expression, ExecutionContext executionContext) {
                Expression e = super.visitExpression(expression, executionContext);
                JavaTemplate.Matcher matcher;
                if ((matcher = urlConstructor.matcher(getCursor())).find()) {
                    System.out.println("e = " + e);
                    System.out.println("s = " + matcher.parameter(0));
                }
                return e;
            }
        })),
      groovy(
        """
          new URL("https://docs.openrewrite.org")
          """
      )
    );
}

[timtebeek]

When I run the following test in rewrite-migrate-java/...URLConstructorsToURITest

@Test
@Issue("https://github.com/openrewrite/rewrite/issues/3737")
void groovyFirstConstructorArgument() {
    rewriteRun(
      groovy(
        """
          new URL("https://docs.openrewrite.org")
          """,
        """
          import java.net.URI

          URI.create("https://docs.openrewrite.org").toURL()
          """
      )
    );
}

Then I get a different error, but still not what you're seeing

diff --git a/file.groovy b/file.groovy
index 9c3fd2a..e0c4163 100644
--- a/file.groovy
+++ b/file.groovy
@@ -1,3 +1 @@ 
-import java.net.URI
-
-URI.create("https://docs.openrewrite.org").toURL()
\ No newline at end of file
+import java.net.URIURI.create("https://docs.openrewrite.org").toURL()
\ No newline at end of file